2024-03-12 13:27:20 +01:00
|
|
|
#define GGML_COMMON_DECL_METAL
|
|
|
|
#define GGML_COMMON_IMPL_METAL
|
|
|
|
#include "ggml-common.h"
|
|
|
|
|
2023-06-25 14:40:30 +02:00
|
|
|
#include <metal_stdlib>
|
|
|
|
|
|
|
|
using namespace metal;
|
|
|
|
|
|
|
|
#define MAX(x, y) ((x) > (y) ? (x) : (y))
|
2023-12-07 21:27:19 +01:00
|
|
|
#define MIN(x, y) ((x) < (y) ? (x) : (y))
|
|
|
|
#define SWAP(x, y) { auto tmp = (x); (x) = (y); (y) = tmp; }
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
#define N_SIMDWIDTH 32 // assuming SIMD group size is 32
|
|
|
|
|
|
|
|
enum ggml_sort_order {
|
2024-04-03 15:07:05 +02:00
|
|
|
GGML_SORT_ORDER_ASC,
|
|
|
|
GGML_SORT_ORDER_DESC,
|
2023-12-07 21:27:19 +01:00
|
|
|
};
|
|
|
|
|
|
|
|
// general-purpose kernel for addition, multiplication and division of two tensors
|
|
|
|
// pros: works for non-contiguous tensors, supports broadcast across all dims
|
2023-11-03 20:35:05 +01:00
|
|
|
// cons: not very efficient
|
2023-06-25 14:40:30 +02:00
|
|
|
kernel void kernel_add(
|
2023-11-03 20:35:05 +01:00
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device char * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
2023-11-03 20:35:05 +01:00
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant uint64_t & nb13,
|
2023-11-03 20:35:05 +01:00
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant int64_t & offs,
|
2023-11-03 20:35:05 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig.z;
|
|
|
|
const int64_t i02 = tgpig.y;
|
|
|
|
const int64_t i01 = tgpig.x;
|
|
|
|
|
|
|
|
const int64_t i13 = i03 % ne13;
|
|
|
|
const int64_t i12 = i02 % ne12;
|
|
|
|
const int64_t i11 = i01 % ne11;
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01 + offs;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11;
|
2023-12-13 20:55:03 +01:00
|
|
|
device char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1 + offs;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) {
|
|
|
|
const int i10 = i0 % ne10;
|
|
|
|
*((device float *)(dst_ptr + i0*nb0)) = *((device float *)(src0_ptr + i0*nb00)) + *((device float *)(src1_ptr + i10*nb10));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_mul(
|
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device char * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant uint64_t & nb13,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
2023-12-07 21:27:19 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig.z;
|
|
|
|
const int64_t i02 = tgpig.y;
|
|
|
|
const int64_t i01 = tgpig.x;
|
|
|
|
|
|
|
|
const int64_t i13 = i03 % ne13;
|
|
|
|
const int64_t i12 = i02 % ne12;
|
|
|
|
const int64_t i11 = i01 % ne11;
|
|
|
|
|
|
|
|
device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01;
|
|
|
|
device const char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11;
|
|
|
|
device char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
|
|
|
for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) {
|
2023-12-07 21:27:19 +01:00
|
|
|
const int i10 = i0 % ne10;
|
|
|
|
*((device float *)(dst_ptr + i0*nb0)) = *((device float *)(src0_ptr + i0*nb00)) * *((device float *)(src1_ptr + i10*nb10));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_div(
|
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device char * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant uint64_t & nb13,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
2023-12-07 21:27:19 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig.z;
|
|
|
|
const int64_t i02 = tgpig.y;
|
|
|
|
const int64_t i01 = tgpig.x;
|
|
|
|
|
|
|
|
const int64_t i13 = i03 % ne13;
|
|
|
|
const int64_t i12 = i02 % ne12;
|
|
|
|
const int64_t i11 = i01 % ne11;
|
|
|
|
|
|
|
|
device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01;
|
|
|
|
device const char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11;
|
|
|
|
device char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) {
|
|
|
|
const int i10 = i0 % ne10;
|
|
|
|
*((device float *)(dst_ptr + i0*nb0)) = *((device float *)(src0_ptr + i0*nb00)) / *((device float *)(src1_ptr + i10*nb10));
|
2023-11-03 20:35:05 +01:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
// assumption: src1 is a row
|
|
|
|
// broadcast src1 into src0
|
|
|
|
kernel void kernel_add_row(
|
2023-09-05 19:57:27 +02:00
|
|
|
device const float4 * src0,
|
|
|
|
device const float4 * src1,
|
|
|
|
device float4 * dst,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb [[buffer(28)]],
|
2023-09-05 12:54:40 +02:00
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
2023-09-05 19:57:27 +02:00
|
|
|
dst[tpig] = src0[tpig] + src1[tpig % nb];
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
kernel void kernel_mul_row(
|
2023-09-05 19:57:27 +02:00
|
|
|
device const float4 * src0,
|
|
|
|
device const float4 * src1,
|
|
|
|
device float4 * dst,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb [[buffer(28)]],
|
2023-06-25 14:40:30 +02:00
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
2023-12-07 21:27:19 +01:00
|
|
|
dst[tpig] = src0[tpig] * src1[tpig % nb];
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
kernel void kernel_div_row(
|
2023-09-05 19:57:27 +02:00
|
|
|
device const float4 * src0,
|
|
|
|
device const float4 * src1,
|
|
|
|
device float4 * dst,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb [[buffer(28)]],
|
2023-06-25 14:40:30 +02:00
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
2023-12-07 21:27:19 +01:00
|
|
|
dst[tpig] = src0[tpig] / src1[tpig % nb];
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_scale(
|
2023-11-03 20:35:05 +01:00
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant float & scale,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
dst[tpig] = src0[tpig] * scale;
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_scale_4(
|
2023-09-15 11:18:18 +02:00
|
|
|
device const float4 * src0,
|
|
|
|
device float4 * dst,
|
2023-11-03 20:35:05 +01:00
|
|
|
constant float & scale,
|
2023-06-25 14:40:30 +02:00
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
dst[tpig] = src0[tpig] * scale;
|
|
|
|
}
|
|
|
|
|
2024-04-14 13:14:19 +02:00
|
|
|
kernel void kernel_clamp(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant float & min,
|
|
|
|
constant float & max,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
dst[tpig] = src0[tpig] < min ? min : (src0[tpig] > max ? max : src0[tpig]);
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
kernel void kernel_relu(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
2023-06-25 14:40:30 +02:00
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
2023-12-13 20:55:03 +01:00
|
|
|
dst[tpig] = max(0.0f, src0[tpig]);
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2024-05-01 23:44:26 +02:00
|
|
|
kernel void kernel_sigmoid(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
dst[tpig] = 1.0f / (1.0f + exp(-src0[tpig]));
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
kernel void kernel_tanh(
|
2023-06-25 14:40:30 +02:00
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
2023-12-13 20:55:03 +01:00
|
|
|
device const float & x = src0[tpig];
|
|
|
|
dst[tpig] = precise::tanh(x);
|
|
|
|
}
|
|
|
|
|
|
|
|
constant float GELU_COEF_A = 0.044715f;
|
|
|
|
constant float GELU_QUICK_COEF = -1.702f;
|
|
|
|
constant float SQRT_2_OVER_PI = 0.79788456080286535587989211986876f;
|
|
|
|
|
|
|
|
kernel void kernel_gelu(
|
2024-04-16 17:40:48 +02:00
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
device const float & x = src0[tpig];
|
|
|
|
|
|
|
|
dst[tpig] = 0.5f*x*(1.0f + precise::tanh(SQRT_2_OVER_PI*x*(1.0f + GELU_COEF_A*x*x)));
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_gelu_4(
|
2023-12-13 20:55:03 +01:00
|
|
|
device const float4 * src0,
|
|
|
|
device float4 * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
device const float4 & x = src0[tpig];
|
|
|
|
|
|
|
|
// BEWARE !!!
|
|
|
|
// Simply using "tanh" instead of "precise::tanh" will sometimes results in NaNs!
|
|
|
|
// This was observed with Falcon 7B and 40B models
|
|
|
|
//
|
|
|
|
dst[tpig] = 0.5f*x*(1.0f + precise::tanh(SQRT_2_OVER_PI*x*(1.0f + GELU_COEF_A*x*x)));
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_gelu_quick(
|
2024-04-16 17:40:48 +02:00
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
device const float & x = src0[tpig];
|
|
|
|
|
|
|
|
dst[tpig] = x*(1.0f/(1.0f+exp(GELU_QUICK_COEF*x)));
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_gelu_quick_4(
|
2023-12-13 20:55:03 +01:00
|
|
|
device const float4 * src0,
|
|
|
|
device float4 * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
device const float4 & x = src0[tpig];
|
|
|
|
|
|
|
|
dst[tpig] = x*(1.0f/(1.0f+exp(GELU_QUICK_COEF*x)));
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_silu(
|
2024-04-16 17:40:48 +02:00
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
device const float & x = src0[tpig];
|
|
|
|
dst[tpig] = x / (1.0f + exp(-x));
|
|
|
|
}
|
|
|
|
|
2024-05-11 15:25:50 +02:00
|
|
|
kernel void kernel_silu_4(
|
2023-12-13 20:55:03 +01:00
|
|
|
device const float4 * src0,
|
|
|
|
device float4 * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
device const float4 & x = src0[tpig];
|
|
|
|
dst[tpig] = x / (1.0f + exp(-x));
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
kernel void kernel_sqr(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
dst[tpig] = src0[tpig] * src0[tpig];
|
|
|
|
}
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
kernel void kernel_sum_rows(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant uint64_t & nb13,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
2023-12-07 21:27:19 +01:00
|
|
|
uint3 tpig[[thread_position_in_grid]]) {
|
|
|
|
int64_t i3 = tpig.z;
|
|
|
|
int64_t i2 = tpig.y;
|
|
|
|
int64_t i1 = tpig.x;
|
|
|
|
|
|
|
|
if (i3 >= ne03 || i2 >= ne02 || i1 >= ne01) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
device const float * src_row = (device const float *) ((device const char *) src0 + i1*nb01 + i2*nb02 + i3*nb03);
|
|
|
|
device float * dst_row = (device float *) ((device char *) dst + i1*nb1 + i2*nb2 + i3*nb3);
|
|
|
|
|
|
|
|
float row_sum = 0;
|
|
|
|
|
|
|
|
for (int64_t i0 = 0; i0 < ne00; i0++) {
|
|
|
|
row_sum += src_row[i0];
|
|
|
|
}
|
|
|
|
|
|
|
|
dst_row[0] = row_sum;
|
|
|
|
}
|
|
|
|
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
template<typename T>
|
2023-06-25 14:40:30 +02:00
|
|
|
kernel void kernel_soft_max(
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device char * dst,
|
2023-06-25 14:40:30 +02:00
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant float & scale,
|
2024-02-19 14:18:09 +01:00
|
|
|
constant float & max_bias,
|
|
|
|
constant float & m0,
|
|
|
|
constant float & m1,
|
|
|
|
constant uint32_t & n_head_log2,
|
|
|
|
threadgroup float * buf [[threadgroup(0)]],
|
2023-11-03 20:35:05 +01:00
|
|
|
uint tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = (tgpig) / (ne02*ne01);
|
|
|
|
const int64_t i02 = (tgpig - i03*ne02*ne01) / ne01;
|
|
|
|
const int64_t i01 = (tgpig - i03*ne02*ne01 - i02*ne01);
|
2023-06-25 14:40:30 +02:00
|
|
|
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
device const float * psrc0 = (device const float *) src0 + (i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00);
|
|
|
|
device const T * pmask = src1 != src0 ? (device const T *) src1 + i01*ne00 : nullptr;
|
|
|
|
device float * pdst = (device float *) dst + (i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00);
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2024-05-11 09:32:41 +02:00
|
|
|
float slope = 1.0f;
|
2024-02-19 14:18:09 +01:00
|
|
|
|
|
|
|
// ALiBi
|
|
|
|
if (max_bias > 0.0f) {
|
|
|
|
const int64_t h = i02;
|
|
|
|
|
|
|
|
const float base = h < n_head_log2 ? m0 : m1;
|
|
|
|
const int exp = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1;
|
|
|
|
|
|
|
|
slope = pow(base, exp);
|
|
|
|
}
|
|
|
|
|
2023-06-25 14:40:30 +02:00
|
|
|
// parallel max
|
2023-12-07 21:27:19 +01:00
|
|
|
float lmax = -INFINITY;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
for (int i00 = tpitg; i00 < ne00; i00 += ntg) {
|
2024-05-11 09:32:41 +02:00
|
|
|
lmax = MAX(lmax, psrc0[i00]*scale + (pmask ? slope*pmask[i00] : 0.0f));
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
// find the max value in the block
|
|
|
|
float max_val = simd_max(lmax);
|
|
|
|
if (ntg > N_SIMDWIDTH) {
|
|
|
|
if (sgitg == 0) {
|
|
|
|
buf[tiisg] = -INFINITY;
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
if (tiisg == 0) {
|
|
|
|
buf[sgitg] = max_val;
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
max_val = buf[tiisg];
|
|
|
|
max_val = simd_max(max_val);
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
// parallel sum
|
2023-09-15 11:18:18 +02:00
|
|
|
float lsum = 0.0f;
|
2023-11-03 20:35:05 +01:00
|
|
|
for (int i00 = tpitg; i00 < ne00; i00 += ntg) {
|
2024-05-11 09:32:41 +02:00
|
|
|
const float exp_psrc0 = exp((psrc0[i00]*scale + (pmask ? slope*pmask[i00] : 0.0f)) - max_val);
|
2023-09-15 11:18:18 +02:00
|
|
|
lsum += exp_psrc0;
|
2023-09-05 19:57:27 +02:00
|
|
|
pdst[i00] = exp_psrc0;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
// This barrier fixes a failing test
|
|
|
|
// ref: https://github.com/ggerganov/ggml/pull/621#discussion_r1425156335
|
|
|
|
threadgroup_barrier(mem_flags::mem_none);
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
float sum = simd_sum(lsum);
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
if (ntg > N_SIMDWIDTH) {
|
|
|
|
if (sgitg == 0) {
|
|
|
|
buf[tiisg] = 0.0f;
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
if (tiisg == 0) {
|
|
|
|
buf[sgitg] = sum;
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
sum = buf[tiisg];
|
|
|
|
sum = simd_sum(sum);
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const float inv_sum = 1.0f/sum;
|
2023-09-15 11:18:18 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
for (int i00 = tpitg; i00 < ne00; i00 += ntg) {
|
2023-12-07 21:27:19 +01:00
|
|
|
pdst[i00] *= inv_sum;
|
2023-09-15 11:18:18 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
template<typename T>
|
2023-09-15 11:18:18 +02:00
|
|
|
kernel void kernel_soft_max_4(
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device char * dst,
|
2023-09-15 11:18:18 +02:00
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant float & scale,
|
2024-02-19 14:18:09 +01:00
|
|
|
constant float & max_bias,
|
|
|
|
constant float & m0,
|
|
|
|
constant float & m1,
|
|
|
|
constant uint32_t & n_head_log2,
|
|
|
|
threadgroup float * buf [[threadgroup(0)]],
|
2023-11-03 20:35:05 +01:00
|
|
|
uint tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = (tgpig) / (ne02*ne01);
|
|
|
|
const int64_t i02 = (tgpig - i03*ne02*ne01) / ne01;
|
|
|
|
const int64_t i01 = (tgpig - i03*ne02*ne01 - i02*ne01);
|
2023-09-15 11:18:18 +02:00
|
|
|
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
device const float4 * psrc4 = (device const float4 *) src0 + (i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00)/4;
|
|
|
|
device const T * pmask = src1 != src0 ? (device const T *) src1 + i01*ne00/4 : nullptr;
|
|
|
|
device float4 * pdst4 = (device float4 *) dst + (i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00)/4;
|
2023-09-15 11:18:18 +02:00
|
|
|
|
2024-05-11 09:32:41 +02:00
|
|
|
float slope = 1.0f;
|
2024-02-19 14:18:09 +01:00
|
|
|
|
|
|
|
if (max_bias > 0.0f) {
|
|
|
|
const int64_t h = i02;
|
|
|
|
|
|
|
|
const float base = h < n_head_log2 ? m0 : m1;
|
|
|
|
const int exp = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1;
|
|
|
|
|
|
|
|
slope = pow(base, exp);
|
|
|
|
}
|
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
// parallel max
|
2023-12-07 21:27:19 +01:00
|
|
|
float4 lmax4 = -INFINITY;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
for (int i00 = tpitg; i00 < ne00/4; i00 += ntg) {
|
2024-05-11 09:32:41 +02:00
|
|
|
lmax4 = fmax(lmax4, psrc4[i00]*scale + (float4)((pmask ? slope*pmask[i00] : 0.0f)));
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
const float lmax = MAX(MAX(lmax4[0], lmax4[1]), MAX(lmax4[2], lmax4[3]));
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
float max_val = simd_max(lmax);
|
|
|
|
if (ntg > N_SIMDWIDTH) {
|
|
|
|
if (sgitg == 0) {
|
|
|
|
buf[tiisg] = -INFINITY;
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
if (tiisg == 0) {
|
|
|
|
buf[sgitg] = max_val;
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
max_val = buf[tiisg];
|
|
|
|
max_val = simd_max(max_val);
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
// parallel sum
|
|
|
|
float4 lsum4 = 0.0f;
|
2023-11-03 20:35:05 +01:00
|
|
|
for (int i00 = tpitg; i00 < ne00/4; i00 += ntg) {
|
2024-05-11 09:32:41 +02:00
|
|
|
const float4 exp_psrc4 = exp((psrc4[i00]*scale + (float4)((pmask ? slope*pmask[i00] : 0.0f))) - max_val);
|
2023-09-15 11:18:18 +02:00
|
|
|
lsum4 += exp_psrc4;
|
|
|
|
pdst4[i00] = exp_psrc4;
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
const float lsum = lsum4[0] + lsum4[1] + lsum4[2] + lsum4[3];
|
2023-12-13 20:55:03 +01:00
|
|
|
|
|
|
|
// This barrier fixes a failing test
|
|
|
|
// ref: https://github.com/ggerganov/ggml/pull/621#discussion_r1425156335
|
|
|
|
threadgroup_barrier(mem_flags::mem_none);
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
float sum = simd_sum(lsum);
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
if (ntg > N_SIMDWIDTH) {
|
|
|
|
if (sgitg == 0) {
|
|
|
|
buf[tiisg] = 0.0f;
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
if (tiisg == 0) {
|
|
|
|
buf[sgitg] = sum;
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
sum = buf[tiisg];
|
|
|
|
sum = simd_sum(sum);
|
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const float inv_sum = 1.0f/sum;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
for (int i00 = tpitg; i00 < ne00/4; i00 += ntg) {
|
2023-12-07 21:27:19 +01:00
|
|
|
pdst4[i00] *= inv_sum;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
typedef decltype(kernel_soft_max<float>) kernel_soft_max_t;
|
|
|
|
typedef decltype(kernel_soft_max_4<float4>) kernel_soft_max_4_t;
|
|
|
|
|
|
|
|
template [[host_name("kernel_soft_max_f16")]] kernel kernel_soft_max_t kernel_soft_max<half>;
|
|
|
|
template [[host_name("kernel_soft_max_f32")]] kernel kernel_soft_max_t kernel_soft_max<float>;
|
|
|
|
template [[host_name("kernel_soft_max_f16_4")]] kernel kernel_soft_max_4_t kernel_soft_max_4<half4>;
|
|
|
|
template [[host_name("kernel_soft_max_f32_4")]] kernel kernel_soft_max_4_t kernel_soft_max_4<float4>;
|
|
|
|
|
2023-06-25 14:40:30 +02:00
|
|
|
kernel void kernel_diag_mask_inf(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int & n_past,
|
|
|
|
uint3 tpig[[thread_position_in_grid]]) {
|
|
|
|
const int64_t i02 = tpig[2];
|
|
|
|
const int64_t i01 = tpig[1];
|
|
|
|
const int64_t i00 = tpig[0];
|
|
|
|
|
|
|
|
if (i00 > n_past + i01) {
|
|
|
|
dst[i02*ne01*ne00 + i01*ne00 + i00] = -INFINITY;
|
|
|
|
} else {
|
|
|
|
dst[i02*ne01*ne00 + i01*ne00 + i00] = src0[i02*ne01*ne00 + i01*ne00 + i00];
|
2023-11-03 20:35:05 +01:00
|
|
|
}
|
2023-09-15 11:18:18 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_diag_mask_inf_8(
|
|
|
|
device const float4 * src0,
|
|
|
|
device float4 * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int & n_past,
|
|
|
|
uint3 tpig[[thread_position_in_grid]]) {
|
|
|
|
|
|
|
|
const int64_t i = 2*tpig[0];
|
|
|
|
|
|
|
|
dst[i+0] = src0[i+0];
|
|
|
|
dst[i+1] = src0[i+1];
|
|
|
|
int64_t i4 = 4*i;
|
|
|
|
const int64_t i02 = i4/(ne00*ne01); i4 -= i02*ne00*ne01;
|
|
|
|
const int64_t i01 = i4/(ne00); i4 -= i01*ne00;
|
|
|
|
const int64_t i00 = i4;
|
|
|
|
for (int k = 3; k >= 0; --k) {
|
|
|
|
if (i00 + 4 + k <= n_past + i01) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
dst[i+1][k] = -INFINITY;
|
|
|
|
if (i00 + k > n_past + i01) {
|
|
|
|
dst[i][k] = -INFINITY;
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_norm(
|
|
|
|
device const void * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant float & eps,
|
|
|
|
threadgroup float * sum [[threadgroup(0)]],
|
|
|
|
uint tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint ntg[[threads_per_threadgroup]]) {
|
|
|
|
device const float * x = (device const float *) ((device const char *) src0 + tgpig*nb01);
|
|
|
|
// MEAN
|
|
|
|
// parallel sum
|
|
|
|
sum[tpitg] = 0.0f;
|
|
|
|
for (int i00 = tpitg; i00 < ne00; i00 += ntg) {
|
|
|
|
sum[tpitg] += x[i00];
|
|
|
|
}
|
|
|
|
// reduce
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
for (uint i = ntg/2; i > 0; i /= 2) {
|
|
|
|
if (tpitg < i) {
|
|
|
|
sum[tpitg] += sum[tpitg + i];
|
|
|
|
}
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
}
|
2023-09-08 16:58:31 +02:00
|
|
|
const float mean = sum[0] / ne00;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
// recenter and VARIANCE
|
2023-09-08 16:58:31 +02:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-06-25 14:40:30 +02:00
|
|
|
device float * y = dst + tgpig*ne00;
|
|
|
|
sum[tpitg] = 0.0f;
|
|
|
|
for (int i00 = tpitg; i00 < ne00; i00 += ntg) {
|
2023-09-05 19:57:27 +02:00
|
|
|
y[i00] = x[i00] - mean;
|
2023-06-25 14:40:30 +02:00
|
|
|
sum[tpitg] += y[i00] * y[i00];
|
|
|
|
}
|
2023-09-05 19:57:27 +02:00
|
|
|
|
2023-06-25 14:40:30 +02:00
|
|
|
// reduce
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
for (uint i = ntg/2; i > 0; i /= 2) {
|
|
|
|
if (tpitg < i) {
|
|
|
|
sum[tpitg] += sum[tpitg + i];
|
|
|
|
}
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
}
|
2023-09-08 16:58:31 +02:00
|
|
|
const float variance = sum[0] / ne00;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
const float scale = 1.0f/sqrt(variance + eps);
|
|
|
|
for (int i00 = tpitg; i00 < ne00; i00 += ntg) {
|
|
|
|
y[i00] = y[i00] * scale;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_rms_norm(
|
|
|
|
device const void * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant float & eps,
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup float * buf [[threadgroup(0)]],
|
2023-06-25 14:40:30 +02:00
|
|
|
uint tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tpitg[[thread_position_in_threadgroup]],
|
2023-09-05 12:54:40 +02:00
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
2023-06-25 14:40:30 +02:00
|
|
|
uint ntg[[threads_per_threadgroup]]) {
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float4 * x = (device const float4 *) ((device const char *) src0 + tgpig*nb01);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
|
|
|
float4 sumf = 0;
|
|
|
|
float all_sum = 0;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
// parallel sum
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i00 = tpitg; i00 < ne00/4; i00 += ntg) {
|
|
|
|
sumf += x[i00] * x[i00];
|
|
|
|
}
|
|
|
|
all_sum = sumf[0] + sumf[1] + sumf[2] + sumf[3];
|
|
|
|
all_sum = simd_sum(all_sum);
|
2023-12-07 21:27:19 +01:00
|
|
|
if (ntg > N_SIMDWIDTH) {
|
|
|
|
if (sgitg == 0) {
|
|
|
|
buf[tiisg] = 0.0f;
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
if (tiisg == 0) {
|
|
|
|
buf[sgitg] = all_sum;
|
2023-11-03 20:35:05 +01:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
all_sum = buf[tiisg];
|
|
|
|
all_sum = simd_sum(all_sum);
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const float mean = all_sum/ne00;
|
2023-06-25 14:40:30 +02:00
|
|
|
const float scale = 1.0f/sqrt(mean + eps);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device float4 * y = (device float4 *) (dst + tgpig*ne00);
|
|
|
|
for (int i00 = tpitg; i00 < ne00/4; i00 += ntg) {
|
2023-06-25 14:40:30 +02:00
|
|
|
y[i00] = x[i00] * scale;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
kernel void kernel_group_norm(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int32_t & n_groups,
|
|
|
|
constant float & eps,
|
|
|
|
threadgroup float * buf [[threadgroup(0)]],
|
|
|
|
uint tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t ne = ne00*ne01*ne02;
|
|
|
|
const int64_t gs = ne00*ne01*((ne02 + n_groups - 1) / n_groups);
|
|
|
|
|
|
|
|
int start = tgpig * gs;
|
|
|
|
int end = start + gs;
|
|
|
|
|
|
|
|
start += tpitg;
|
|
|
|
|
|
|
|
if (end >= ne) {
|
|
|
|
end = ne;
|
|
|
|
}
|
|
|
|
|
|
|
|
float tmp = 0.0f; // partial sum for thread in warp
|
|
|
|
|
|
|
|
for (int j = start; j < end; j += ntg) {
|
|
|
|
tmp += src0[j];
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
tmp = simd_sum(tmp);
|
|
|
|
if (ntg > N_SIMDWIDTH) {
|
|
|
|
if (sgitg == 0) {
|
|
|
|
buf[tiisg] = 0.0f;
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
if (tiisg == 0) {
|
|
|
|
buf[sgitg] = tmp;
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
tmp = buf[tiisg];
|
|
|
|
tmp = simd_sum(tmp);
|
|
|
|
}
|
|
|
|
|
|
|
|
const float mean = tmp / gs;
|
|
|
|
tmp = 0.0f;
|
|
|
|
|
|
|
|
for (int j = start; j < end; j += ntg) {
|
|
|
|
float xi = src0[j] - mean;
|
|
|
|
dst[j] = xi;
|
|
|
|
tmp += xi * xi;
|
|
|
|
}
|
|
|
|
|
|
|
|
tmp = simd_sum(tmp);
|
|
|
|
if (ntg > N_SIMDWIDTH) {
|
|
|
|
if (sgitg == 0) {
|
|
|
|
buf[tiisg] = 0.0f;
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
if (tiisg == 0) {
|
|
|
|
buf[sgitg] = tmp;
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
tmp = buf[tiisg];
|
|
|
|
tmp = simd_sum(tmp);
|
|
|
|
}
|
|
|
|
|
|
|
|
const float variance = tmp / gs;
|
|
|
|
const float scale = 1.0f/sqrt(variance + eps);
|
|
|
|
for (int j = start; j < end; j += ntg) {
|
|
|
|
dst[j] *= scale;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
// function for calculate inner product between half a q4_0 block and 16 floats (yl), sumy is SUM(yl[i])
|
|
|
|
// il indicates where the q4 quants begin (0 or QK4_0/4)
|
|
|
|
// we assume that the yl's have been multiplied with the appropriate scale factor
|
|
|
|
// that corresponds to the missing bit shifts (1, 1/16, 1/256, 1/4096)
|
|
|
|
inline float block_q_n_dot_y(device const block_q4_0 * qb_curr, float sumy, thread float * yl, int il) {
|
|
|
|
float d = qb_curr->d;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float2 acc = 0.f;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint16_t * qs = ((device const uint16_t *)qb_curr + 1 + il/2);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = 0; i < 8; i+=2) {
|
|
|
|
acc[0] += yl[i + 0] * (qs[i / 2] & 0x000F)
|
|
|
|
+ yl[i + 1] * (qs[i / 2] & 0x0F00);
|
|
|
|
acc[1] += yl[i + 8] * (qs[i / 2] & 0x00F0)
|
|
|
|
+ yl[i + 9] * (qs[i / 2] & 0xF000);
|
|
|
|
}
|
|
|
|
return d * (sumy * -8.f + acc[0] + acc[1]);
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
// function for calculate inner product between half a q4_1 block and 16 floats (yl), sumy is SUM(yl[i])
|
|
|
|
// il indicates where the q4 quants begin (0 or QK4_0/4)
|
|
|
|
// we assume that the yl's have been multiplied with the appropriate scale factor
|
|
|
|
// that corresponds to the missing bit shifts (1, 1/16, 1/256, 1/4096)
|
|
|
|
inline float block_q_n_dot_y(device const block_q4_1 * qb_curr, float sumy, thread float * yl, int il) {
|
|
|
|
float d = qb_curr->d;
|
|
|
|
float m = qb_curr->m;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float2 acc = 0.f;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
|
|
|
device const uint16_t * qs = ((device const uint16_t *)qb_curr + 2 + il/2);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = 0; i < 8; i+=2) {
|
|
|
|
acc[0] += yl[i + 0] * (qs[i / 2] & 0x000F)
|
|
|
|
+ yl[i + 1] * (qs[i / 2] & 0x0F00);
|
|
|
|
acc[1] += yl[i + 8] * (qs[i / 2] & 0x00F0)
|
|
|
|
+ yl[i + 9] * (qs[i / 2] & 0xF000);
|
|
|
|
}
|
|
|
|
return d * (acc[0] + acc[1]) + sumy * m;
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
// function for calculate inner product between half a q5_0 block and 16 floats (yl), sumy is SUM(yl[i])
|
|
|
|
// il indicates where the q5 quants begin (0 or QK5_0/4)
|
|
|
|
// we assume that the yl's have been multiplied with the appropriate scale factor
|
|
|
|
// that corresponds to the missing bit shifts (1, 1/16, 1/256, 1/4096)
|
|
|
|
inline float block_q_n_dot_y(device const block_q5_0 * qb_curr, float sumy, thread float * yl, int il) {
|
|
|
|
float d = qb_curr->d;
|
|
|
|
|
|
|
|
float2 acc = 0.f;
|
|
|
|
|
|
|
|
device const uint16_t * qs = ((device const uint16_t *)qb_curr + 3 + il/2);
|
|
|
|
const uint32_t qh = *((device const uint32_t *)qb_curr->qh);
|
|
|
|
|
|
|
|
for (int i = 0; i < 8; i+=2) {
|
|
|
|
acc[0] += yl[i + 0] * ((qs[i / 2] & 0x000F) | ((qh >> (i+0+il ) << 4 ) & 0x00010))
|
|
|
|
+ yl[i + 1] * ((qs[i / 2] & 0x0F00) | ((qh >> (i+1+il ) << 12) & 0x01000));
|
|
|
|
acc[1] += yl[i + 8] * ((qs[i / 2] & 0x00F0) | ((qh >> (i+0+il+QK5_0/2) << 8 ) & 0x00100))
|
|
|
|
+ yl[i + 9] * ((qs[i / 2] & 0xF000) | ((qh >> (i+1+il+QK5_0/2) << 16) & 0x10000));
|
|
|
|
}
|
|
|
|
return d * (sumy * -16.f + acc[0] + acc[1]);
|
|
|
|
}
|
|
|
|
|
|
|
|
// function for calculate inner product between half a q5_1 block and 16 floats (yl), sumy is SUM(yl[i])
|
|
|
|
// il indicates where the q5 quants begin (0 or QK5_1/4)
|
|
|
|
// we assume that the yl's have been multiplied with the appropriate scale factor
|
|
|
|
// that corresponds to the missing bit shifts (1, 1/16, 1/256, 1/4096)
|
|
|
|
inline float block_q_n_dot_y(device const block_q5_1 * qb_curr, float sumy, thread float * yl, int il) {
|
|
|
|
float d = qb_curr->d;
|
|
|
|
float m = qb_curr->m;
|
|
|
|
|
|
|
|
float2 acc = 0.f;
|
|
|
|
|
|
|
|
device const uint16_t * qs = ((device const uint16_t *)qb_curr + 4 + il/2);
|
|
|
|
const uint32_t qh = *((device const uint32_t *)qb_curr->qh);
|
|
|
|
|
|
|
|
for (int i = 0; i < 8; i+=2) {
|
|
|
|
acc[0] += yl[i + 0] * ((qs[i / 2] & 0x000F) | ((qh >> (i+0+il ) << 4 ) & 0x00010))
|
|
|
|
+ yl[i + 1] * ((qs[i / 2] & 0x0F00) | ((qh >> (i+1+il ) << 12) & 0x01000));
|
|
|
|
acc[1] += yl[i + 8] * ((qs[i / 2] & 0x00F0) | ((qh >> (i+0+il+QK5_0/2) << 8 ) & 0x00100))
|
|
|
|
+ yl[i + 9] * ((qs[i / 2] & 0xF000) | ((qh >> (i+1+il+QK5_0/2) << 16) & 0x10000));
|
|
|
|
}
|
|
|
|
return d * (acc[0] + acc[1]) + sumy * m;
|
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
// putting them in the kernel cause a significant performance penalty
|
2023-11-03 20:35:05 +01:00
|
|
|
#define N_DST 4 // each SIMD group works on 4 rows
|
|
|
|
#define N_SIMDGROUP 2 // number of SIMD groups in a thread group
|
2023-09-05 12:54:40 +02:00
|
|
|
//Note: This is a template, but strictly speaking it only applies to
|
|
|
|
// quantizations where the block size is 32. It also does not
|
2024-01-02 20:07:47 +01:00
|
|
|
// guard against the number of rows not being divisible by
|
2023-09-05 12:54:40 +02:00
|
|
|
// N_DST, so this is another explicit assumption of the implementation.
|
|
|
|
template<typename block_q_type, int nr, int nsg, int nw>
|
2023-12-13 20:55:03 +01:00
|
|
|
void mul_vec_q_n_f32_impl(
|
2023-12-07 21:27:19 +01:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
2023-12-07 21:27:19 +01:00
|
|
|
uint3 tgpig, uint tiisg, uint sgitg) {
|
2023-09-05 12:54:40 +02:00
|
|
|
const int nb = ne00/QK4_0;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int first_row = (r0 * nsg + sgitg) * nr;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q_type * x = (device const block_q_type *) src0 + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
float yl[16]; // src1 vector cache
|
|
|
|
float sumf[nr] = {0.f};
|
|
|
|
|
|
|
|
const int ix = (tiisg/2);
|
|
|
|
const int il = (tiisg%2)*8;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
device const float * yb = y + ix * QK4_0 + il;
|
|
|
|
|
|
|
|
// each thread in a SIMD group deals with half a block.
|
|
|
|
for (int ib = ix; ib < nb; ib += nw/2) {
|
|
|
|
float sumy = 0;
|
|
|
|
for (int i = 0; i < 8; i += 2) {
|
|
|
|
sumy += yb[i] + yb[i+1];
|
|
|
|
yl[i+0] = yb[i+ 0];
|
|
|
|
yl[i+1] = yb[i+ 1]/256.f;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
sumy += yb[i+16] + yb[i+17];
|
|
|
|
yl[i+8] = yb[i+16]/16.f;
|
|
|
|
yl[i+9] = yb[i+17]/4096.f;
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < nr; row++) {
|
|
|
|
sumf[row] += block_q_n_dot_y(x+ib+row*nb, sumy, yl, il);
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
yb += QK4_0 * 16;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < nr; ++row) {
|
|
|
|
const float tot = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0 && first_row + row < ne01) {
|
2023-11-03 20:35:05 +01:00
|
|
|
dst[im*ne0*ne1 + r1*ne0 + first_row + row] = tot;
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
kernel void kernel_mul_mv_q4_0_f32(
|
2023-06-25 14:40:30 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-09-05 12:54:40 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
2023-11-03 20:35:05 +01:00
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2024-04-12 18:13:20 +02:00
|
|
|
mul_vec_q_n_f32_impl<block_q4_0, N_DST, N_SIMDGROUP, N_SIMDWIDTH>(src0,src1,dst,ne00,ne01,ne02,ne10,ne12,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg);
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
kernel void kernel_mul_mv_q4_1_f32(
|
2023-09-05 12:54:40 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-09-05 12:54:40 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2024-04-12 18:13:20 +02:00
|
|
|
mul_vec_q_n_f32_impl<block_q4_1, N_DST, N_SIMDGROUP, N_SIMDWIDTH>(src0,src1,dst,ne00,ne01,ne02,ne10,ne12,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg);
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
kernel void kernel_mul_mv_q5_0_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-11-03 20:35:05 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2024-04-12 18:13:20 +02:00
|
|
|
mul_vec_q_n_f32_impl<block_q5_0, N_DST, N_SIMDGROUP, N_SIMDWIDTH>(src0,src1,dst,ne00,ne01,ne02,ne10,ne12,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg);
|
2023-11-03 20:35:05 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_mul_mv_q5_1_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-11-03 20:35:05 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2024-04-12 18:13:20 +02:00
|
|
|
mul_vec_q_n_f32_impl<block_q5_1, N_DST, N_SIMDGROUP, N_SIMDWIDTH>(src0,src1,dst,ne00,ne01,ne02,ne10,ne12,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg);
|
2023-11-03 20:35:05 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
#define NB_Q8_0 8
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_q8_0_f32_impl(
|
2023-09-05 12:54:40 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2023-09-05 12:54:40 +02:00
|
|
|
const int nr = N_DST;
|
|
|
|
const int nsg = N_SIMDGROUP;
|
|
|
|
const int nw = N_SIMDWIDTH;
|
|
|
|
|
|
|
|
const int nb = ne00/QK8_0;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int first_row = (r0 * nsg + sgitg) * nr;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = first_row * nb + (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q8_0 * x = (device const block_q8_0 *) src0 + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
float yl[NB_Q8_0];
|
2023-09-05 12:54:40 +02:00
|
|
|
float sumf[nr]={0.f};
|
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
const int ix = tiisg/4;
|
|
|
|
const int il = tiisg%4;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
device const float * yb = y + ix * QK8_0 + NB_Q8_0*il;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
// each thread in a SIMD group deals with NB_Q8_0 quants at a time
|
|
|
|
for (int ib = ix; ib < nb; ib += nw/4) {
|
|
|
|
for (int i = 0; i < NB_Q8_0; ++i) {
|
2023-09-05 12:54:40 +02:00
|
|
|
yl[i] = yb[i];
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < nr; row++) {
|
2023-09-05 19:57:27 +02:00
|
|
|
device const int8_t * qs = x[ib+row*nb].qs + NB_Q8_0*il;
|
2023-09-05 12:54:40 +02:00
|
|
|
float sumq = 0.f;
|
2023-09-05 19:57:27 +02:00
|
|
|
for (int iq = 0; iq < NB_Q8_0; ++iq) {
|
2023-09-05 12:54:40 +02:00
|
|
|
sumq += qs[iq] * yl[iq];
|
|
|
|
}
|
|
|
|
sumf[row] += sumq*x[ib+row*nb].d;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
yb += NB_Q8_0 * nw;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < nr; ++row) {
|
|
|
|
const float tot = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0 && first_row + row < ne01) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = tot;
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_q8_0_f32")]]
|
|
|
|
kernel void kernel_mul_mv_q8_0_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant int64_t & ne10,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne11,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant int64_t & ne12,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2024-04-12 18:13:20 +02:00
|
|
|
kernel_mul_mv_q8_0_f32_impl(src0,src1,dst,ne00,ne01,ne02,ne10,ne12,ne0,ne1,r2,r3,nullptr,tgpig,tiisg,sgitg);
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
|
|
|
|
2023-09-15 12:56:08 +02:00
|
|
|
#define N_F32_F32 4
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_f32_f32_impl(
|
2023-09-15 12:56:08 +02:00
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
uint64_t nb00,
|
|
|
|
uint64_t nb01,
|
|
|
|
uint64_t nb02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne11,
|
|
|
|
int64_t ne12,
|
|
|
|
uint64_t nb10,
|
|
|
|
uint64_t nb11,
|
|
|
|
uint64_t nb12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg) {
|
2023-09-15 12:56:08 +02:00
|
|
|
|
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t rb = tgpig.y*N_F32_F32;
|
|
|
|
const int64_t im = tgpig.z;
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb02*ne02;
|
|
|
|
|
|
|
|
device const float * x = (device const float *) (src0 + offset0);
|
2023-09-15 12:56:08 +02:00
|
|
|
|
|
|
|
if (ne00 < 128) {
|
|
|
|
for (int row = 0; row < N_F32_F32; ++row) {
|
|
|
|
int r1 = rb + row;
|
|
|
|
if (r1 >= ne11) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
device const float * y = (device const float *) (src1 + r1*nb11 + im*nb12);
|
|
|
|
|
|
|
|
float sumf = 0;
|
|
|
|
for (int i = tiisg; i < ne00; i += 32) {
|
|
|
|
sumf += (float) x[i] * (float) y[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
device const float4 * x4 = (device const float4 *)x;
|
|
|
|
for (int row = 0; row < N_F32_F32; ++row) {
|
|
|
|
int r1 = rb + row;
|
|
|
|
if (r1 >= ne11) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
device const float * y = (device const float *) (src1 + r1*nb11 + im*nb12);
|
|
|
|
device const float4 * y4 = (device const float4 *) y;
|
|
|
|
|
|
|
|
float sumf = 0;
|
|
|
|
for (int i = tiisg; i < ne00/4; i += 32) {
|
|
|
|
for (int k = 0; k < 4; ++k) sumf += (float) x4[i][k] * y4[i][k];
|
|
|
|
}
|
|
|
|
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
for (int i = 4*(ne00/4); i < ne00; ++i) all_sum += (float) x[i] * y[i];
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_f32_f32")]]
|
|
|
|
kernel void kernel_mul_mv_f32_f32(
|
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]]) {
|
|
|
|
kernel_mul_mv_f32_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb00, nb01, nb02, ne10, ne11, ne12, nb10, nb11, nb12, ne0, ne1, r2, r3, tgpig, tiisg);
|
|
|
|
}
|
|
|
|
|
2023-11-12 14:31:08 +01:00
|
|
|
#define N_F16_F16 4
|
|
|
|
|
|
|
|
kernel void kernel_mul_mv_f16_f16(
|
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-11-12 14:31:08 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]]) {
|
|
|
|
|
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t rb = tgpig.y*N_F16_F16;
|
|
|
|
const int64_t im = tgpig.z;
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb02*ne02;
|
|
|
|
|
|
|
|
device const half * x = (device const half *) (src0 + offset0);
|
2023-11-12 14:31:08 +01:00
|
|
|
|
|
|
|
if (ne00 < 128) {
|
|
|
|
for (int row = 0; row < N_F16_F16; ++row) {
|
|
|
|
int r1 = rb + row;
|
|
|
|
if (r1 >= ne11) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
device const half * y = (device const half *) (src1 + r1*nb11 + im*nb12);
|
|
|
|
|
|
|
|
float sumf = 0;
|
|
|
|
for (int i = tiisg; i < ne00; i += 32) {
|
|
|
|
sumf += (half) x[i] * (half) y[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
device const half4 * x4 = (device const half4 *)x;
|
|
|
|
for (int row = 0; row < N_F16_F16; ++row) {
|
|
|
|
int r1 = rb + row;
|
|
|
|
if (r1 >= ne11) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
device const half * y = (device const half *) (src1 + r1*nb11 + im*nb12);
|
|
|
|
device const half4 * y4 = (device const half4 *) y;
|
|
|
|
|
|
|
|
float sumf = 0;
|
|
|
|
for (int i = tiisg; i < ne00/4; i += 32) {
|
|
|
|
for (int k = 0; k < 4; ++k) sumf += (half) x4[i][k] * y4[i][k];
|
|
|
|
}
|
|
|
|
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
for (int i = 4*(ne00/4); i < ne00; ++i) all_sum += (half) x[i] * y[i];
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_f16_f32_1row_impl(
|
2023-06-25 14:40:30 +02:00
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
2023-09-05 12:54:40 +02:00
|
|
|
constant int64_t & ne02,
|
2023-06-25 14:40:30 +02:00
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
2023-09-05 12:54:40 +02:00
|
|
|
constant int64_t & ne12,
|
2023-06-25 14:40:30 +02:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-06-25 14:40:30 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
2023-11-03 20:35:05 +01:00
|
|
|
uint tiisg[[thread_index_in_simdgroup]]) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t r1 = tgpig.y;
|
|
|
|
const int64_t im = tgpig.z;
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb02*ne02;
|
|
|
|
|
|
|
|
device const half * x = (device const half *) (src0 + offset0);
|
2023-06-25 14:40:30 +02:00
|
|
|
device const float * y = (device const float *) (src1 + r1*nb11 + im*nb12);
|
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
float sumf = 0;
|
|
|
|
if (ne00 < 128) {
|
|
|
|
for (int i = tiisg; i < ne00; i += 32) {
|
|
|
|
sumf += (float) x[i] * (float) y[i];
|
|
|
|
}
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
device const half4 * x4 = (device const half4 *) x;
|
|
|
|
device const float4 * y4 = (device const float4 *) y;
|
|
|
|
for (int i = tiisg; i < ne00/4; i += 32) {
|
|
|
|
for (int k = 0; k < 4; ++k) sumf += (float)x4[i][k] * y4[i][k];
|
|
|
|
}
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
for (int i = 4*(ne00/4); i < ne00; ++i) all_sum += (float) x[i] * y[i];
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_f16_f32_1row")]]
|
|
|
|
kernel void kernel_mul_mv_f16_f32_1row(
|
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]]) {
|
|
|
|
kernel_mul_mv_f16_f32_1row_impl(src0, src1, dst, ne00, ne01, ne02, nb00, nb01, nb02, ne10, ne11, ne12, nb10, nb11, nb12, ne0, ne1, r2, r3, tgpig, tiisg);
|
2023-09-05 19:57:27 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
#define N_F16_F32 4
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_f16_f32_impl(
|
2023-09-05 19:57:27 +02:00
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
uint64_t nb00,
|
|
|
|
uint64_t nb01,
|
|
|
|
uint64_t nb02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne11,
|
|
|
|
int64_t ne12,
|
|
|
|
uint64_t nb10,
|
|
|
|
uint64_t nb11,
|
|
|
|
uint64_t nb12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg) {
|
2023-09-05 19:57:27 +02:00
|
|
|
|
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t rb = tgpig.y*N_F16_F32;
|
|
|
|
const int64_t im = tgpig.z;
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb02*ne02;
|
|
|
|
|
|
|
|
device const half * x = (device const half *) (src0 + offset0);
|
2023-09-05 19:57:27 +02:00
|
|
|
|
|
|
|
if (ne00 < 128) {
|
|
|
|
for (int row = 0; row < N_F16_F32; ++row) {
|
|
|
|
int r1 = rb + row;
|
|
|
|
if (r1 >= ne11) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
device const float * y = (device const float *) (src1 + r1*nb11 + im*nb12);
|
|
|
|
|
|
|
|
float sumf = 0;
|
|
|
|
for (int i = tiisg; i < ne00; i += 32) {
|
|
|
|
sumf += (float) x[i] * (float) y[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-09-05 19:57:27 +02:00
|
|
|
} else {
|
|
|
|
device const half4 * x4 = (device const half4 *)x;
|
|
|
|
for (int row = 0; row < N_F16_F32; ++row) {
|
|
|
|
int r1 = rb + row;
|
|
|
|
if (r1 >= ne11) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
device const float * y = (device const float *) (src1 + r1*nb11 + im*nb12);
|
|
|
|
device const float4 * y4 = (device const float4 *) y;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 19:57:27 +02:00
|
|
|
float sumf = 0;
|
|
|
|
for (int i = tiisg; i < ne00/4; i += 32) {
|
|
|
|
for (int k = 0; k < 4; ++k) sumf += (float) x4[i][k] * y4[i][k];
|
|
|
|
}
|
|
|
|
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
for (int i = 4*(ne00/4); i < ne00; ++i) all_sum += (float) x[i] * y[i];
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_f16_f32")]]
|
|
|
|
kernel void kernel_mul_mv_f16_f32(
|
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]]) {
|
|
|
|
kernel_mul_mv_f16_f32_impl(src0, src1, dst, ne00, ne01, ne02, nb00, nb01, nb02, ne10, ne11, ne12, nb10, nb11, nb12, ne0, ne1, r2, r3, tgpig, tiisg);
|
|
|
|
}
|
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
// Assumes row size (ne00) is a multiple of 4
|
2023-11-03 20:35:05 +01:00
|
|
|
kernel void kernel_mul_mv_f16_f32_l4(
|
2023-09-15 11:18:18 +02:00
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-09-15 11:18:18 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]]) {
|
|
|
|
|
|
|
|
const int nrows = ne11;
|
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t im = tgpig.z;
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = r0*nb01 + (i12/r2)*nb02 + (i13/r3)*nb02*ne02;
|
|
|
|
|
|
|
|
device const half4 * x4 = (device const half4 *) (src0 + offset0);
|
2023-09-15 11:18:18 +02:00
|
|
|
|
|
|
|
for (int r1 = 0; r1 < nrows; ++r1) {
|
|
|
|
device const float4 * y4 = (device const float4 *) (src1 + r1*nb11 + im*nb12);
|
|
|
|
|
|
|
|
float sumf = 0;
|
|
|
|
for (int i = tiisg; i < ne00/4; i += 32) {
|
|
|
|
for (int k = 0; k < 4; ++k) sumf += (float) x4[i][k] * y4[i][k];
|
|
|
|
}
|
|
|
|
|
|
|
|
float all_sum = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[im*ne1*ne0 + r1*ne0 + r0] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
static float rope_yarn_ramp(const float low, const float high, const int i0) {
|
|
|
|
const float y = (i0 / 2 - low) / max(0.001f, high - low);
|
|
|
|
return 1.0f - min(1.0f, max(0.0f, y));
|
|
|
|
}
|
|
|
|
|
|
|
|
// YaRN algorithm based on LlamaYaRNScaledRotaryEmbedding.py from https://github.com/jquesnelle/yarn
|
|
|
|
// MIT licensed. Copyright (c) 2023 Jeffrey Quesnelle and Bowen Peng.
|
|
|
|
static void rope_yarn(
|
|
|
|
float theta_extrap, float freq_scale, float corr_dims[2], int64_t i0, float ext_factor, float mscale,
|
|
|
|
thread float * cos_theta, thread float * sin_theta
|
|
|
|
) {
|
|
|
|
// Get n-d rotational scaling corrected for extrapolation
|
|
|
|
float theta_interp = freq_scale * theta_extrap;
|
|
|
|
float theta = theta_interp;
|
|
|
|
if (ext_factor != 0.0f) {
|
|
|
|
float ramp_mix = rope_yarn_ramp(corr_dims[0], corr_dims[1], i0) * ext_factor;
|
|
|
|
theta = theta_interp * (1 - ramp_mix) + theta_extrap * ramp_mix;
|
|
|
|
|
|
|
|
// Get n-d magnitude scaling corrected for interpolation
|
|
|
|
mscale *= 1.0f + 0.1f * log(1.0f / freq_scale);
|
|
|
|
}
|
|
|
|
*cos_theta = cos(theta) * mscale;
|
|
|
|
*sin_theta = sin(theta) * mscale;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Apparently solving `n_rot = 2pi * x * base^((2 * max_pos_emb) / n_dims)` for x, we get
|
|
|
|
// `corr_fac(n_rot) = n_dims * log(max_pos_emb / (n_rot * 2pi)) / (2 * log(base))`
|
|
|
|
static float rope_yarn_corr_factor(int n_dims, int n_orig_ctx, float n_rot, float base) {
|
|
|
|
return n_dims * log(n_orig_ctx / (n_rot * 2 * M_PI_F)) / (2 * log(base));
|
|
|
|
}
|
|
|
|
|
|
|
|
static void rope_yarn_corr_dims(
|
|
|
|
int n_dims, int n_orig_ctx, float freq_base, float beta_fast, float beta_slow, float dims[2]
|
|
|
|
) {
|
|
|
|
// start and end correction dims
|
|
|
|
dims[0] = max(0.0f, floor(rope_yarn_corr_factor(n_dims, n_orig_ctx, beta_fast, freq_base)));
|
|
|
|
dims[1] = min(n_dims - 1.0f, ceil(rope_yarn_corr_factor(n_dims, n_orig_ctx, beta_slow, freq_base)));
|
|
|
|
}
|
|
|
|
|
|
|
|
typedef void (rope_t)(
|
|
|
|
device const void * src0,
|
|
|
|
device const int32_t * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
constant int & n_past,
|
|
|
|
constant int & n_dims,
|
|
|
|
constant int & mode,
|
|
|
|
constant int & n_orig_ctx,
|
|
|
|
constant float & freq_base,
|
|
|
|
constant float & freq_scale,
|
|
|
|
constant float & ext_factor,
|
|
|
|
constant float & attn_factor,
|
|
|
|
constant float & beta_fast,
|
|
|
|
constant float & beta_slow,
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint3 tptg[[threads_per_threadgroup]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]]);
|
|
|
|
|
|
|
|
template<typename T>
|
2023-06-25 14:40:30 +02:00
|
|
|
kernel void kernel_rope(
|
2023-11-03 20:35:05 +01:00
|
|
|
device const void * src0,
|
|
|
|
device const int32_t * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
constant int & n_past,
|
|
|
|
constant int & n_dims,
|
|
|
|
constant int & mode,
|
|
|
|
constant int & n_orig_ctx,
|
|
|
|
constant float & freq_base,
|
|
|
|
constant float & freq_scale,
|
|
|
|
constant float & ext_factor,
|
|
|
|
constant float & attn_factor,
|
|
|
|
constant float & beta_fast,
|
|
|
|
constant float & beta_slow,
|
2023-09-08 16:58:31 +02:00
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint3 tptg[[threads_per_threadgroup]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]]) {
|
|
|
|
const int64_t i3 = tgpig[2];
|
|
|
|
const int64_t i2 = tgpig[1];
|
|
|
|
const int64_t i1 = tgpig[0];
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
const bool is_neox = mode & 2;
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
float corr_dims[2];
|
|
|
|
rope_yarn_corr_dims(n_dims, n_orig_ctx, freq_base, beta_fast, beta_slow, corr_dims);
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
device const int32_t * pos = src1;
|
|
|
|
|
|
|
|
const int64_t p = pos[i2];
|
|
|
|
|
|
|
|
const float theta_0 = (float)p;
|
2023-09-08 16:58:31 +02:00
|
|
|
const float inv_ndims = -1.f/n_dims;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
if (!is_neox) {
|
2023-09-08 16:58:31 +02:00
|
|
|
for (int64_t i0 = 2*tiitg; i0 < ne0; i0 += 2*tptg.x) {
|
|
|
|
|
|
|
|
const float theta = theta_0 * pow(freq_base, inv_ndims*i0);
|
2023-11-03 20:35:05 +01:00
|
|
|
float cos_theta, sin_theta;
|
|
|
|
rope_yarn(theta, freq_scale, corr_dims, i0, ext_factor, attn_factor, &cos_theta, &sin_theta);
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
|
|
|
|
device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
const T x0 = src[0];
|
|
|
|
const T x1 = src[1];
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
dst_data[0] = x0*cos_theta - x1*sin_theta;
|
|
|
|
dst_data[1] = x0*sin_theta + x1*cos_theta;
|
|
|
|
}
|
|
|
|
} else {
|
2023-12-22 16:53:39 +01:00
|
|
|
for (int64_t ic = 2*tiitg; ic < ne0; ic += 2*tptg.x) {
|
|
|
|
if (ic < n_dims) {
|
|
|
|
const int64_t ib = 0;
|
2023-09-08 16:58:31 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
// simplified from `(ib * n_dims + ic) * inv_ndims`
|
|
|
|
const float cur_rot = inv_ndims*ic - ib;
|
|
|
|
|
|
|
|
const float theta = theta_0 * pow(freq_base, cur_rot);
|
|
|
|
float cos_theta, sin_theta;
|
|
|
|
rope_yarn(theta, freq_scale, corr_dims, cur_rot, ext_factor, attn_factor, &cos_theta, &sin_theta);
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
const int64_t i0 = ib*n_dims + ic/2;
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
|
|
|
|
device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
const float x0 = src[0];
|
|
|
|
const float x1 = src[n_dims/2];
|
|
|
|
|
|
|
|
dst_data[0] = x0*cos_theta - x1*sin_theta;
|
|
|
|
dst_data[n_dims/2] = x0*sin_theta + x1*cos_theta;
|
2023-12-22 16:53:39 +01:00
|
|
|
} else {
|
|
|
|
const int64_t i0 = ic;
|
|
|
|
|
|
|
|
device const T * const src = (device T *)((device char *) src0 + i3*nb03 + i2*nb02 + i1*nb01 + i0*nb00);
|
|
|
|
device T * dst_data = (device T *)((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
dst_data[0] = src[0];
|
|
|
|
dst_data[1] = src[1];
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
template [[host_name("kernel_rope_f32")]] kernel rope_t kernel_rope<float>;
|
|
|
|
template [[host_name("kernel_rope_f16")]] kernel rope_t kernel_rope<half>;
|
|
|
|
|
2024-01-31 14:35:41 +01:00
|
|
|
typedef void (im2col_t)(
|
2023-11-12 14:31:08 +01:00
|
|
|
device const float * x,
|
2024-01-31 14:35:41 +01:00
|
|
|
device char * dst,
|
|
|
|
constant int32_t & ofs0,
|
|
|
|
constant int32_t & ofs1,
|
|
|
|
constant int32_t & IW,
|
|
|
|
constant int32_t & IH,
|
|
|
|
constant int32_t & CHW,
|
|
|
|
constant int32_t & s0,
|
|
|
|
constant int32_t & s1,
|
|
|
|
constant int32_t & p0,
|
|
|
|
constant int32_t & p1,
|
|
|
|
constant int32_t & d0,
|
|
|
|
constant int32_t & d1,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tgpg[[threadgroups_per_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]);
|
|
|
|
|
|
|
|
template <typename T>
|
|
|
|
kernel void kernel_im2col(
|
|
|
|
device const float * x,
|
|
|
|
device char * dst,
|
2023-11-12 14:31:08 +01:00
|
|
|
constant int32_t & ofs0,
|
|
|
|
constant int32_t & ofs1,
|
|
|
|
constant int32_t & IW,
|
|
|
|
constant int32_t & IH,
|
|
|
|
constant int32_t & CHW,
|
|
|
|
constant int32_t & s0,
|
|
|
|
constant int32_t & s1,
|
|
|
|
constant int32_t & p0,
|
|
|
|
constant int32_t & p1,
|
|
|
|
constant int32_t & d0,
|
|
|
|
constant int32_t & d1,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tgpg[[threadgroups_per_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int32_t iiw = tgpig[2] * s0 + tpitg[2] * d0 - p0;
|
|
|
|
const int32_t iih = tgpig[1] * s1 + tpitg[1] * d1 - p1;
|
|
|
|
|
|
|
|
const int32_t offset_dst =
|
|
|
|
(tpitg[0] * tgpg[1] * tgpg[2] + tgpig[1] * tgpg[2] + tgpig[2]) * CHW +
|
|
|
|
(tgpig[0] * (ntg[1] * ntg[2]) + tpitg[1] * ntg[2] + tpitg[2]);
|
|
|
|
|
2024-01-31 14:35:41 +01:00
|
|
|
device T * pdst = (device T *) (dst);
|
|
|
|
|
2023-11-12 14:31:08 +01:00
|
|
|
if (iih < 0 || iih >= IH || iiw < 0 || iiw >= IW) {
|
2024-01-31 14:35:41 +01:00
|
|
|
pdst[offset_dst] = 0.0f;
|
2023-11-12 14:31:08 +01:00
|
|
|
} else {
|
|
|
|
const int32_t offset_src = tpitg[0] * ofs0 + tgpig[0] * ofs1;
|
2024-01-31 14:35:41 +01:00
|
|
|
pdst[offset_dst] = x[offset_src + iih * IW + iiw];
|
2023-11-12 14:31:08 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-01-31 14:35:41 +01:00
|
|
|
template [[host_name("kernel_im2col_f32")]] kernel im2col_t kernel_im2col<float>;
|
|
|
|
template [[host_name("kernel_im2col_f16")]] kernel im2col_t kernel_im2col<half>;
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
kernel void kernel_upscale_f32(
|
|
|
|
device const char * src0,
|
|
|
|
device char * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
constant int32_t & sf,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
|
|
|
|
const int64_t i3 = tgpig.z;
|
|
|
|
const int64_t i2 = tgpig.y;
|
|
|
|
const int64_t i1 = tgpig.x;
|
|
|
|
|
|
|
|
const int64_t i03 = i3;
|
|
|
|
const int64_t i02 = i2;
|
|
|
|
const int64_t i01 = i1/sf;
|
|
|
|
|
|
|
|
device const float * src0_ptr = (device const float *) (src0 + i03*nb03 + i02*nb02 + i01*nb01);
|
|
|
|
device float * dst_ptr = (device float *) (dst + i3*nb3 + i2*nb2 + i1*nb1);
|
|
|
|
|
|
|
|
for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) {
|
|
|
|
dst_ptr[i0] = src0_ptr[i0/sf];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_pad_f32(
|
|
|
|
device const char * src0,
|
|
|
|
device char * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
|
|
|
|
const int64_t i3 = tgpig.z;
|
|
|
|
const int64_t i2 = tgpig.y;
|
|
|
|
const int64_t i1 = tgpig.x;
|
|
|
|
|
|
|
|
const int64_t i03 = i3;
|
|
|
|
const int64_t i02 = i2;
|
|
|
|
const int64_t i01 = i1;
|
|
|
|
|
|
|
|
device const float * src0_ptr = (device const float *) (src0 + i03*nb03 + i02*nb02 + i01*nb01);
|
|
|
|
device float * dst_ptr = (device float *) (dst + i3*nb3 + i2*nb2 + i1*nb1);
|
|
|
|
|
|
|
|
if (i1 < ne01 && i2 < ne02 && i3 < ne03) {
|
|
|
|
for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) {
|
|
|
|
if (i0 < ne00) {
|
|
|
|
dst_ptr[i0] = src0_ptr[i0];
|
|
|
|
} else {
|
|
|
|
dst_ptr[i0] = 0.0f;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) {
|
|
|
|
dst_ptr[i0] = 0.0f;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-03-03 13:23:52 +01:00
|
|
|
kernel void kernel_arange_f32(
|
|
|
|
device char * dst,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant float & start,
|
|
|
|
constant float & step,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
|
|
|
|
device float * dst_ptr = (device float *) dst;
|
|
|
|
|
|
|
|
for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) {
|
|
|
|
dst_ptr[i0] = start + step * i0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_timestep_embedding_f32(
|
|
|
|
device const char * src0,
|
|
|
|
device char * dst,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant int & dim,
|
|
|
|
constant int & max_period,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
|
|
|
|
int i = tgpig.x;
|
|
|
|
device float * embed_data = (device float *)(dst + i*nb1);
|
|
|
|
|
|
|
|
int half_ = dim / 2;
|
|
|
|
for (int j = tpitg.x; j < half_; j += ntg.x) {
|
|
|
|
float timestep = ((device float *)src0)[i];
|
|
|
|
float freq = (float)exp(-log((float)max_period) * j / half_);
|
|
|
|
float arg = timestep * freq;
|
|
|
|
embed_data[j ] = cos(arg);
|
|
|
|
embed_data[j + half_] = sin(arg);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (dim % 2 != 0 && tpitg.x == 0) {
|
|
|
|
embed_data[dim] = 0.f;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
// bitonic sort implementation following the CUDA kernels as reference
|
|
|
|
typedef void (argsort_t)(
|
2024-04-03 15:07:05 +02:00
|
|
|
device const float * x,
|
|
|
|
device int32_t * dst,
|
|
|
|
constant int64_t & ncols,
|
|
|
|
constant int64_t & ncols_pad,
|
|
|
|
threadgroup int32_t * shared_values [[threadgroup(0)]],
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]]);
|
|
|
|
|
|
|
|
template<ggml_sort_order order>
|
|
|
|
kernel void kernel_argsort_f32_i32(
|
|
|
|
device const float * x,
|
|
|
|
device int32_t * dst,
|
|
|
|
constant int64_t & ncols,
|
2024-04-03 15:07:05 +02:00
|
|
|
constant int64_t & ncols_pad,
|
|
|
|
threadgroup int32_t * shared_values [[threadgroup(0)]],
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]]) {
|
|
|
|
// bitonic sort
|
|
|
|
int col = tpitg[0];
|
|
|
|
int row = tgpig[1];
|
|
|
|
|
2024-04-03 15:07:05 +02:00
|
|
|
if (col >= ncols_pad) return;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
2024-04-03 15:07:05 +02:00
|
|
|
device const float * x_row = x + row * ncols;
|
|
|
|
threadgroup int32_t * dst_row = shared_values;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
// initialize indices
|
2024-04-03 15:07:05 +02:00
|
|
|
dst_row[col] = col;
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
2024-04-03 15:07:05 +02:00
|
|
|
for (int k = 2; k <= ncols_pad; k *= 2) {
|
2023-12-07 21:27:19 +01:00
|
|
|
for (int j = k / 2; j > 0; j /= 2) {
|
|
|
|
int ixj = col ^ j;
|
|
|
|
if (ixj > col) {
|
|
|
|
if ((col & k) == 0) {
|
2024-04-03 15:07:05 +02:00
|
|
|
if (dst_row[col] >= ncols ||
|
|
|
|
(dst_row[ixj] < ncols && (order == GGML_SORT_ORDER_ASC ?
|
|
|
|
x_row[dst_row[col]] > x_row[dst_row[ixj]] :
|
|
|
|
x_row[dst_row[col]] < x_row[dst_row[ixj]]))
|
|
|
|
) {
|
2023-12-07 21:27:19 +01:00
|
|
|
SWAP(dst_row[col], dst_row[ixj]);
|
|
|
|
}
|
|
|
|
} else {
|
2024-04-03 15:07:05 +02:00
|
|
|
if (dst_row[ixj] >= ncols ||
|
|
|
|
(dst_row[col] < ncols && (order == GGML_SORT_ORDER_ASC ?
|
|
|
|
x_row[dst_row[col]] < x_row[dst_row[ixj]] :
|
|
|
|
x_row[dst_row[col]] > x_row[dst_row[ixj]]))
|
|
|
|
) {
|
2023-12-07 21:27:19 +01:00
|
|
|
SWAP(dst_row[col], dst_row[ixj]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
}
|
|
|
|
}
|
2024-04-03 15:07:05 +02:00
|
|
|
|
|
|
|
// copy the result to dst without the padding
|
|
|
|
if (col < ncols) {
|
|
|
|
dst[row * ncols + col] = dst_row[col];
|
|
|
|
}
|
2023-12-07 21:27:19 +01:00
|
|
|
}
|
|
|
|
|
2024-04-03 15:07:05 +02:00
|
|
|
template [[host_name("kernel_argsort_f32_i32_asc")]] kernel argsort_t kernel_argsort_f32_i32<GGML_SORT_ORDER_ASC>;
|
|
|
|
template [[host_name("kernel_argsort_f32_i32_desc")]] kernel argsort_t kernel_argsort_f32_i32<GGML_SORT_ORDER_DESC>;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
kernel void kernel_leaky_relu_f32(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant float & slope,
|
|
|
|
uint tpig[[thread_position_in_grid]]) {
|
|
|
|
dst[tpig] = src0[tpig] > 0.0f ? src0[tpig] : src0[tpig] * slope;
|
|
|
|
}
|
|
|
|
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
typedef void (flash_attn_ext_f16_t)(
|
|
|
|
device const char * q,
|
|
|
|
device const char * k,
|
|
|
|
device const char * v,
|
|
|
|
device const char * mask,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant uint64_t & nb13,
|
|
|
|
constant uint64_t & nb31,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant float & scale,
|
2024-05-11 09:32:41 +02:00
|
|
|
constant float & max_bias,
|
|
|
|
constant float & m0,
|
|
|
|
constant float & m1,
|
|
|
|
constant uint32_t & n_head_log2,
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
threadgroup half * shared,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]],
|
|
|
|
ushort tiisg[[thread_index_in_simdgroup]],
|
|
|
|
ushort sgitg[[simdgroup_index_in_threadgroup]]);
|
|
|
|
|
|
|
|
// ref: https://arxiv.org/pdf/2307.08691.pdf
|
|
|
|
template<int64_t D, int64_t Q = 8, int64_t C = 32> // head size, queries per threadgroup, cache items per threadgroup
|
|
|
|
kernel void kernel_flash_attn_ext_f16(
|
|
|
|
device const char * q,
|
|
|
|
device const char * k,
|
|
|
|
device const char * v,
|
|
|
|
device const char * mask,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant uint64_t & nb13,
|
|
|
|
constant uint64_t & nb31,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant float & scale,
|
2024-05-11 09:32:41 +02:00
|
|
|
constant float & max_bias,
|
|
|
|
constant float & m0,
|
|
|
|
constant float & m1,
|
|
|
|
constant uint32_t & n_head_log2,
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
threadgroup half * shared [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]],
|
|
|
|
ushort tiisg[[thread_index_in_simdgroup]],
|
|
|
|
ushort sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
const short nsg = ntg.y; // number of simdgroups
|
|
|
|
|
|
|
|
const short iq3 = tgpig[2];
|
|
|
|
const short iq2 = tgpig[1];
|
|
|
|
const short iq1 = tgpig[0]*Q;
|
|
|
|
|
|
|
|
const short D4 = D/4;
|
|
|
|
const short D8 = D/8;
|
2024-05-08 08:14:50 +02:00
|
|
|
//const short Q8 = Q/8;
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
const short NW = N_SIMDWIDTH;
|
|
|
|
const short SH = (C + Q); // shared memory per simdgroup in (half)
|
|
|
|
|
|
|
|
const short T = D + 2*nsg*SH; // shared memory size per query in (half)
|
|
|
|
const short TF = T/2; // shared memory size per query in (float)
|
|
|
|
const short T4 = T/4; // shared memory size per query in (half4)
|
|
|
|
|
|
|
|
threadgroup half * sq = (threadgroup half *) (shared + 0*D); // holds the query data
|
|
|
|
threadgroup half4 * sq4 = (threadgroup half4 *) (shared + 0*D); // same as above but in half4
|
|
|
|
threadgroup float * ss = (threadgroup float *) (shared + 2*sgitg*SH + 1*D); // scratch buffer for attention and diagonal matrix
|
|
|
|
|
|
|
|
// store the result for all queries in local memory in 8x8 matrices (the O matrix from the paper)
|
|
|
|
simdgroup_half8x8 lo[D8];
|
|
|
|
|
|
|
|
// load heads from Q to shared memory
|
|
|
|
for (short j = sgitg; j < Q; j += nsg) {
|
|
|
|
device const float4 * q4 = (device const float4 *) ((device const char *) q + ((iq1 + j)*nb01 + iq2*nb02 + iq3*nb03));
|
|
|
|
|
|
|
|
for (short i = tiisg; i < D4; i += NW) {
|
|
|
|
if (iq1 + j < ne01) {
|
|
|
|
sq4[j*T4 + i] = (half4) q4[i];
|
|
|
|
} else {
|
|
|
|
sq4[j*T4 + i] = 0.0h;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// zero out lo
|
|
|
|
for (short i = 0; i < D8; ++i) {
|
|
|
|
lo[i] = make_filled_simdgroup_matrix<half, 8>(0.0h);
|
|
|
|
}
|
|
|
|
|
|
|
|
// zero out shared memory SH
|
|
|
|
for (short j = 0; j < Q; ++j) {
|
|
|
|
for (short i = tiisg; i < SH; i += NW) {
|
|
|
|
ss[j*TF + i] = 0.0f;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
{
|
|
|
|
float S[Q] = { [0 ... Q-1] = 0.0h };
|
|
|
|
float M[Q] = { [0 ... Q-1] = -FLT_MAX/2 };
|
|
|
|
|
|
|
|
// assume K and V are same shape
|
|
|
|
const short ne22 = ne12;
|
|
|
|
const short ne23 = ne13;
|
|
|
|
|
|
|
|
const uint nb21 = nb11;
|
|
|
|
const uint nb22 = nb12;
|
|
|
|
const uint nb23 = nb13;
|
|
|
|
|
|
|
|
// broadcast
|
|
|
|
const short rk2 = ne02/ne12;
|
|
|
|
const short rk3 = ne03/ne13;
|
|
|
|
|
|
|
|
const short rv2 = ne02/ne22;
|
|
|
|
const short rv3 = ne03/ne23;
|
|
|
|
|
|
|
|
// k indices
|
|
|
|
const short ik2 = iq2/rk2;
|
|
|
|
const short ik3 = iq3/rk3;
|
|
|
|
|
|
|
|
// v indices
|
|
|
|
const short iv2 = iq2/rv2;
|
|
|
|
const short iv3 = iq3/rv3;
|
|
|
|
|
|
|
|
// load the queries from shared memory into local memory
|
|
|
|
simdgroup_half8x8 mq[D8];
|
|
|
|
|
|
|
|
for (short i = 0; i < D8; ++i) {
|
|
|
|
simdgroup_load(mq[i], sq + i*8, T);
|
|
|
|
}
|
|
|
|
|
|
|
|
// pointer to the mask
|
|
|
|
device const half * mp = (device const half *) (mask + iq1*nb31);
|
|
|
|
|
|
|
|
// prepare diagonal scale matrix
|
|
|
|
simdgroup_float8x8 mscale(scale);
|
|
|
|
|
2024-05-11 09:32:41 +02:00
|
|
|
// prepare diagonal slope matrix
|
|
|
|
simdgroup_float8x8 mslope(1.0f);
|
|
|
|
|
|
|
|
// ALiBi
|
|
|
|
if (max_bias > 0.0f) {
|
2024-05-11 15:25:50 +02:00
|
|
|
const uint32_t h = iq2;
|
2024-05-11 09:32:41 +02:00
|
|
|
|
|
|
|
const float base = h < n_head_log2 ? m0 : m1;
|
|
|
|
const int exph = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1;
|
|
|
|
|
|
|
|
mslope = simdgroup_float8x8(pow(base, exph));
|
|
|
|
}
|
|
|
|
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
// loop over the KV cache
|
|
|
|
// each simdgroup handles blocks of Q rows and C columns
|
|
|
|
for (int ic0 = 0; ic0 < ne11; ic0 += C*nsg) {
|
|
|
|
const int ic = ic0 + C*sgitg;
|
|
|
|
if (ic >= ne11) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Q*K^T
|
|
|
|
{
|
|
|
|
for (short cc = 0; cc < C/8; ++cc) {
|
|
|
|
simdgroup_float8x8 mqk = make_filled_simdgroup_matrix<float, 8>(0.h);
|
|
|
|
|
|
|
|
device const half * pk = (device const half *) ((device const char *) k + ((ic + 8*cc)*nb11 + ik2*nb12 + ik3*nb13));
|
|
|
|
|
|
|
|
for (short i = 0; i < D8; ++i) {
|
|
|
|
simdgroup_half8x8 mk;
|
|
|
|
simdgroup_load(mk, pk + i*8, nb11/sizeof(half), 0, true); // transpose
|
|
|
|
|
|
|
|
simdgroup_multiply_accumulate(mqk, mq[i], mk, mqk);
|
|
|
|
}
|
|
|
|
|
2024-05-11 09:32:41 +02:00
|
|
|
// mqk = mqk*scale + mask*slope
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
simdgroup_half8x8 mm;
|
|
|
|
simdgroup_load(mm, mp + ic + 8*cc, nb31/sizeof(half), 0, false);
|
2024-05-11 09:32:41 +02:00
|
|
|
simdgroup_multiply(mm, mslope, mm);
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
simdgroup_multiply_accumulate(mqk, mqk, mscale, mm);
|
|
|
|
|
|
|
|
simdgroup_store(mqk, ss + 8*cc, TF, 0, false);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// used to detect blocks full of -INF
|
|
|
|
float smax = -INFINITY;
|
|
|
|
|
|
|
|
// online softmax
|
|
|
|
{
|
|
|
|
float ms[Q];
|
|
|
|
|
|
|
|
for (short j = 0; j < Q; ++j) {
|
|
|
|
const short p = tiisg;
|
|
|
|
|
|
|
|
const float m = M[j];
|
|
|
|
const float s = ss[j*TF + p];
|
|
|
|
|
|
|
|
smax = simd_max(max(smax, s));
|
|
|
|
M[j] = simd_max(max(M[j], s));
|
|
|
|
|
|
|
|
ms[j] = exp(m - M[j]);
|
|
|
|
const float vs = exp(s - M[j]);
|
|
|
|
|
|
|
|
S[j] = S[j]*ms[j] + simd_sum(vs);
|
|
|
|
|
|
|
|
// the P matrix from the paper (Q rows, C columns)
|
|
|
|
ss[j*TF + p] = vs;
|
|
|
|
}
|
|
|
|
|
|
|
|
// create a QxQ diagonal matrix for rescaling the output
|
|
|
|
if (tiisg < Q) {
|
|
|
|
ss[tiisg*TF + C + tiisg] = ms[tiisg];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// skip -INF blocks
|
|
|
|
if (smax == -INFINITY) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
// O = diag(ms)*O
|
|
|
|
{
|
|
|
|
simdgroup_float8x8 mm;
|
|
|
|
simdgroup_load(mm, ss + C, TF, 0, false);
|
|
|
|
|
|
|
|
for (short i = 0; i < D8; ++i) {
|
|
|
|
simdgroup_multiply(lo[i], mm, lo[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// O = O + (Q*K^T)*V
|
|
|
|
{
|
|
|
|
for (short cc = 0; cc < C/8; ++cc) {
|
|
|
|
device const half * pv = (device const half *) ((device const char *) v + ((ic + 8*cc)*nb21 + iv2*nb22 + iv3*nb23));
|
|
|
|
|
|
|
|
for (short i = 0; i < D8; ++i) {
|
|
|
|
simdgroup_half8x8 mk;
|
|
|
|
simdgroup_load(mk, pv + i*8, nb21/sizeof(half), 0, false);
|
|
|
|
|
|
|
|
simdgroup_float8x8 mv;
|
|
|
|
simdgroup_load(mv, ss + 8*cc, TF, 0, false);
|
|
|
|
|
|
|
|
simdgroup_multiply_accumulate(lo[i], mv, mk, lo[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// these are needed for reducing the results from the simdgroups (reuse the ss buffer)
|
|
|
|
for (short j = 0; j < Q; ++j) {
|
|
|
|
if (tiisg == 0) {
|
|
|
|
ss[j*TF + 0] = S[j];
|
|
|
|
ss[j*TF + 1] = M[j];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// reduce the warps sequentially
|
|
|
|
for (short sg = 1; sg < nsg; ++sg) {
|
|
|
|
float S = { 0.0h };
|
|
|
|
float M = { -FLT_MAX/2 };
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
// each simdgroup stores its output to shared memory, reusing sq
|
|
|
|
if (sgitg == sg) {
|
|
|
|
for (short i = 0; i < D8; ++i) {
|
|
|
|
simdgroup_store(lo[i], sq + i*8, T, 0, false);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
// the first simdgroup accumulates the results from the other simdgroups
|
|
|
|
if (sgitg == 0) {
|
|
|
|
for (short j = 0; j < Q; ++j) {
|
|
|
|
const float S0 = ss[j*TF + 0];
|
|
|
|
const float S1 = ss[j*TF + sg*SH + 0];
|
|
|
|
|
|
|
|
const float M0 = ss[j*TF + 1];
|
|
|
|
const float M1 = ss[j*TF + sg*SH + 1];
|
|
|
|
|
|
|
|
M = max(M0, M1);
|
|
|
|
|
|
|
|
const float ms0 = exp(M0 - M);
|
|
|
|
const float ms1 = exp(M1 - M);
|
|
|
|
|
|
|
|
S = S0*ms0 + S1*ms1;
|
|
|
|
|
|
|
|
if (tiisg == 0) {
|
|
|
|
ss[j*TF + 0] = S;
|
|
|
|
ss[j*TF + 1] = M;
|
|
|
|
|
|
|
|
ss[j*TF + C + j ] = ms0;
|
|
|
|
ss[j*TF + C + j + sg*SH] = ms1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// O_0 = diag(ms0)*O_0 + diag(ms1)*O_1
|
|
|
|
{
|
|
|
|
simdgroup_half8x8 t;
|
|
|
|
simdgroup_float8x8 ms0;
|
|
|
|
simdgroup_float8x8 ms1;
|
|
|
|
|
|
|
|
simdgroup_load(ms0, ss + C, TF, 0, false);
|
|
|
|
simdgroup_load(ms1, ss + C + sg*SH, TF, 0, false);
|
|
|
|
|
|
|
|
for (short i = 0; i < D8; ++i) {
|
|
|
|
simdgroup_load (t, sq + i*8, T, 0, false);
|
|
|
|
simdgroup_multiply(t, ms1, t);
|
|
|
|
|
|
|
|
simdgroup_multiply_accumulate(lo[i], ms0, lo[i], t);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// store result to shared memory (reuse sq)
|
|
|
|
if (sgitg == 0) {
|
|
|
|
for (short i = 0; i < D8; ++i) {
|
|
|
|
simdgroup_store(lo[i], sq + i*8, T, 0, false);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
device float4 * dst4 = (device float4 *) dst;
|
|
|
|
|
|
|
|
// final rescale with 1/S and store to global memory
|
|
|
|
if (sgitg == 0) {
|
|
|
|
for (short j = 0; j < Q && iq1 + j < ne01; ++j) {
|
|
|
|
const float S = ss[j*TF + 0];
|
|
|
|
|
|
|
|
for (short i = tiisg; i < D4; i += NW) {
|
|
|
|
dst4[(iq3*ne2*ne1 + iq2 + (iq1 + j)*ne1)*D4 + i] = (float4) sq4[j*T4 + i]/S;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
template [[host_name("kernel_flash_attn_ext_f16_h64" )]] kernel flash_attn_ext_f16_t kernel_flash_attn_ext_f16<64>;
|
|
|
|
template [[host_name("kernel_flash_attn_ext_f16_h80" )]] kernel flash_attn_ext_f16_t kernel_flash_attn_ext_f16<80>;
|
|
|
|
template [[host_name("kernel_flash_attn_ext_f16_h96" )]] kernel flash_attn_ext_f16_t kernel_flash_attn_ext_f16<96>;
|
|
|
|
template [[host_name("kernel_flash_attn_ext_f16_h112")]] kernel flash_attn_ext_f16_t kernel_flash_attn_ext_f16<112>;
|
|
|
|
template [[host_name("kernel_flash_attn_ext_f16_h128")]] kernel flash_attn_ext_f16_t kernel_flash_attn_ext_f16<128>;
|
|
|
|
template [[host_name("kernel_flash_attn_ext_f16_h256")]] kernel flash_attn_ext_f16_t kernel_flash_attn_ext_f16<256>;
|
|
|
|
|
|
|
|
template<int64_t D, int64_t Q = 1, int64_t C = 32> // head size, queries per threadgroup, cache items per threadgroup
|
|
|
|
kernel void kernel_flash_attn_ext_vec_f16(
|
|
|
|
device const char * q,
|
|
|
|
device const char * k,
|
|
|
|
device const char * v,
|
|
|
|
device const char * mask,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant uint64_t & nb13,
|
|
|
|
constant uint64_t & nb31,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant float & scale,
|
2024-05-11 09:32:41 +02:00
|
|
|
constant float & max_bias,
|
|
|
|
constant float & m0,
|
|
|
|
constant float & m1,
|
|
|
|
constant uint32_t & n_head_log2,
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
threadgroup half * shared [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]],
|
|
|
|
ushort tiisg[[thread_index_in_simdgroup]],
|
|
|
|
ushort sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
const short nsg = ntg.y; // number of simdgroups
|
|
|
|
|
|
|
|
const short iq3 = tgpig[2];
|
|
|
|
const short iq2 = tgpig[1];
|
|
|
|
const short iq1 = tgpig[0];
|
|
|
|
|
|
|
|
const short D4 = D/4;
|
|
|
|
const short NW = N_SIMDWIDTH;
|
|
|
|
const short SH = (C + Q); // shared memory per simdgroup in (half)
|
|
|
|
|
|
|
|
const short T = D + 2*nsg*SH; // shared memory size per query in (half)
|
|
|
|
|
2024-05-11 09:32:41 +02:00
|
|
|
float slope = 1.0f;
|
|
|
|
|
|
|
|
// ALiBi
|
|
|
|
if (max_bias > 0.0f) {
|
2024-05-11 15:25:50 +02:00
|
|
|
const uint32_t h = iq2;
|
2024-05-11 09:32:41 +02:00
|
|
|
|
|
|
|
const float base = h < n_head_log2 ? m0 : m1;
|
|
|
|
const int exp = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1;
|
|
|
|
|
|
|
|
slope = pow(base, exp);
|
|
|
|
}
|
|
|
|
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
//threadgroup half * sq = (threadgroup half *) (shared + 0*D); // holds the query data
|
|
|
|
threadgroup half4 * sq4 = (threadgroup half4 *) (shared + 0*D); // same as above but in half4
|
|
|
|
threadgroup float * ss = (threadgroup float *) (shared + 2*sgitg*SH + 1*D); // scratch buffer for attention and diagonal matrix
|
|
|
|
threadgroup float4 * ss4 = (threadgroup float4 *) (shared + 2*sgitg*SH + 1*D); // same as above but in half4
|
|
|
|
threadgroup half4 * sr4 = (threadgroup half4 *) (shared + sgitg*D + 1*T); // scratch buffer for the results
|
|
|
|
|
|
|
|
// store the result for all queries in local memory in 8x8 matrices (the O matrix from the paper)
|
|
|
|
half4 lo[D4/NW];
|
|
|
|
|
|
|
|
// load heads from Q to shared memory
|
|
|
|
device const float4 * q4 = (device const float4 *) ((device const char *) q + (iq1*nb01 + iq2*nb02 + iq3*nb03));
|
|
|
|
|
|
|
|
for (short i = tiisg; i < D4; i += NW) {
|
|
|
|
if (iq1 < ne01) {
|
|
|
|
sq4[i] = (half4) q4[i];
|
|
|
|
} else {
|
|
|
|
sq4[i] = 0.0h;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// zero out lo
|
|
|
|
for (short i = tiisg; i < D4; i += NW) {
|
|
|
|
lo[i/NW] = 0.0h;
|
|
|
|
}
|
|
|
|
|
|
|
|
// zero out shared memory SH
|
|
|
|
for (short i = tiisg; i < SH/4; i += NW) {
|
|
|
|
ss4[i] = 0.0h;
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
{
|
|
|
|
float S = { 0.0h };
|
|
|
|
float M = { -FLT_MAX/2 };
|
|
|
|
|
|
|
|
// assume K and V are same shape
|
|
|
|
const short ne22 = ne12;
|
|
|
|
const short ne23 = ne13;
|
|
|
|
|
|
|
|
const uint nb21 = nb11;
|
|
|
|
const uint nb22 = nb12;
|
|
|
|
const uint nb23 = nb13;
|
|
|
|
|
|
|
|
// broadcast
|
|
|
|
const short rk2 = ne02/ne12;
|
|
|
|
const short rk3 = ne03/ne13;
|
|
|
|
|
|
|
|
const short rv2 = ne02/ne22;
|
|
|
|
const short rv3 = ne03/ne23;
|
|
|
|
|
|
|
|
// k indices
|
|
|
|
const short ik2 = iq2 / rk2;
|
|
|
|
const short ik3 = iq3 / rk3;
|
|
|
|
|
|
|
|
// v indices
|
|
|
|
const short iv2 = iq2 / rv2;
|
|
|
|
const short iv3 = iq3 / rv3;
|
|
|
|
|
|
|
|
// load the queries from shared memory into local memory
|
|
|
|
half4 mq[D4];
|
|
|
|
|
|
|
|
for (short ii = 0; ii < D4; ii += NW) {
|
|
|
|
short i = ii + tiisg;
|
|
|
|
mq[i] = sq4[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
// pointer to the mask
|
|
|
|
device const half4 * mp4 = (device const half4 *) (mask + iq1*nb31);
|
|
|
|
|
|
|
|
// loop over the KV cache
|
|
|
|
// each simdgroup handles blocks of Q rows and C columns
|
|
|
|
for (int ic0 = 0; ic0 < ne11; ic0 += C*nsg) {
|
|
|
|
const int ic = ic0 + C*sgitg;
|
|
|
|
if (ic >= ne11) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Q*K^T
|
|
|
|
{
|
|
|
|
#pragma unroll
|
|
|
|
for (short cc = 0; cc < C/4; ++cc) {
|
|
|
|
float4 mqk = { 0.0h };
|
|
|
|
|
|
|
|
device const half4 * pk4 = (device const half4 *) ((device const char *) k + ((ic + 4*cc)*nb11 + ik2*nb12 + ik3*nb13));
|
|
|
|
|
|
|
|
#pragma unroll
|
|
|
|
for (short ii = 0; ii < D4; ii += NW) {
|
|
|
|
const short i = ii + tiisg;
|
|
|
|
|
|
|
|
half4x4 mk;
|
|
|
|
mk[0] = pk4[i + 0*(nb11/8)];
|
|
|
|
mk[1] = pk4[i + 1*(nb11/8)];
|
|
|
|
mk[2] = pk4[i + 2*(nb11/8)];
|
|
|
|
mk[3] = pk4[i + 3*(nb11/8)];
|
|
|
|
|
|
|
|
mqk += (float4) (mq[i] * mk);
|
|
|
|
}
|
|
|
|
|
|
|
|
// reduce the results from the threads in the simdgroup
|
|
|
|
mqk += simd_shuffle_down(mqk, 16);
|
|
|
|
mqk += simd_shuffle_down(mqk, 8);
|
|
|
|
mqk += simd_shuffle_down(mqk, 4);
|
|
|
|
mqk += simd_shuffle_down(mqk, 2);
|
|
|
|
mqk += simd_shuffle_down(mqk, 1);
|
|
|
|
|
2024-05-11 09:32:41 +02:00
|
|
|
// mqk = mqk*scale + mask*slope
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
if (tiisg == 0) {
|
|
|
|
float4 mm = (float4) mp4[ic/4 + cc];
|
2024-05-11 09:32:41 +02:00
|
|
|
mqk = mqk*scale + mm*slope;
|
ggml : add Flash Attention (llama/5021)
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (llama/6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (llama/6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
2024-04-30 11:16:08 +02:00
|
|
|
|
|
|
|
ss4[cc] = mqk;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// online softmax
|
|
|
|
{
|
|
|
|
const short p = tiisg;
|
|
|
|
|
|
|
|
const float m = M;
|
|
|
|
const float s = ss[p];
|
|
|
|
|
|
|
|
M = simd_max(max(M, s));
|
|
|
|
|
|
|
|
const float ms = exp(m - M);
|
|
|
|
const float vs = exp(s - M);
|
|
|
|
|
|
|
|
S = S*ms + simd_sum(vs);
|
|
|
|
|
|
|
|
// the P matrix from the paper (Q rows, C columns)
|
|
|
|
ss[p] = vs;
|
|
|
|
|
|
|
|
// O = diag(ms)*O
|
|
|
|
#pragma unroll
|
|
|
|
for (short ii = 0; ii < D4; ii += NW) {
|
|
|
|
const short i = ii + tiisg;
|
|
|
|
lo[i/NW] *= ms;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// O = O + (Q*K^T)*V
|
|
|
|
{
|
|
|
|
#pragma unroll
|
|
|
|
for (short cc = 0; cc < C/4; ++cc) {
|
|
|
|
device const half4 * pv4 = (device const half4 *) ((device const char *) v + ((ic + 4*cc)*nb21 + iv2*nb22 + iv3*nb23));
|
|
|
|
|
|
|
|
#pragma unroll
|
|
|
|
for (short ii = 0; ii < D4; ii += NW) {
|
|
|
|
const short i = ii + tiisg;
|
|
|
|
|
|
|
|
lo[i/NW] += pv4[i + 0*(nb21/8)] * ss[4*cc + 0];
|
|
|
|
lo[i/NW] += pv4[i + 1*(nb21/8)] * ss[4*cc + 1];
|
|
|
|
lo[i/NW] += pv4[i + 2*(nb21/8)] * ss[4*cc + 2];
|
|
|
|
lo[i/NW] += pv4[i + 3*(nb21/8)] * ss[4*cc + 3];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
// these are needed for reducing the results from the simdgroups (reuse the ss buffer)
|
|
|
|
if (tiisg == 0) {
|
|
|
|
ss[0] = S;
|
|
|
|
ss[1] = M;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// store results to shared memory
|
|
|
|
for (short ii = 0; ii < D4; ii += NW) {
|
|
|
|
short i = ii + tiisg;
|
|
|
|
sr4[i] = lo[ii/NW];
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
// parallel reduce
|
|
|
|
for (short r = nsg/2; r > 0; r >>= 1) {
|
|
|
|
if (sgitg < r) {
|
|
|
|
const float S0 = ss[ 0];
|
|
|
|
const float S1 = ss[r*SH + 0];
|
|
|
|
|
|
|
|
const float M0 = ss[ 1];
|
|
|
|
const float M1 = ss[r*SH + 1];
|
|
|
|
|
|
|
|
const float M = max(M0, M1);
|
|
|
|
|
|
|
|
const float ms0 = exp(M0 - M);
|
|
|
|
const float ms1 = exp(M1 - M);
|
|
|
|
|
|
|
|
const float S = S0*ms0 + S1*ms1;
|
|
|
|
|
|
|
|
if (tiisg == 0) {
|
|
|
|
ss[0] = S;
|
|
|
|
ss[1] = M;
|
|
|
|
}
|
|
|
|
|
|
|
|
// O_0 = diag(ms0)*O_0 + diag(ms1)*O_1
|
|
|
|
for (short ii = 0; ii < D4; ii += NW) {
|
|
|
|
short i = ii + tiisg;
|
|
|
|
sr4[i] = sr4[i]*ms0 + sr4[i + r*D4]*ms1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
}
|
|
|
|
|
|
|
|
device float4 * dst4 = (device float4 *) dst;
|
|
|
|
|
|
|
|
// final rescale with 1/S and store to global memory
|
|
|
|
if (sgitg == 0) {
|
|
|
|
const float S = ss[0];
|
|
|
|
|
|
|
|
for (short ii = 0; ii < D4; ii += NW) {
|
|
|
|
short i = ii + tiisg;
|
|
|
|
dst4[(iq3*ne2*ne1 + iq2 + (iq1)*ne1)*D4 + i] = (float4) sr4[i]/S;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
template [[host_name("kernel_flash_attn_ext_vec_f16_h128")]] kernel flash_attn_ext_f16_t kernel_flash_attn_ext_vec_f16<128>;
|
|
|
|
template [[host_name("kernel_flash_attn_ext_vec_f16_h256")]] kernel flash_attn_ext_f16_t kernel_flash_attn_ext_vec_f16<256>;
|
|
|
|
|
2023-06-25 14:40:30 +02:00
|
|
|
kernel void kernel_cpy_f16_f16(
|
2023-12-13 20:55:03 +01:00
|
|
|
device const half * src0,
|
|
|
|
device half * dst,
|
2023-06-25 14:40:30 +02:00
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0);
|
|
|
|
|
|
|
|
device half * dst_data = (device half *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x; i00 < ne00; i00 += ntg.x) {
|
|
|
|
device const half * src = (device half *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
dst_data[i00] = src[0];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
kernel void kernel_cpy_f16_f32(
|
|
|
|
device const half * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0);
|
|
|
|
|
|
|
|
device float * dst_data = (device float *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x; i00 < ne00; i00 += ntg.x) {
|
|
|
|
device const half * src = (device half *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
dst_data[i00] = src[0];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-06-25 14:40:30 +02:00
|
|
|
kernel void kernel_cpy_f32_f16(
|
|
|
|
device const float * src0,
|
|
|
|
device half * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0);
|
|
|
|
|
|
|
|
device half * dst_data = (device half *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x; i00 < ne00; i00 += ntg.x) {
|
|
|
|
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
|
2024-05-11 09:32:41 +02:00
|
|
|
// TODO: is there a better way to handle -INFINITY?
|
|
|
|
dst_data[i00] = src[0] == -INFINITY ? -MAXHALF : src[0];
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_cpy_f32_f32(
|
|
|
|
device const float * src0,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0);
|
|
|
|
|
|
|
|
device float * dst_data = (device float *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x; i00 < ne00; i00 += ntg.x) {
|
|
|
|
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
|
|
|
|
dst_data[i00] = src[0];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
kernel void kernel_cpy_f32_q8_0(
|
|
|
|
device const float * src0,
|
|
|
|
device void * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK8_0;
|
|
|
|
|
|
|
|
device block_q8_0 * dst_data = (device block_q8_0 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x*QK8_0; i00 < ne00; i00 += ntg.x*QK8_0) {
|
|
|
|
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
|
|
|
|
float amax = 0.0f; // absolute max
|
|
|
|
|
|
|
|
for (int j = 0; j < QK8_0; j++) {
|
|
|
|
const float v = src[j];
|
|
|
|
amax = MAX(amax, fabs(v));
|
|
|
|
}
|
|
|
|
|
|
|
|
const float d = amax / ((1 << 7) - 1);
|
|
|
|
const float id = d ? 1.0f/d : 0.0f;
|
|
|
|
|
|
|
|
dst_data[i00/QK8_0].d = d;
|
|
|
|
|
|
|
|
for (int j = 0; j < QK8_0; ++j) {
|
|
|
|
const float x0 = src[j]*id;
|
|
|
|
|
|
|
|
dst_data[i00/QK8_0].qs[j] = round(x0);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_cpy_f32_q4_0(
|
|
|
|
device const float * src0,
|
|
|
|
device void * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK4_0;
|
|
|
|
|
|
|
|
device block_q4_0 * dst_data = (device block_q4_0 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x*QK4_0; i00 < ne00; i00 += ntg.x*QK4_0) {
|
|
|
|
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
|
|
|
|
float amax = 0.0f; // absolute max
|
|
|
|
float max = 0.0f;
|
|
|
|
|
|
|
|
for (int j = 0; j < QK4_0; j++) {
|
|
|
|
const float v = src[j];
|
|
|
|
if (amax < fabs(v)) {
|
|
|
|
amax = fabs(v);
|
|
|
|
max = v;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
const float d = max / -8;
|
|
|
|
const float id = d ? 1.0f/d : 0.0f;
|
|
|
|
|
|
|
|
dst_data[i00/QK4_0].d = d;
|
|
|
|
|
|
|
|
for (int j = 0; j < QK4_0/2; ++j) {
|
|
|
|
const float x0 = src[0 + j]*id;
|
|
|
|
const float x1 = src[QK4_0/2 + j]*id;
|
|
|
|
|
|
|
|
const uint8_t xi0 = MIN(15, (int8_t)(x0 + 8.5f));
|
|
|
|
const uint8_t xi1 = MIN(15, (int8_t)(x1 + 8.5f));
|
|
|
|
|
|
|
|
dst_data[i00/QK4_0].qs[j] = xi0;
|
|
|
|
dst_data[i00/QK4_0].qs[j] |= xi1 << 4;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_cpy_f32_q4_1(
|
|
|
|
device const float * src0,
|
|
|
|
device void * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK4_1;
|
|
|
|
|
|
|
|
device block_q4_1 * dst_data = (device block_q4_1 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x*QK4_1; i00 < ne00; i00 += ntg.x*QK4_1) {
|
|
|
|
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
|
|
|
|
float min = FLT_MAX;
|
|
|
|
float max = -FLT_MAX;
|
|
|
|
|
|
|
|
for (int j = 0; j < QK4_1; j++) {
|
|
|
|
const float v = src[j];
|
|
|
|
if (min > v) min = v;
|
|
|
|
if (max < v) max = v;
|
|
|
|
}
|
|
|
|
|
|
|
|
const float d = (max - min) / ((1 << 4) - 1);
|
|
|
|
const float id = d ? 1.0f/d : 0.0f;
|
|
|
|
|
|
|
|
dst_data[i00/QK4_1].d = d;
|
|
|
|
dst_data[i00/QK4_1].m = min;
|
|
|
|
|
|
|
|
for (int j = 0; j < QK4_1/2; ++j) {
|
|
|
|
const float x0 = (src[0 + j] - min)*id;
|
|
|
|
const float x1 = (src[QK4_1/2 + j] - min)*id;
|
|
|
|
|
|
|
|
const uint8_t xi0 = MIN(15, (int8_t)(x0 + 0.5f));
|
|
|
|
const uint8_t xi1 = MIN(15, (int8_t)(x1 + 0.5f));
|
|
|
|
|
|
|
|
dst_data[i00/QK4_1].qs[j] = xi0;
|
|
|
|
dst_data[i00/QK4_1].qs[j] |= xi1 << 4;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-03-27 17:55:10 +01:00
|
|
|
kernel void kernel_cpy_f32_q5_0(
|
|
|
|
device const float * src0,
|
|
|
|
device void * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK5_0;
|
|
|
|
|
|
|
|
device block_q5_0 * dst_data = (device block_q5_0 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x*QK5_0; i00 < ne00; i00 += ntg.x*QK5_0) {
|
|
|
|
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
|
|
|
|
float amax = 0.0f; // absolute max
|
|
|
|
float max = 0.0f;
|
|
|
|
|
|
|
|
for (int j = 0; j < QK5_0; j++) {
|
|
|
|
const float v = src[j];
|
|
|
|
if (amax < fabs(v)) {
|
|
|
|
amax = fabs(v);
|
|
|
|
max = v;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
const float d = max / -16;
|
|
|
|
const float id = d ? 1.0f/d : 0.0f;
|
|
|
|
|
|
|
|
dst_data[i00/QK5_0].d = d;
|
|
|
|
|
|
|
|
uint32_t qh = 0;
|
|
|
|
for (int j = 0; j < QK5_0/2; ++j) {
|
|
|
|
const float x0 = src[0 + j]*id;
|
|
|
|
const float x1 = src[QK5_0/2 + j]*id;
|
|
|
|
|
|
|
|
const uint8_t xi0 = MIN(31, (int8_t)(x0 + 16.5f));
|
|
|
|
const uint8_t xi1 = MIN(31, (int8_t)(x1 + 16.5f));
|
|
|
|
|
|
|
|
dst_data[i00/QK5_0].qs[j] = (xi0 & 0xf) | ((xi1 & 0xf) << 4);
|
|
|
|
qh |= ((xi0 & 0x10u) >> 4) << (j + 0);
|
|
|
|
qh |= ((xi1 & 0x10u) >> 4) << (j + QK5_0/2);
|
|
|
|
}
|
|
|
|
thread const uint8_t * qh8 = (thread const uint8_t *)&qh;
|
|
|
|
for (int j = 0; j < 4; ++j) {
|
|
|
|
dst_data[i00/QK5_0].qh[j] = qh8[j];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_cpy_f32_q5_1(
|
|
|
|
device const float * src0,
|
|
|
|
device void * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK5_1;
|
|
|
|
|
|
|
|
device block_q5_1 * dst_data = (device block_q5_1 *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x*QK5_1; i00 < ne00; i00 += ntg.x*QK5_1) {
|
|
|
|
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
|
|
|
|
float max = src[0];
|
|
|
|
float min = src[0];
|
|
|
|
|
|
|
|
for (int j = 1; j < QK5_1; j++) {
|
|
|
|
const float v = src[j];
|
|
|
|
min = v < min ? v : min;
|
|
|
|
max = v > max ? v : max;
|
|
|
|
}
|
|
|
|
|
|
|
|
const float d = (max - min) / 31;
|
|
|
|
const float id = d ? 1.0f/d : 0.0f;
|
|
|
|
|
|
|
|
dst_data[i00/QK5_1].d = d;
|
|
|
|
dst_data[i00/QK5_1].m = min;
|
|
|
|
|
|
|
|
uint32_t qh = 0;
|
|
|
|
for (int j = 0; j < QK5_1/2; ++j) {
|
|
|
|
const float x0 = (src[0 + j] - min)*id;
|
|
|
|
const float x1 = (src[QK5_1/2 + j] - min)*id;
|
|
|
|
|
|
|
|
const uint8_t xi0 = (uint8_t)(x0 + 0.5f);
|
|
|
|
const uint8_t xi1 = (uint8_t)(x1 + 0.5f);
|
|
|
|
|
|
|
|
dst_data[i00/QK5_1].qs[j] = (xi0 & 0xf) | ((xi1 & 0xf) << 4);
|
|
|
|
qh |= ((xi0 & 0x10u) >> 4) << (j + 0);
|
|
|
|
qh |= ((xi1 & 0x10u) >> 4) << (j + QK5_1/2);
|
|
|
|
}
|
|
|
|
thread const uint8_t * qh8 = (thread const uint8_t *)&qh;
|
|
|
|
for (int j = 0; j < 4; ++j) {
|
|
|
|
dst_data[i00/QK5_1].qh[j] = qh8[j];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline int best_index_int8(int n, constant float * val, float x) {
|
|
|
|
if (x <= val[0]) return 0;
|
|
|
|
if (x >= val[n-1]) return n-1;
|
|
|
|
int ml = 0, mu = n-1;
|
|
|
|
while (mu-ml > 1) {
|
|
|
|
int mav = (ml+mu)/2;
|
|
|
|
if (x < val[mav]) mu = mav; else ml = mav;
|
|
|
|
}
|
|
|
|
return x - val[mu-1] < val[mu] - x ? mu-1 : mu;
|
|
|
|
}
|
|
|
|
|
|
|
|
constexpr constant static float kvalues_iq4nl_f[16] = {
|
|
|
|
-127.f, -104.f, -83.f, -65.f, -49.f, -35.f, -22.f, -10.f, 1.f, 13.f, 25.f, 38.f, 53.f, 69.f, 89.f, 113.f
|
|
|
|
};
|
|
|
|
|
|
|
|
kernel void kernel_cpy_f32_iq4_nl(
|
|
|
|
device const float * src0,
|
|
|
|
device void * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i03 = tgpig[2];
|
|
|
|
const int64_t i02 = tgpig[1];
|
|
|
|
const int64_t i01 = tgpig[0];
|
|
|
|
|
|
|
|
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
|
|
|
|
|
|
|
|
const int64_t i3 = n / (ne2*ne1*ne0);
|
|
|
|
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
|
|
|
|
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
|
|
|
|
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0)/QK4_NL;
|
|
|
|
|
|
|
|
device block_iq4_nl * dst_data = (device block_iq4_nl *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
|
|
|
|
|
|
|
|
for (int64_t i00 = tpitg.x*QK4_NL; i00 < ne00; i00 += ntg.x*QK4_NL) {
|
|
|
|
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
|
|
|
|
|
|
|
|
float amax = 0.0f; // absolute max
|
|
|
|
float max = 0.0f;
|
|
|
|
|
|
|
|
for (int j = 0; j < QK4_0; j++) {
|
|
|
|
const float v = src[j];
|
|
|
|
if (amax < fabs(v)) {
|
|
|
|
amax = fabs(v);
|
|
|
|
max = v;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
const float d = max / kvalues_iq4nl_f[0];
|
|
|
|
const float id = d ? 1.0f/d : 0.0f;
|
|
|
|
|
|
|
|
float sumqx = 0, sumq2 = 0;
|
|
|
|
for (int j = 0; j < QK4_NL/2; ++j) {
|
|
|
|
const float x0 = src[0 + j]*id;
|
|
|
|
const float x1 = src[QK4_NL/2 + j]*id;
|
|
|
|
|
|
|
|
const uint8_t xi0 = best_index_int8(16, kvalues_iq4nl_f, x0);
|
|
|
|
const uint8_t xi1 = best_index_int8(16, kvalues_iq4nl_f, x1);
|
|
|
|
|
|
|
|
dst_data[i00/QK4_NL].qs[j] = xi0 | (xi1 << 4);
|
|
|
|
|
|
|
|
const float v0 = kvalues_iq4nl_f[xi0];
|
|
|
|
const float v1 = kvalues_iq4nl_f[xi1];
|
|
|
|
const float w0 = src[0 + j]*src[0 + j];
|
|
|
|
const float w1 = src[QK4_NL/2 + j]*src[QK4_NL/2 + j];
|
|
|
|
sumqx += w0*v0*src[j] + w1*v1*src[QK4_NL/2 + j];
|
|
|
|
sumq2 += w0*v0*v0 + w1*v1*v1;
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
dst_data[i00/QK4_NL].d = sumq2 > 0 ? sumqx/sumq2 : d;
|
|
|
|
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
kernel void kernel_concat(
|
2023-12-13 20:55:03 +01:00
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device char * dst,
|
2023-11-03 20:35:05 +01:00
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne03,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant uint64_t & nb03,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant uint64_t & nb13,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant int64_t & ne2,
|
|
|
|
constant int64_t & ne3,
|
|
|
|
constant uint64_t & nb0,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
constant uint64_t & nb3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint3 tpitg[[thread_position_in_threadgroup]],
|
|
|
|
uint3 ntg[[threads_per_threadgroup]]) {
|
|
|
|
|
|
|
|
const int64_t i03 = tgpig.z;
|
|
|
|
const int64_t i02 = tgpig.y;
|
|
|
|
const int64_t i01 = tgpig.x;
|
|
|
|
|
|
|
|
const int64_t i13 = i03 % ne13;
|
|
|
|
const int64_t i12 = i02 % ne12;
|
|
|
|
const int64_t i11 = i01 % ne11;
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
device const char * src0_ptr = src0 + i03*nb03 + i02*nb02 + i01*nb01 + tpitg.x*nb00;
|
2023-11-03 20:35:05 +01:00
|
|
|
device const char * src1_ptr = src1 + i13*nb13 + i12*nb12 + i11*nb11 + tpitg.x*nb10;
|
|
|
|
device char * dst_ptr = dst + i03*nb3 + i02*nb2 + i01*nb1 + tpitg.x*nb0;
|
|
|
|
|
|
|
|
for (int i0 = tpitg.x; i0 < ne0; i0 += ntg.x) {
|
|
|
|
if (i02 < ne02) {
|
|
|
|
((device float *)dst_ptr)[0] = ((device float *)src0_ptr)[0];
|
|
|
|
src0_ptr += ntg.x*nb00;
|
|
|
|
} else {
|
|
|
|
((device float *)dst_ptr)[0] = ((device float *)src1_ptr)[0];
|
|
|
|
src1_ptr += ntg.x*nb10;
|
|
|
|
}
|
|
|
|
dst_ptr += ntg.x*nb0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_q2_K_f32_impl(
|
2023-09-05 12:54:40 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int im = tgpig.z;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int ib_row = first_row * nb;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q2_K * x = (device const block_q2_K *) src0 + ib_row + offset0;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float yl[32];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int step = sizeof(block_q2_K) * nb;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-07-02 20:45:27 +02:00
|
|
|
#if QK_K == 256
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ix = tiisg/8; // 0...3
|
|
|
|
const int it = tiisg%8; // 0...7
|
2023-12-07 21:27:19 +01:00
|
|
|
const int iq = it/4; // 0 or 1
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ir = it%4; // 0...3
|
|
|
|
const int is = (8*ir)/16;// 0 or 1
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * y4 = y + ix * QK_K + 128 * iq + 8 * ir;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
for (int ib = ix; ib < nb; ib += 4) {
|
|
|
|
|
|
|
|
float4 sumy = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
yl[i+ 0] = y4[i+ 0]; sumy[0] += yl[i+ 0];
|
|
|
|
yl[i+ 8] = y4[i+32]; sumy[1] += yl[i+ 8];
|
|
|
|
yl[i+16] = y4[i+64]; sumy[2] += yl[i+16];
|
|
|
|
yl[i+24] = y4[i+96]; sumy[3] += yl[i+24];
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
device const uint8_t * sc = (device const uint8_t *)x[ib].scales + 8*iq + is;
|
|
|
|
device const uint16_t * qs = (device const uint16_t *)x[ib].qs + 16 * iq + 4 * ir;
|
2023-09-05 12:54:40 +02:00
|
|
|
device const half * dh = &x[ib].d;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
float4 acc1 = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
float4 acc2 = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int i = 0; i < 8; i += 2) {
|
|
|
|
acc1[0] += yl[i+ 0] * (qs[i/2] & 0x0003);
|
|
|
|
acc2[0] += yl[i+ 1] * (qs[i/2] & 0x0300);
|
|
|
|
acc1[1] += yl[i+ 8] * (qs[i/2] & 0x000c);
|
|
|
|
acc2[1] += yl[i+ 9] * (qs[i/2] & 0x0c00);
|
|
|
|
acc1[2] += yl[i+16] * (qs[i/2] & 0x0030);
|
|
|
|
acc2[2] += yl[i+17] * (qs[i/2] & 0x3000);
|
|
|
|
acc1[3] += yl[i+24] * (qs[i/2] & 0x00c0);
|
|
|
|
acc2[3] += yl[i+25] * (qs[i/2] & 0xc000);
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
float dall = dh[0];
|
|
|
|
float dmin = dh[1] * 1.f/16.f;
|
|
|
|
sumf[row] += dall * ((acc1[0] + 1.f/256.f * acc2[0]) * (sc[0] & 0xF) * 1.f/ 1.f +
|
|
|
|
(acc1[1] + 1.f/256.f * acc2[1]) * (sc[2] & 0xF) * 1.f/ 4.f +
|
|
|
|
(acc1[2] + 1.f/256.f * acc2[2]) * (sc[4] & 0xF) * 1.f/16.f +
|
|
|
|
(acc1[3] + 1.f/256.f * acc2[3]) * (sc[6] & 0xF) * 1.f/64.f) -
|
|
|
|
dmin * (sumy[0] * (sc[0] & 0xF0) + sumy[1] * (sc[2] & 0xF0) + sumy[2] * (sc[4] & 0xF0) + sumy[3] * (sc[6] & 0xF0));
|
|
|
|
|
|
|
|
qs += step/2;
|
|
|
|
sc += step;
|
|
|
|
dh += step/2;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
y4 += 4 * QK_K;
|
|
|
|
}
|
2023-07-02 20:45:27 +02:00
|
|
|
#else
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ix = tiisg/2; // 0...15
|
|
|
|
const int it = tiisg%2; // 0...1
|
|
|
|
|
|
|
|
device const float * y4 = y + ix * QK_K + 8 * it;
|
|
|
|
|
|
|
|
for (int ib = ix; ib < nb; ib += 16) {
|
|
|
|
|
|
|
|
float4 sumy = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
yl[i+ 0] = y4[i+ 0]; sumy[0] += yl[i+ 0];
|
|
|
|
yl[i+ 8] = y4[i+16]; sumy[1] += yl[i+ 8];
|
|
|
|
yl[i+16] = y4[i+32]; sumy[2] += yl[i+16];
|
|
|
|
yl[i+24] = y4[i+48]; sumy[3] += yl[i+24];
|
|
|
|
}
|
|
|
|
|
|
|
|
device const uint8_t * sc = (device const uint8_t *)x[ib].scales;
|
|
|
|
device const uint16_t * qs = (device const uint16_t *)x[ib].qs + 4 * it;
|
|
|
|
device const half * dh = &x[ib].d;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
float4 acc1 = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
float4 acc2 = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int i = 0; i < 8; i += 2) {
|
|
|
|
acc1[0] += yl[i+ 0] * (qs[i/2] & 0x0003);
|
|
|
|
acc2[0] += yl[i+ 1] * (qs[i/2] & 0x0300);
|
|
|
|
acc1[1] += yl[i+ 8] * (qs[i/2] & 0x000c);
|
|
|
|
acc2[1] += yl[i+ 9] * (qs[i/2] & 0x0c00);
|
|
|
|
acc1[2] += yl[i+16] * (qs[i/2] & 0x0030);
|
|
|
|
acc2[2] += yl[i+17] * (qs[i/2] & 0x3000);
|
|
|
|
acc1[3] += yl[i+24] * (qs[i/2] & 0x00c0);
|
|
|
|
acc2[3] += yl[i+25] * (qs[i/2] & 0xc000);
|
|
|
|
}
|
|
|
|
|
|
|
|
float dall = dh[0];
|
|
|
|
float dmin = dh[1];
|
|
|
|
sumf[row] += dall * ((acc1[0] + 1.f/256.f * acc2[0]) * (sc[0] & 0xF) * 1.f/ 1.f +
|
|
|
|
(acc1[1] + 1.f/256.f * acc2[1]) * (sc[1] & 0xF) * 1.f/ 4.f +
|
|
|
|
(acc1[2] + 1.f/256.f * acc2[2]) * (sc[2] & 0xF) * 1.f/16.f +
|
|
|
|
(acc1[3] + 1.f/256.f * acc2[3]) * (sc[3] & 0xF) * 1.f/64.f) -
|
|
|
|
dmin * (sumy[0] * (sc[0] >> 4) + sumy[1] * (sc[1] >> 4) + sumy[2] * (sc[2] >> 4) + sumy[3] * (sc[3] >> 4));
|
|
|
|
|
|
|
|
qs += step/2;
|
|
|
|
sc += step;
|
|
|
|
dh += step/2;
|
2023-07-02 20:45:27 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
y4 += 16 * QK_K;
|
|
|
|
}
|
2023-07-02 20:45:27 +02:00
|
|
|
#endif
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
2023-12-07 21:27:19 +01:00
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum;
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_q2_K_f32")]]
|
|
|
|
kernel void kernel_mul_mv_q2_K_f32(
|
2023-09-05 12:54:40 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-09-05 12:54:40 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
2023-12-13 20:55:03 +01:00
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
kernel_mul_mv_q2_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg);
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
#if QK_K == 256
|
|
|
|
void kernel_mul_mv_q3_K_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
|
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t r1 = tgpig.y;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int64_t im = tgpig.z;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * 2;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q3_K * x = (device const block_q3_K *) src0 + first_row*nb + offset0;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * yy = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
float yl[32];
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
//const uint16_t kmask1 = 0x3030;
|
|
|
|
//const uint16_t kmask2 = 0x0f0f;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
const int tid = tiisg/4;
|
|
|
|
const int ix = tiisg%4;
|
|
|
|
const int ip = tid/4; // 0 or 1
|
|
|
|
const int il = 2*((tid%4)/2); // 0 or 2
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ir = tid%2;
|
|
|
|
const int n = 8;
|
|
|
|
const int l0 = n*ir;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
// One would think that the Metal compiler would figure out that ip and il can only have
|
|
|
|
// 4 possible states, and optimize accordingly. Well, no. It needs help, and we do it
|
|
|
|
// with these two tales.
|
|
|
|
//
|
|
|
|
// Possible masks for the high bit
|
|
|
|
const ushort4 mm[4] = {{0x0001, 0x0100, 0x0002, 0x0200}, // ip = 0, il = 0
|
|
|
|
{0x0004, 0x0400, 0x0008, 0x0800}, // ip = 0, il = 2
|
|
|
|
{0x0010, 0x1000, 0x0020, 0x2000}, // ip = 1, il = 0
|
|
|
|
{0x0040, 0x4000, 0x0080, 0x8000}}; // ip = 1, il = 2
|
|
|
|
|
|
|
|
// Possible masks for the low 2 bits
|
|
|
|
const int4 qm[2] = {{0x0003, 0x0300, 0x000c, 0x0c00}, {0x0030, 0x3000, 0x00c0, 0xc000}};
|
|
|
|
|
|
|
|
const ushort4 hm = mm[2*ip + il/2];
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int shift = 2*il;
|
2023-09-15 11:18:18 +02:00
|
|
|
const float v1 = il == 0 ? 4.f : 64.f;
|
|
|
|
const float v2 = 4.f * v1;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const uint16_t s_shift1 = 4*ip;
|
2023-09-15 11:18:18 +02:00
|
|
|
const uint16_t s_shift2 = s_shift1 + il;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int q_offset = 32*ip + l0;
|
|
|
|
const int y_offset = 128*ip + 32*il + l0;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int step = sizeof(block_q3_K) * nb / 2;
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const float * y1 = yy + ix*QK_K + y_offset;
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
uint32_t scales32, aux32;
|
|
|
|
thread uint16_t * scales16 = (thread uint16_t *)&scales32;
|
|
|
|
thread const int8_t * scales = (thread const int8_t *)&scales32;
|
|
|
|
|
|
|
|
float sumf1[2] = {0.f};
|
|
|
|
float sumf2[2] = {0.f};
|
|
|
|
for (int i = ix; i < nb; i += 4) {
|
2023-07-02 20:45:27 +02:00
|
|
|
|
|
|
|
for (int l = 0; l < 8; ++l) {
|
2023-09-15 11:18:18 +02:00
|
|
|
yl[l+ 0] = y1[l+ 0];
|
|
|
|
yl[l+ 8] = y1[l+16];
|
|
|
|
yl[l+16] = y1[l+32];
|
|
|
|
yl[l+24] = y1[l+48];
|
2023-07-02 20:45:27 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint16_t * q = (device const uint16_t *)(x[i].qs + q_offset);
|
|
|
|
device const uint16_t * h = (device const uint16_t *)(x[i].hmask + l0);
|
|
|
|
device const uint16_t * a = (device const uint16_t *)(x[i].scales);
|
|
|
|
device const half * dh = &x[i].d;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < 2; ++row) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const float d_all = (float)dh[0];
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
scales16[0] = a[4];
|
|
|
|
scales16[1] = a[5];
|
|
|
|
aux32 = ((scales32 >> s_shift2) << 4) & 0x30303030;
|
|
|
|
scales16[0] = a[il+0];
|
|
|
|
scales16[1] = a[il+1];
|
|
|
|
scales32 = ((scales32 >> s_shift1) & 0x0f0f0f0f) | aux32;
|
|
|
|
|
|
|
|
float s1 = 0, s2 = 0, s3 = 0, s4 = 0, s5 = 0, s6 = 0;
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int l = 0; l < n; l += 2) {
|
2023-09-15 11:18:18 +02:00
|
|
|
const int32_t qs = q[l/2];
|
|
|
|
s1 += yl[l+0] * (qs & qm[il/2][0]);
|
|
|
|
s2 += yl[l+1] * (qs & qm[il/2][1]);
|
|
|
|
s3 += ((h[l/2] & hm[0]) ? 0.f : yl[l+0]) + ((h[l/2] & hm[1]) ? 0.f : yl[l+1]);
|
|
|
|
s4 += yl[l+16] * (qs & qm[il/2][2]);
|
|
|
|
s5 += yl[l+17] * (qs & qm[il/2][3]);
|
|
|
|
s6 += ((h[l/2] & hm[2]) ? 0.f : yl[l+16]) + ((h[l/2] & hm[3]) ? 0.f : yl[l+17]);
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-09-15 11:18:18 +02:00
|
|
|
float d1 = d_all * (s1 + 1.f/256.f * s2 - s3*v1);
|
|
|
|
float d2 = d_all * (s4 + 1.f/256.f * s5 - s6*v2);
|
|
|
|
sumf1[row] += d1 * (scales[0] - 32);
|
|
|
|
sumf2[row] += d2 * (scales[2] - 32);
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
s1 = s2 = s3 = s4 = s5 = s6 = 0;
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int l = 0; l < n; l += 2) {
|
2023-09-15 11:18:18 +02:00
|
|
|
const int32_t qs = q[l/2+8];
|
|
|
|
s1 += yl[l+8] * (qs & qm[il/2][0]);
|
|
|
|
s2 += yl[l+9] * (qs & qm[il/2][1]);
|
|
|
|
s3 += ((h[l/2+8] & hm[0]) ? 0.f : yl[l+8]) + ((h[l/2+8] & hm[1]) ? 0.f : yl[l+9]);
|
|
|
|
s4 += yl[l+24] * (qs & qm[il/2][2]);
|
|
|
|
s5 += yl[l+25] * (qs & qm[il/2][3]);
|
|
|
|
s6 += ((h[l/2+8] & hm[2]) ? 0.f : yl[l+24]) + ((h[l/2+8] & hm[3]) ? 0.f : yl[l+25]);
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-09-15 11:18:18 +02:00
|
|
|
d1 = d_all * (s1 + 1.f/256.f * s2 - s3*v1);
|
|
|
|
d2 = d_all * (s4 + 1.f/256.f * s5 - s6*v2);
|
|
|
|
sumf1[row] += d1 * (scales[1] - 32);
|
|
|
|
sumf2[row] += d2 * (scales[3] - 32);
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
q += step;
|
|
|
|
h += step;
|
|
|
|
a += step;
|
|
|
|
dh += step;
|
2023-07-02 20:45:27 +02:00
|
|
|
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
y1 += 4 * QK_K;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < 2; ++row) {
|
2023-09-15 11:18:18 +02:00
|
|
|
const float sumf = (sumf1[row] + 0.25f * sumf2[row]) / (1 << shift);
|
|
|
|
sumf1[row] = simd_sum(sumf);
|
|
|
|
}
|
|
|
|
if (tiisg == 0) {
|
|
|
|
for (int row = 0; row < 2; ++row) {
|
2023-12-07 21:27:19 +01:00
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = sumf1[row];
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-07-02 20:45:27 +02:00
|
|
|
#else
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_q3_K_f32_impl(
|
2023-09-05 12:54:40 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2024-04-12 18:13:20 +02:00
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
2023-09-05 12:54:40 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
2023-12-07 21:27:19 +01:00
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int nb = ne00/QK_K;
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t r1 = tgpig.y;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int64_t im = tgpig.z;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
const int row = 2 * r0 + sgitg;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q3_K * x = (device const block_q3_K *) src0 + row*nb + offset0;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * yy = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ix = tiisg/4;
|
|
|
|
const int il = 4 * (tiisg%4);// 0, 4, 8, 12
|
2023-12-07 21:27:19 +01:00
|
|
|
const int iq = il/8; // 0, 0, 1, 1
|
2023-09-05 12:54:40 +02:00
|
|
|
const int in = il%8; // 0, 4, 0, 4
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float2 sum = {0.f, 0.f};
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = ix; i < nb; i += 8) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const float d_all = (float)(x[i].d);
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint16_t * q = (device const uint16_t *)(x[i].qs + il);
|
|
|
|
device const uint16_t * h = (device const uint16_t *)(x[i].hmask + in);
|
|
|
|
device const uint16_t * s = (device const uint16_t *)(x[i].scales);
|
|
|
|
device const float * y = yy + i * QK_K + il;
|
|
|
|
|
|
|
|
const float d1 = d_all * ((int32_t)(s[0] & 0x000F) - 8);
|
|
|
|
const float d2 = d_all * ((int32_t)(s[0] & 0x00F0) - 128) * 1.f/64.f;
|
|
|
|
const float d3 = d_all * ((int32_t)(s[0] & 0x0F00) - 2048) * 1.f/4096.f;
|
|
|
|
const float d4 = d_all * ((int32_t)(s[0] & 0xF000) - 32768) * 1.f/262144.f;
|
|
|
|
|
|
|
|
for (int l = 0; l < 4; l += 2) {
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint16_t hm = h[l/2] >> iq;
|
2023-09-05 12:54:40 +02:00
|
|
|
sum[0] += y[l+ 0] * d1 * ((int32_t)(q[l/2] & 0x0003) - ((hm & 0x0001) ? 0 : 4))
|
|
|
|
+ y[l+16] * d2 * ((int32_t)(q[l/2] & 0x000c) - ((hm & 0x0004) ? 0 : 16))
|
|
|
|
+ y[l+32] * d3 * ((int32_t)(q[l/2] & 0x0030) - ((hm & 0x0010) ? 0 : 64))
|
|
|
|
+ y[l+48] * d4 * ((int32_t)(q[l/2] & 0x00c0) - ((hm & 0x0040) ? 0 : 256));
|
|
|
|
sum[1] += y[l+ 1] * d1 * ((int32_t)(q[l/2] & 0x0300) - ((hm & 0x0100) ? 0 : 1024))
|
|
|
|
+ y[l+17] * d2 * ((int32_t)(q[l/2] & 0x0c00) - ((hm & 0x0400) ? 0 : 4096))
|
|
|
|
+ y[l+33] * d3 * ((int32_t)(q[l/2] & 0x3000) - ((hm & 0x1000) ? 0 : 16384))
|
|
|
|
+ y[l+49] * d4 * ((int32_t)(q[l/2] & 0xc000) - ((hm & 0x4000) ? 0 : 65536));
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
const float sumf = sum[0] + sum[1] * 1.f/256.f;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const float tot = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
2023-12-07 21:27:19 +01:00
|
|
|
dst[r1*ne0 + im*ne0*ne1 + row] = tot;
|
2023-07-02 20:45:27 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
#endif
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_q3_K_f32")]]
|
|
|
|
kernel void kernel_mul_mv_q3_K_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
kernel_mul_mv_q3_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg);
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
#if QK_K == 256
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_q4_K_f32_impl(
|
2023-06-25 14:40:30 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const uint16_t kmask1 = 0x3f3f;
|
2023-07-02 20:45:27 +02:00
|
|
|
const uint16_t kmask2 = 0x0f0f;
|
2023-09-05 12:54:40 +02:00
|
|
|
const uint16_t kmask3 = 0xc0c0;
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ix = tiisg/8; // 0...3
|
|
|
|
const int it = tiisg%8; // 0...7
|
2023-12-07 21:27:19 +01:00
|
|
|
const int iq = it/4; // 0 or 1
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ir = it%4; // 0...3
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int im = tgpig.z;
|
2023-09-05 19:57:27 +02:00
|
|
|
//const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int first_row = r0 * N_DST;
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ib_row = first_row * nb;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q4_K * x = (device const block_q4_K *) src0 + ib_row + offset0;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float yl[16];
|
|
|
|
float yh[16];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int step = sizeof(block_q4_K) * nb / 2;
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * y4 = y + ix * QK_K + 64 * iq + 8 * ir;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
uint16_t sc16[4];
|
|
|
|
thread const uint8_t * sc8 = (thread const uint8_t *)sc16;
|
|
|
|
|
|
|
|
for (int ib = ix; ib < nb; ib += 4) {
|
|
|
|
|
|
|
|
float4 sumy = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
yl[i+0] = y4[i+ 0]; sumy[0] += yl[i+0];
|
|
|
|
yl[i+8] = y4[i+ 32]; sumy[1] += yl[i+8];
|
|
|
|
yh[i+0] = y4[i+128]; sumy[2] += yh[i+0];
|
|
|
|
yh[i+8] = y4[i+160]; sumy[3] += yh[i+8];
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
device const uint16_t * sc = (device const uint16_t *)x[ib].scales + iq;
|
|
|
|
device const uint16_t * q1 = (device const uint16_t *)x[ib].qs + 16 * iq + 4 * ir;
|
2023-09-05 12:54:40 +02:00
|
|
|
device const half * dh = &x[ib].d;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
sc16[0] = sc[0] & kmask1;
|
|
|
|
sc16[1] = sc[2] & kmask1;
|
|
|
|
sc16[2] = ((sc[4] >> 0) & kmask2) | ((sc[0] & kmask3) >> 2);
|
|
|
|
sc16[3] = ((sc[4] >> 4) & kmask2) | ((sc[2] & kmask3) >> 2);
|
|
|
|
|
|
|
|
device const uint16_t * q2 = q1 + 32;
|
|
|
|
|
|
|
|
float4 acc1 = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
float4 acc2 = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int i = 0; i < 8; i += 2) {
|
|
|
|
acc1[0] += yl[i+0] * (q1[i/2] & 0x000F);
|
|
|
|
acc1[1] += yl[i+1] * (q1[i/2] & 0x0F00);
|
|
|
|
acc1[2] += yl[i+8] * (q1[i/2] & 0x00F0);
|
|
|
|
acc1[3] += yl[i+9] * (q1[i/2] & 0xF000);
|
|
|
|
acc2[0] += yh[i+0] * (q2[i/2] & 0x000F);
|
|
|
|
acc2[1] += yh[i+1] * (q2[i/2] & 0x0F00);
|
|
|
|
acc2[2] += yh[i+8] * (q2[i/2] & 0x00F0);
|
|
|
|
acc2[3] += yh[i+9] * (q2[i/2] & 0xF000);
|
|
|
|
}
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float dall = dh[0];
|
|
|
|
float dmin = dh[1];
|
|
|
|
sumf[row] += dall * ((acc1[0] + 1.f/256.f * acc1[1]) * sc8[0] +
|
|
|
|
(acc1[2] + 1.f/256.f * acc1[3]) * sc8[1] * 1.f/16.f +
|
|
|
|
(acc2[0] + 1.f/256.f * acc2[1]) * sc8[4] +
|
|
|
|
(acc2[2] + 1.f/256.f * acc2[3]) * sc8[5] * 1.f/16.f) -
|
|
|
|
dmin * (sumy[0] * sc8[2] + sumy[1] * sc8[3] + sumy[2] * sc8[6] + sumy[3] * sc8[7]);
|
|
|
|
|
|
|
|
q1 += step;
|
|
|
|
sc += step;
|
|
|
|
dh += step;
|
2023-07-02 20:45:27 +02:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
y4 += 4 * QK_K;
|
2023-07-02 20:45:27 +02:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
2023-12-07 21:27:19 +01:00
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum;
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
#else
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_q4_K_f32_impl(
|
2023-06-25 14:40:30 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2024-04-12 18:13:20 +02:00
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
2023-09-05 12:54:40 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
2024-01-02 09:57:44 +01:00
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ix = tiisg/4; // 0...7
|
|
|
|
const int it = tiisg%4; // 0...3
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int im = tgpig.z;
|
2024-01-02 09:57:44 +01:00
|
|
|
const int first_row = r0 * N_DST;
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ib_row = first_row * nb;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q4_K * x = (device const block_q4_K *) src0 + ib_row + offset0;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float yl[8];
|
|
|
|
float yh[8];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int step = sizeof(block_q4_K) * nb / 2;
|
|
|
|
|
|
|
|
device const float * y4 = y + ix * QK_K + 8 * it;
|
|
|
|
|
|
|
|
uint16_t sc16[4];
|
|
|
|
|
|
|
|
for (int ib = ix; ib < nb; ib += 8) {
|
|
|
|
|
|
|
|
float2 sumy = {0.f, 0.f};
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
yl[i] = y4[i+ 0]; sumy[0] += yl[i];
|
|
|
|
yh[i] = y4[i+32]; sumy[1] += yh[i];
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint16_t * sc = (device const uint16_t *)x[ib].scales;
|
|
|
|
device const uint16_t * qs = (device const uint16_t *)x[ib].qs + 4 * it;
|
|
|
|
device const half * dh = x[ib].d;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < N_DST; row++) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
sc16[0] = sc[0] & 0x000f;
|
|
|
|
sc16[1] = sc[0] & 0x0f00;
|
|
|
|
sc16[2] = sc[0] & 0x00f0;
|
|
|
|
sc16[3] = sc[0] & 0xf000;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float2 acc1 = {0.f, 0.f};
|
|
|
|
float2 acc2 = {0.f, 0.f};
|
|
|
|
for (int i = 0; i < 8; i += 2) {
|
|
|
|
acc1[0] += yl[i+0] * (qs[i/2] & 0x000F);
|
|
|
|
acc1[1] += yl[i+1] * (qs[i/2] & 0x0F00);
|
|
|
|
acc2[0] += yh[i+0] * (qs[i/2] & 0x00F0);
|
|
|
|
acc2[1] += yh[i+1] * (qs[i/2] & 0xF000);
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float dall = dh[0];
|
|
|
|
float dmin = dh[1];
|
|
|
|
sumf[row] += dall * ((acc1[0] + 1.f/256.f * acc1[1]) * sc16[0] +
|
|
|
|
(acc2[0] + 1.f/256.f * acc2[1]) * sc16[1] * 1.f/4096.f) -
|
|
|
|
dmin * 1.f/16.f * (sumy[0] * sc16[2] + sumy[1] * sc16[3] * 1.f/256.f);
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
qs += step;
|
|
|
|
sc += step;
|
|
|
|
dh += step;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
y4 += 8 * QK_K;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
2024-01-02 09:57:44 +01:00
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum;
|
2023-07-02 20:45:27 +02:00
|
|
|
}
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
#endif
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_q4_K_f32")]]
|
|
|
|
kernel void kernel_mul_mv_q4_K_f32(
|
2023-06-25 14:40:30 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-09-05 12:54:40 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
kernel_mul_mv_q4_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg);
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
void kernel_mul_mv_q5_K_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2023-06-25 14:40:30 +02:00
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
|
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t r1 = tgpig.y;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int im = tgpig.z;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * 2;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q5_K * x = (device const block_q5_K *) src0 + first_row*nb + offset0;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * yy = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
float sumf[2]={0.f};
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int step = sizeof(block_q5_K) * nb;
|
2023-07-02 20:45:27 +02:00
|
|
|
|
|
|
|
#if QK_K == 256
|
2023-09-05 12:54:40 +02:00
|
|
|
#
|
|
|
|
float yl[16], yh[16];
|
2023-07-02 20:45:27 +02:00
|
|
|
|
|
|
|
const uint16_t kmask1 = 0x3f3f;
|
|
|
|
const uint16_t kmask2 = 0x0f0f;
|
|
|
|
const uint16_t kmask3 = 0xc0c0;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int tid = tiisg/4;
|
|
|
|
const int ix = tiisg%4;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int iq = tid/4;
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ir = tid%4;
|
|
|
|
const int n = 8;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int l0 = n*ir;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int q_offset = 32*iq + l0;
|
|
|
|
const int y_offset = 64*iq + l0;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint8_t hm1 = 1u << (2*iq);
|
2023-06-25 14:40:30 +02:00
|
|
|
const uint8_t hm2 = hm1 << 1;
|
|
|
|
const uint8_t hm3 = hm1 << 4;
|
|
|
|
const uint8_t hm4 = hm2 << 4;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
uint16_t sc16[4];
|
|
|
|
thread const uint8_t * sc8 = (thread const uint8_t *)sc16;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const float * y1 = yy + ix*QK_K + y_offset;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = ix; i < nb; i += 4) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint8_t * q1 = x[i].qs + q_offset;
|
|
|
|
device const uint8_t * qh = x[i].qh + l0;
|
|
|
|
device const half * dh = &x[i].d;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const uint16_t * a = (device const uint16_t *)x[i].scales + iq;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const float * y2 = y1 + 128;
|
|
|
|
float4 sumy = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int l = 0; l < 8; ++l) {
|
|
|
|
yl[l+0] = y1[l+ 0]; sumy[0] += yl[l+0];
|
|
|
|
yl[l+8] = y1[l+32]; sumy[1] += yl[l+8];
|
|
|
|
yh[l+0] = y2[l+ 0]; sumy[2] += yh[l+0];
|
|
|
|
yh[l+8] = y2[l+32]; sumy[3] += yh[l+8];
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < 2; ++row) {
|
|
|
|
|
|
|
|
device const uint8_t * q2 = q1 + 64;
|
|
|
|
|
|
|
|
sc16[0] = a[0] & kmask1;
|
|
|
|
sc16[1] = a[2] & kmask1;
|
|
|
|
sc16[2] = ((a[4] >> 0) & kmask2) | ((a[0] & kmask3) >> 2);
|
|
|
|
sc16[3] = ((a[4] >> 4) & kmask2) | ((a[2] & kmask3) >> 2);
|
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
float4 acc1 = {0.f};
|
|
|
|
float4 acc2 = {0.f};
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int l = 0; l < n; ++l) {
|
|
|
|
uint8_t h = qh[l];
|
2023-09-15 11:18:18 +02:00
|
|
|
acc1[0] += yl[l+0] * (q1[l] & 0x0F);
|
|
|
|
acc1[1] += yl[l+8] * (q1[l] & 0xF0);
|
|
|
|
acc1[2] += yh[l+0] * (q2[l] & 0x0F);
|
|
|
|
acc1[3] += yh[l+8] * (q2[l] & 0xF0);
|
|
|
|
acc2[0] += h & hm1 ? yl[l+0] : 0.f;
|
|
|
|
acc2[1] += h & hm2 ? yl[l+8] : 0.f;
|
|
|
|
acc2[2] += h & hm3 ? yh[l+0] : 0.f;
|
|
|
|
acc2[3] += h & hm4 ? yh[l+8] : 0.f;
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
const float dall = dh[0];
|
|
|
|
const float dmin = dh[1];
|
2023-09-15 11:18:18 +02:00
|
|
|
sumf[row] += dall * (sc8[0] * (acc1[0] + 16.f*acc2[0]) +
|
|
|
|
sc8[1] * (acc1[1]/16.f + 16.f*acc2[1]) +
|
|
|
|
sc8[4] * (acc1[2] + 16.f*acc2[2]) +
|
|
|
|
sc8[5] * (acc1[3]/16.f + 16.f*acc2[3])) -
|
2023-09-05 12:54:40 +02:00
|
|
|
dmin * (sumy[0] * sc8[2] + sumy[1] * sc8[3] + sumy[2] * sc8[6] + sumy[3] * sc8[7]);
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
q1 += step;
|
|
|
|
qh += step;
|
|
|
|
dh += step/2;
|
|
|
|
a += step/2;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
y1 += 4 * QK_K;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
|
|
|
}
|
2023-07-02 20:45:27 +02:00
|
|
|
#else
|
2023-09-05 12:54:40 +02:00
|
|
|
float yl[8], yh[8];
|
|
|
|
|
|
|
|
const int il = 4 * (tiisg/8); // 0, 4, 8, 12
|
|
|
|
const int ix = tiisg%8;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int iq = il/8; // 0, 0, 1, 1
|
2023-09-05 12:54:40 +02:00
|
|
|
const int in = il%8; // 0, 4, 0, 4
|
|
|
|
|
|
|
|
device const float * y = yy + ix*QK_K + il;
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = ix; i < nb; i += 8) {
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int l = 0; l < 4; ++l) {
|
|
|
|
yl[l+0] = y[l+ 0];
|
|
|
|
yl[l+4] = y[l+16];
|
|
|
|
yh[l+0] = y[l+32];
|
|
|
|
yh[l+4] = y[l+48];
|
|
|
|
}
|
|
|
|
|
|
|
|
device const half * dh = &x[i].d;
|
2023-07-02 20:45:27 +02:00
|
|
|
device const uint8_t * q = x[i].qs + il;
|
|
|
|
device const uint8_t * h = x[i].qh + in;
|
|
|
|
device const int8_t * s = x[i].scales;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < 2; ++row) {
|
|
|
|
|
|
|
|
const float d = dh[0];
|
|
|
|
|
|
|
|
float2 acc = {0.f, 0.f};
|
|
|
|
for (int l = 0; l < 4; ++l) {
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint8_t hl = h[l] >> iq;
|
2023-09-05 12:54:40 +02:00
|
|
|
acc[0] += yl[l+0] * s[0] * ((int16_t)(q[l+ 0] & 0x0F) - (hl & 0x01 ? 0 : 16))
|
|
|
|
+ yl[l+4] * s[1] * ((int16_t)(q[l+16] & 0x0F) - (hl & 0x04 ? 0 : 16));
|
|
|
|
acc[1] += yh[l+0] * s[2] * ((int16_t)(q[l+ 0] & 0xF0) - (hl & 0x10 ? 0 : 256))
|
|
|
|
+ yh[l+4] * s[3] * ((int16_t)(q[l+16] & 0xF0) - (hl & 0x40 ? 0 : 256));
|
|
|
|
}
|
|
|
|
sumf[row] += d * (acc[0] + 1.f/16.f * acc[1]);
|
|
|
|
|
|
|
|
q += step;
|
|
|
|
h += step;
|
|
|
|
s += step;
|
|
|
|
dh += step/2;
|
|
|
|
|
2023-07-02 20:45:27 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
y += 8 * QK_K;
|
2023-07-02 20:45:27 +02:00
|
|
|
}
|
|
|
|
#endif
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int row = 0; row < 2; ++row) {
|
|
|
|
const float tot = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
2023-12-07 21:27:19 +01:00
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = tot;
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_q5_K_f32")]]
|
|
|
|
kernel void kernel_mul_mv_q5_K_f32(
|
2023-06-25 14:40:30 +02:00
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-09-05 12:54:40 +02:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
2023-12-13 20:55:03 +01:00
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
kernel_mul_mv_q5_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg);
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
void kernel_mul_mv_q6_K_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2023-12-13 20:55:03 +01:00
|
|
|
|
|
|
|
const uint8_t kmask1 = 0x03;
|
|
|
|
const uint8_t kmask2 = 0x0C;
|
|
|
|
const uint8_t kmask3 = 0x30;
|
|
|
|
const uint8_t kmask4 = 0xC0;
|
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
|
|
|
|
const int64_t r0 = tgpig.x;
|
|
|
|
const int64_t r1 = tgpig.y;
|
2023-12-07 21:27:19 +01:00
|
|
|
const int im = tgpig.z;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const int row = 2 * r0 + sgitg;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const block_q6_K * x = (device const block_q6_K *) src0 + row * nb + offset0;
|
2023-12-07 21:27:19 +01:00
|
|
|
device const float * yy = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-07-02 20:45:27 +02:00
|
|
|
float sumf = 0;
|
|
|
|
|
|
|
|
#if QK_K == 256
|
2023-09-05 12:54:40 +02:00
|
|
|
const int tid = tiisg/2;
|
|
|
|
const int ix = tiisg%2;
|
|
|
|
const int ip = tid/8; // 0 or 1
|
|
|
|
const int il = tid%8;
|
2023-06-25 14:40:30 +02:00
|
|
|
const int n = 4;
|
|
|
|
const int l0 = n*il;
|
|
|
|
const int is = 8*ip + l0/16;
|
|
|
|
|
|
|
|
const int y_offset = 128*ip + l0;
|
|
|
|
const int q_offset_l = 64*ip + l0;
|
|
|
|
const int q_offset_h = 32*ip + l0;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = ix; i < nb; i += 2) {
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint8_t * q1 = x[i].ql + q_offset_l;
|
|
|
|
device const uint8_t * q2 = q1 + 32;
|
2023-06-25 14:40:30 +02:00
|
|
|
device const uint8_t * qh = x[i].qh + q_offset_h;
|
|
|
|
device const int8_t * sc = x[i].scales + is;
|
|
|
|
|
|
|
|
device const float * y = yy + i * QK_K + y_offset;
|
|
|
|
|
|
|
|
const float dall = x[i].d;
|
|
|
|
|
|
|
|
float4 sums = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int l = 0; l < n; ++l) {
|
2023-09-05 12:54:40 +02:00
|
|
|
sums[0] += y[l+ 0] * ((int8_t)((q1[l] & 0xF) | ((qh[l] & kmask1) << 4)) - 32);
|
|
|
|
sums[1] += y[l+32] * ((int8_t)((q2[l] & 0xF) | ((qh[l] & kmask2) << 2)) - 32);
|
|
|
|
sums[2] += y[l+64] * ((int8_t)((q1[l] >> 4) | ((qh[l] & kmask3) << 0)) - 32);
|
|
|
|
sums[3] += y[l+96] * ((int8_t)((q2[l] >> 4) | ((qh[l] & kmask4) >> 2)) - 32);
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
sumf += dall * (sums[0] * sc[0] + sums[1] * sc[2] + sums[2] * sc[4] + sums[3] * sc[6]);
|
|
|
|
|
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-07-02 20:45:27 +02:00
|
|
|
#else
|
2023-09-05 12:54:40 +02:00
|
|
|
const int ix = tiisg/4;
|
|
|
|
const int il = 4*(tiisg%4);
|
2023-07-02 20:45:27 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = ix; i < nb; i += 8) {
|
2023-07-02 20:45:27 +02:00
|
|
|
device const float * y = yy + i * QK_K + il;
|
|
|
|
device const uint8_t * ql = x[i].ql + il;
|
|
|
|
device const uint8_t * qh = x[i].qh + il;
|
|
|
|
device const int8_t * s = x[i].scales;
|
|
|
|
|
|
|
|
const float d = x[i].d;
|
|
|
|
|
|
|
|
float4 sums = {0.f, 0.f, 0.f, 0.f};
|
|
|
|
for (int l = 0; l < 4; ++l) {
|
|
|
|
sums[0] += y[l+ 0] * ((int8_t)((ql[l+ 0] & 0xF) | ((qh[l] & kmask1) << 4)) - 32);
|
|
|
|
sums[1] += y[l+16] * ((int8_t)((ql[l+16] & 0xF) | ((qh[l] & kmask2) << 2)) - 32);
|
|
|
|
sums[2] += y[l+32] * ((int8_t)((ql[l+ 0] >> 4) | ((qh[l] & kmask3) >> 0)) - 32);
|
|
|
|
sums[3] += y[l+48] * ((int8_t)((ql[l+16] >> 4) | ((qh[l] & kmask4) >> 2)) - 32);
|
|
|
|
}
|
|
|
|
sumf += d * (sums[0] * s[0] + sums[1] * s[1] + sums[2] * s[2] + sums[3] * s[3]);
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
const float tot = simd_sum(sumf);
|
|
|
|
if (tiisg == 0) {
|
2023-12-07 21:27:19 +01:00
|
|
|
dst[r1*ne0 + im*ne0*ne1 + row] = tot;
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
[[host_name("kernel_mul_mv_q6_K_f32")]]
|
|
|
|
kernel void kernel_mul_mv_q6_K_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
kernel_mul_mv_q6_K_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg);
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
|
|
|
|
2024-01-08 16:02:32 +01:00
|
|
|
// ======================= "True" 2-bit
|
|
|
|
|
|
|
|
void kernel_mul_mv_iq2_xxs_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-01-08 16:02:32 +01:00
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
|
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
|
|
|
device const block_iq2_xxs * x = (device const block_iq2_xxs *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
|
|
|
float yl[32];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int nb32 = nb * (QK_K / 32);
|
|
|
|
|
|
|
|
threadgroup uint64_t * values = (threadgroup uint64_t *)shared_values;
|
|
|
|
threadgroup uint8_t * shared_signs = (threadgroup uint8_t *)(values + 256);
|
|
|
|
{
|
|
|
|
int nval = 4;
|
|
|
|
int pos = (32*sgitg + tiisg)*nval;
|
2024-01-11 20:39:39 +01:00
|
|
|
for (int i = 0; i < nval; ++i) values[pos + i] = iq2xxs_grid[pos + i];
|
2024-01-08 16:02:32 +01:00
|
|
|
nval = 2;
|
|
|
|
pos = (32*sgitg + tiisg)*nval;
|
|
|
|
for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i];
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ix = tiisg;
|
|
|
|
|
|
|
|
device const float * y4 = y + 32 * ix;
|
|
|
|
|
|
|
|
for (int ib32 = ix; ib32 < nb32; ib32 += 32) {
|
|
|
|
|
|
|
|
for (int i = 0; i < 32; ++i) {
|
|
|
|
yl[i] = y4[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ibl = ib32 / (QK_K / 32);
|
|
|
|
const int ib = ib32 % (QK_K / 32);
|
|
|
|
|
|
|
|
device const block_iq2_xxs * xr = x + ibl;
|
|
|
|
device const uint16_t * q2 = xr->qs + 4 * ib;
|
|
|
|
device const half * dh = &xr->d;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
const float db = dh[0];
|
|
|
|
device const uint8_t * aux8 = (device const uint8_t *)q2;
|
|
|
|
const uint32_t aux32 = q2[2] | (q2[3] << 16);
|
|
|
|
const float d = db * (0.5f + (aux32 >> 28));
|
|
|
|
|
|
|
|
float sum = 0;
|
|
|
|
for (int l = 0; l < 4; ++l) {
|
|
|
|
const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(values + aux8[l]);
|
|
|
|
const uint8_t signs = shared_signs[(aux32 >> 7*l) & 127];
|
|
|
|
for (int j = 0; j < 8; ++j) {
|
|
|
|
sum += yl[8*l + j] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
sumf[row] += d * sum;
|
|
|
|
|
|
|
|
dh += nb*sizeof(block_iq2_xxs)/2;
|
|
|
|
q2 += nb*sizeof(block_iq2_xxs)/2;
|
|
|
|
}
|
|
|
|
|
|
|
|
y4 += 32 * 32;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.25f;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
[[host_name("kernel_mul_mv_iq2_xxs_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq2_xxs_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
|
|
|
kernel_mul_mv_iq2_xxs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
|
|
|
}
|
|
|
|
|
2024-01-11 20:39:39 +01:00
|
|
|
void kernel_mul_mv_iq2_xs_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-01-11 20:39:39 +01:00
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
|
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
|
|
|
device const block_iq2_xs * x = (device const block_iq2_xs *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
|
|
|
float yl[32];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int nb32 = nb * (QK_K / 32);
|
|
|
|
|
|
|
|
threadgroup uint64_t * values = (threadgroup uint64_t *)shared_values;
|
|
|
|
threadgroup uint8_t * shared_signs = (threadgroup uint8_t *)(values + 512);
|
|
|
|
{
|
|
|
|
int nval = 8;
|
|
|
|
int pos = (32*sgitg + tiisg)*nval;
|
|
|
|
for (int i = 0; i < nval; ++i) values[pos + i] = iq2xs_grid[pos + i];
|
|
|
|
nval = 2;
|
|
|
|
pos = (32*sgitg + tiisg)*nval;
|
|
|
|
for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i];
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ix = tiisg;
|
|
|
|
|
|
|
|
device const float * y4 = y + 32 * ix;
|
|
|
|
|
|
|
|
for (int ib32 = ix; ib32 < nb32; ib32 += 32) {
|
|
|
|
|
|
|
|
for (int i = 0; i < 32; ++i) {
|
|
|
|
yl[i] = y4[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ibl = ib32 / (QK_K / 32);
|
|
|
|
const int ib = ib32 % (QK_K / 32);
|
|
|
|
|
|
|
|
device const block_iq2_xs * xr = x + ibl;
|
|
|
|
device const uint16_t * q2 = xr->qs + 4 * ib;
|
|
|
|
device const uint8_t * sc = xr->scales + ib;
|
|
|
|
device const half * dh = &xr->d;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
const float db = dh[0];
|
|
|
|
const uint8_t ls1 = sc[0] & 0xf;
|
|
|
|
const uint8_t ls2 = sc[0] >> 4;
|
|
|
|
const float d1 = db * (0.5f + ls1);
|
|
|
|
const float d2 = db * (0.5f + ls2);
|
|
|
|
|
|
|
|
float sum1 = 0, sum2 = 0;
|
|
|
|
for (int l = 0; l < 2; ++l) {
|
|
|
|
const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(values + (q2[l] & 511));
|
|
|
|
const uint8_t signs = shared_signs[(q2[l] >> 9)];
|
|
|
|
for (int j = 0; j < 8; ++j) {
|
|
|
|
sum1 += yl[8*l + j] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
for (int l = 2; l < 4; ++l) {
|
|
|
|
const threadgroup uint8_t * grid = (const threadgroup uint8_t *)(values + (q2[l] & 511));
|
|
|
|
const uint8_t signs = shared_signs[(q2[l] >> 9)];
|
|
|
|
for (int j = 0; j < 8; ++j) {
|
|
|
|
sum2 += yl[8*l + j] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
sumf[row] += d1 * sum1 + d2 * sum2;
|
|
|
|
|
|
|
|
dh += nb*sizeof(block_iq2_xs)/2;
|
|
|
|
q2 += nb*sizeof(block_iq2_xs)/2;
|
|
|
|
sc += nb*sizeof(block_iq2_xs);
|
|
|
|
}
|
|
|
|
|
|
|
|
y4 += 32 * 32;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.25f;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
[[host_name("kernel_mul_mv_iq2_xs_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq2_xs_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
|
|
|
kernel_mul_mv_iq2_xs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
|
|
|
}
|
|
|
|
|
2024-01-30 14:14:12 +01:00
|
|
|
void kernel_mul_mv_iq3_xxs_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-01-30 14:14:12 +01:00
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
|
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
|
|
|
device const block_iq3_xxs * x = (device const block_iq3_xxs *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
|
|
|
float yl[32];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int nb32 = nb * (QK_K / 32);
|
|
|
|
|
|
|
|
threadgroup uint32_t * values = (threadgroup uint32_t *)shared_values;
|
|
|
|
threadgroup uint8_t * shared_signs = (threadgroup uint8_t *)(values + 256);
|
|
|
|
{
|
|
|
|
int nval = 4;
|
|
|
|
int pos = (32*sgitg + tiisg)*nval;
|
|
|
|
for (int i = 0; i < nval; ++i) values[pos + i] = iq3xxs_grid[pos + i];
|
|
|
|
nval = 2;
|
|
|
|
pos = (32*sgitg + tiisg)*nval;
|
|
|
|
for (int i = 0; i < nval; ++i) shared_signs[pos+i] = ksigns_iq2xs[pos+i];
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ix = tiisg;
|
|
|
|
|
|
|
|
device const float * y4 = y + 32 * ix;
|
|
|
|
|
|
|
|
for (int ib32 = ix; ib32 < nb32; ib32 += 32) {
|
|
|
|
|
|
|
|
for (int i = 0; i < 32; ++i) {
|
|
|
|
yl[i] = y4[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ibl = ib32 / (QK_K / 32);
|
|
|
|
const int ib = ib32 % (QK_K / 32);
|
|
|
|
|
|
|
|
device const block_iq3_xxs * xr = x + ibl;
|
|
|
|
device const uint8_t * q3 = xr->qs + 8 * ib;
|
|
|
|
device const uint16_t * gas = (device const uint16_t *)(xr->qs + QK_K/4) + 2 * ib;
|
|
|
|
device const half * dh = &xr->d;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
const float db = dh[0];
|
|
|
|
const uint32_t aux32 = gas[0] | (gas[1] << 16);
|
|
|
|
const float d = db * (0.5f + (aux32 >> 28));
|
|
|
|
|
|
|
|
float2 sum = {0};
|
|
|
|
for (int l = 0; l < 4; ++l) {
|
|
|
|
const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(values + q3[2*l+0]);
|
|
|
|
const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(values + q3[2*l+1]);
|
|
|
|
const uint8_t signs = shared_signs[(aux32 >> 7*l) & 127];
|
|
|
|
for (int j = 0; j < 4; ++j) {
|
|
|
|
sum[0] += yl[8*l + j + 0] * grid1[j] * (signs & kmask_iq2xs[j+0] ? -1.f : 1.f);
|
|
|
|
sum[1] += yl[8*l + j + 4] * grid2[j] * (signs & kmask_iq2xs[j+4] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
sumf[row] += d * (sum[0] + sum[1]);
|
|
|
|
|
|
|
|
dh += nb*sizeof(block_iq3_xxs)/2;
|
|
|
|
q3 += nb*sizeof(block_iq3_xxs);
|
|
|
|
gas += nb*sizeof(block_iq3_xxs)/2;
|
|
|
|
}
|
|
|
|
|
|
|
|
y4 += 32 * 32;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.5f;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
[[host_name("kernel_mul_mv_iq3_xxs_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq3_xxs_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
|
|
|
kernel_mul_mv_iq3_xxs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
|
|
|
}
|
|
|
|
|
2024-02-24 15:23:52 +01:00
|
|
|
void kernel_mul_mv_iq3_s_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-02-24 15:23:52 +01:00
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
|
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
|
|
|
device const block_iq3_s * x = (device const block_iq3_s *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
|
|
|
float yl[32];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int nb32 = nb * (QK_K / 32);
|
|
|
|
|
|
|
|
threadgroup uint32_t * values = (threadgroup uint32_t *)shared_values;
|
|
|
|
{
|
|
|
|
int nval = 8;
|
|
|
|
int pos = (32*sgitg + tiisg)*nval;
|
2024-03-02 16:00:51 +01:00
|
|
|
for (int i = 0; i < nval; ++i) values[pos + i] = iq3s_grid[pos + i];
|
2024-02-24 15:23:52 +01:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ix = tiisg;
|
|
|
|
|
|
|
|
device const float * y4 = y + 32 * ix;
|
|
|
|
|
|
|
|
for (int ib32 = ix; ib32 < nb32; ib32 += 32) {
|
|
|
|
|
|
|
|
for (int i = 0; i < 32; ++i) {
|
|
|
|
yl[i] = y4[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ibl = ib32 / (QK_K / 32);
|
|
|
|
const int ib = ib32 % (QK_K / 32);
|
|
|
|
|
|
|
|
device const block_iq3_s * xr = x + ibl;
|
|
|
|
device const uint8_t * qs = xr->qs + 8 * ib;
|
|
|
|
device const uint8_t * qh = xr->qh + ib;
|
|
|
|
device const uint8_t * sc = xr->scales + (ib/2);
|
|
|
|
device const uint8_t * signs = xr->signs + 4 * ib;
|
|
|
|
device const half * dh = &xr->d;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
const float db = dh[0];
|
2024-03-02 16:00:51 +01:00
|
|
|
const float d = db * (1 + 2*((sc[0] >> 4*(ib%2)) & 0xf));
|
2024-02-24 15:23:52 +01:00
|
|
|
|
|
|
|
float2 sum = {0};
|
|
|
|
for (int l = 0; l < 4; ++l) {
|
2024-03-02 16:00:51 +01:00
|
|
|
const threadgroup uint32_t * table1 = qh[0] & kmask_iq2xs[2*l+0] ? values + 256 : values;
|
|
|
|
const threadgroup uint32_t * table2 = qh[0] & kmask_iq2xs[2*l+1] ? values + 256 : values;
|
|
|
|
const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(table1 + qs[2*l+0]);
|
|
|
|
const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(table2 + qs[2*l+1]);
|
2024-02-24 15:23:52 +01:00
|
|
|
for (int j = 0; j < 4; ++j) {
|
|
|
|
sum[0] += yl[8*l + j + 0] * grid1[j] * select(1, -1, signs[l] & kmask_iq2xs[j+0]);
|
|
|
|
sum[1] += yl[8*l + j + 4] * grid2[j] * select(1, -1, signs[l] & kmask_iq2xs[j+4]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
sumf[row] += d * (sum[0] + sum[1]);
|
|
|
|
|
|
|
|
dh += nb*sizeof(block_iq3_s)/2;
|
|
|
|
qs += nb*sizeof(block_iq3_s);
|
|
|
|
qh += nb*sizeof(block_iq3_s);
|
|
|
|
sc += nb*sizeof(block_iq3_s);
|
|
|
|
signs += nb*sizeof(block_iq3_s);
|
|
|
|
}
|
|
|
|
|
|
|
|
y4 += 32 * 32;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
2024-03-02 16:00:51 +01:00
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum;
|
2024-02-24 15:23:52 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
[[host_name("kernel_mul_mv_iq3_s_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq3_s_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
|
|
|
kernel_mul_mv_iq3_s_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
|
|
|
}
|
|
|
|
|
2024-02-26 17:28:38 +01:00
|
|
|
void kernel_mul_mv_iq2_s_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-02-26 17:28:38 +01:00
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
|
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
|
|
|
|
device const block_iq2_s * x = (device const block_iq2_s *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
|
|
|
float yl[32];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int nb32 = nb * (QK_K / 32);
|
|
|
|
|
|
|
|
//threadgroup uint64_t * values = (threadgroup uint64_t *)shared_values;
|
|
|
|
//{
|
|
|
|
// int nval = 32;
|
|
|
|
// int pos = (32*sgitg + tiisg)*nval;
|
|
|
|
// for (int i = 0; i < nval; ++i) values[pos + i] = iq2s_grid[pos + i];
|
|
|
|
// threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
//}
|
|
|
|
|
|
|
|
const int ix = tiisg;
|
|
|
|
|
|
|
|
device const float * y4 = y + 32 * ix;
|
|
|
|
|
|
|
|
for (int ib32 = ix; ib32 < nb32; ib32 += 32) {
|
|
|
|
|
|
|
|
for (int i = 0; i < 32; ++i) {
|
|
|
|
yl[i] = y4[i];
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ibl = ib32 / (QK_K / 32);
|
|
|
|
const int ib = ib32 % (QK_K / 32);
|
|
|
|
|
|
|
|
device const block_iq2_s * xr = x + ibl;
|
|
|
|
device const uint8_t * qs = xr->qs + 4 * ib;
|
|
|
|
device const uint8_t * qh = xr->qh + ib;
|
|
|
|
device const uint8_t * sc = xr->scales + ib;
|
|
|
|
device const uint8_t * signs = qs + QK_K/8;
|
|
|
|
device const half * dh = &xr->d;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
const float db = dh[0];
|
|
|
|
const float d1 = db * (0.5f + (sc[0] & 0xf));
|
|
|
|
const float d2 = db * (0.5f + (sc[0] >> 4));
|
|
|
|
|
|
|
|
float2 sum = {0};
|
|
|
|
for (int l = 0; l < 2; ++l) {
|
|
|
|
//const threadgroup uint8_t * grid1 = (const threadgroup uint8_t *)(values + (qs[l+0] | ((qh[0] << (8-2*l)) & 0x300)));
|
|
|
|
//const threadgroup uint8_t * grid2 = (const threadgroup uint8_t *)(values + (qs[l+2] | ((qh[0] << (4-2*l)) & 0x300)));
|
|
|
|
constant uint8_t * grid1 = (constant uint8_t *)(iq2s_grid + (qs[l+0] | ((qh[0] << (8-2*l)) & 0x300)));
|
|
|
|
constant uint8_t * grid2 = (constant uint8_t *)(iq2s_grid + (qs[l+2] | ((qh[0] << (4-2*l)) & 0x300)));
|
|
|
|
for (int j = 0; j < 8; ++j) {
|
|
|
|
sum[0] += yl[8*l + j + 0] * grid1[j] * select(1, -1, signs[l+0] & kmask_iq2xs[j]);
|
|
|
|
sum[1] += yl[8*l + j + 16] * grid2[j] * select(1, -1, signs[l+2] & kmask_iq2xs[j]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
sumf[row] += d1 * sum[0] + d2 * sum[1];
|
|
|
|
|
|
|
|
dh += nb*sizeof(block_iq2_s)/2;
|
|
|
|
qs += nb*sizeof(block_iq2_s);
|
|
|
|
qh += nb*sizeof(block_iq2_s);
|
|
|
|
sc += nb*sizeof(block_iq2_s);
|
|
|
|
signs += nb*sizeof(block_iq2_s);
|
|
|
|
}
|
|
|
|
|
|
|
|
y4 += 32 * 32;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum * 0.25f;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
[[host_name("kernel_mul_mv_iq2_s_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq2_s_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
|
|
|
kernel_mul_mv_iq2_s_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
|
|
|
}
|
|
|
|
|
2024-02-18 17:16:55 +01:00
|
|
|
void kernel_mul_mv_iq1_s_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_value,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-02-18 17:16:55 +01:00
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
|
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
device const block_iq1_s * x = (device const block_iq1_s *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
2024-03-11 07:51:49 +01:00
|
|
|
float yl[32];
|
2024-02-18 17:16:55 +01:00
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int nb32 = nb * (QK_K / 32);
|
|
|
|
|
2024-03-11 07:51:49 +01:00
|
|
|
const int ix = tiisg;
|
2024-02-18 17:16:55 +01:00
|
|
|
|
2024-03-11 07:51:49 +01:00
|
|
|
device const float * y4 = y + 32 * ix;
|
2024-02-18 17:16:55 +01:00
|
|
|
|
2024-03-11 07:51:49 +01:00
|
|
|
for (int ib32 = ix; ib32 < nb32; ib32 += 32) {
|
2024-02-18 17:16:55 +01:00
|
|
|
|
2024-03-11 07:51:49 +01:00
|
|
|
float sumy = 0;
|
|
|
|
for (int i = 0; i < 32; ++i) {
|
2024-02-18 17:16:55 +01:00
|
|
|
yl[i] = y4[i];
|
2024-03-11 07:51:49 +01:00
|
|
|
sumy += yl[i];
|
2024-02-18 17:16:55 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
const int ibl = ib32 / (QK_K / 32);
|
|
|
|
const int ib = ib32 % (QK_K / 32);
|
|
|
|
|
|
|
|
device const block_iq1_s * xr = x + ibl;
|
2024-03-11 07:51:49 +01:00
|
|
|
device const uint8_t * qs = xr->qs + 4 * ib;
|
|
|
|
device const uint16_t * qh = xr->qh + ib;
|
|
|
|
device const half * dh = &xr->d;
|
2024-02-18 17:16:55 +01:00
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
2024-03-11 07:51:49 +01:00
|
|
|
constant uint8_t * grid1 = (constant uint8_t *)(iq1s_grid_gpu + (qs[0] | ((qh[0] << 8) & 0x700)));
|
|
|
|
constant uint8_t * grid2 = (constant uint8_t *)(iq1s_grid_gpu + (qs[1] | ((qh[0] << 5) & 0x700)));
|
|
|
|
constant uint8_t * grid3 = (constant uint8_t *)(iq1s_grid_gpu + (qs[2] | ((qh[0] << 2) & 0x700)));
|
|
|
|
constant uint8_t * grid4 = (constant uint8_t *)(iq1s_grid_gpu + (qs[3] | ((qh[0] >> 1) & 0x700)));
|
2024-02-18 17:16:55 +01:00
|
|
|
|
2024-03-11 07:51:49 +01:00
|
|
|
float sum = 0;
|
|
|
|
for (int j = 0; j < 4; ++j) {
|
|
|
|
sum += yl[j+ 0] * (grid1[j] & 0xf) + yl[j+ 4] * (grid1[j] >> 4)
|
|
|
|
+ yl[j+ 8] * (grid2[j] & 0xf) + yl[j+12] * (grid2[j] >> 4)
|
|
|
|
+ yl[j+16] * (grid3[j] & 0xf) + yl[j+20] * (grid3[j] >> 4)
|
|
|
|
+ yl[j+24] * (grid4[j] & 0xf) + yl[j+28] * (grid4[j] >> 4);
|
2024-02-18 17:16:55 +01:00
|
|
|
}
|
2024-03-11 16:53:15 +01:00
|
|
|
sumf[row] += (float)dh[0] * (sum + sumy * (qh[0] & 0x8000 ? -1 - IQ1S_DELTA : -1 + IQ1S_DELTA)) * (2*((qh[0] >> 12) & 7) + 1);
|
2024-02-18 17:16:55 +01:00
|
|
|
|
|
|
|
dh += nb*sizeof(block_iq1_s)/2;
|
|
|
|
qs += nb*sizeof(block_iq1_s);
|
2024-03-11 07:51:49 +01:00
|
|
|
qh += nb*sizeof(block_iq1_s)/2;
|
2024-02-18 17:16:55 +01:00
|
|
|
}
|
|
|
|
|
2024-03-11 07:51:49 +01:00
|
|
|
y4 += 32 * 32;
|
2024-02-18 17:16:55 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-03-27 17:55:10 +01:00
|
|
|
void kernel_mul_mv_iq1_m_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_value,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-03-27 17:55:10 +01:00
|
|
|
|
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
|
|
|
|
const int first_row = (r0 * N_SIMDGROUP + sgitg) * N_DST;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
device const block_iq1_m * x = (device const block_iq1_m *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
|
|
|
float yl[32];
|
|
|
|
float sumf[N_DST]={0.f}, all_sum;
|
|
|
|
|
|
|
|
const int nb32 = nb * (QK_K / 32);
|
|
|
|
|
|
|
|
const int ix = tiisg;
|
|
|
|
|
|
|
|
device const float * y4 = y + 32 * ix;
|
|
|
|
|
|
|
|
#if QK_K != 64
|
|
|
|
iq1m_scale_t scale;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
for (int ib32 = ix; ib32 < nb32; ib32 += 32) {
|
|
|
|
|
|
|
|
float4 sumy = {0.f};
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
yl[i+ 0] = y4[i+ 0]; sumy[0] += yl[i+ 0];
|
|
|
|
yl[i+ 8] = y4[i+ 8]; sumy[1] += yl[i+ 8];
|
|
|
|
yl[i+16] = y4[i+16]; sumy[2] += yl[i+16];
|
|
|
|
yl[i+24] = y4[i+24]; sumy[3] += yl[i+24];
|
|
|
|
}
|
|
|
|
|
|
|
|
const int ibl = ib32 / (QK_K / 32);
|
|
|
|
const int ib = ib32 % (QK_K / 32);
|
|
|
|
|
|
|
|
device const block_iq1_m * xr = x + ibl;
|
|
|
|
device const uint8_t * qs = xr->qs + 4 * ib;
|
|
|
|
device const uint8_t * qh = xr->qh + 2 * ib;
|
|
|
|
device const uint16_t * sc = (device const uint16_t *)xr->scales;
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; row++) {
|
|
|
|
|
|
|
|
#if QK_K != 64
|
|
|
|
scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000);
|
|
|
|
#endif
|
|
|
|
|
|
|
|
constant uint8_t * grid1 = (constant uint8_t *)(iq1s_grid_gpu + (qs[0] | ((qh[0] << 8) & 0x700)));
|
|
|
|
constant uint8_t * grid2 = (constant uint8_t *)(iq1s_grid_gpu + (qs[1] | ((qh[0] << 4) & 0x700)));
|
|
|
|
constant uint8_t * grid3 = (constant uint8_t *)(iq1s_grid_gpu + (qs[2] | ((qh[1] << 8) & 0x700)));
|
|
|
|
constant uint8_t * grid4 = (constant uint8_t *)(iq1s_grid_gpu + (qs[3] | ((qh[1] << 4) & 0x700)));
|
|
|
|
|
|
|
|
float2 sum = {0.f};
|
|
|
|
for (int j = 0; j < 4; ++j) {
|
|
|
|
sum[0] += yl[j+ 0] * (grid1[j] & 0xf) + yl[j+ 4] * (grid1[j] >> 4)
|
|
|
|
+ yl[j+ 8] * (grid2[j] & 0xf) + yl[j+12] * (grid2[j] >> 4);
|
|
|
|
sum[1] += yl[j+16] * (grid3[j] & 0xf) + yl[j+20] * (grid3[j] >> 4)
|
|
|
|
+ yl[j+24] * (grid4[j] & 0xf) + yl[j+28] * (grid4[j] >> 4);
|
|
|
|
}
|
|
|
|
const float delta1 = sumy[0] * (qh[0] & 0x08 ? -1 - IQ1M_DELTA : -1 + IQ1M_DELTA) + sumy[1] * (qh[0] & 0x80 ? -1 - IQ1M_DELTA : -1 + IQ1M_DELTA);
|
|
|
|
const float delta2 = sumy[2] * (qh[1] & 0x08 ? -1 - IQ1M_DELTA : -1 + IQ1M_DELTA) + sumy[3] * (qh[1] & 0x80 ? -1 - IQ1M_DELTA : -1 + IQ1M_DELTA);
|
|
|
|
#if QK_K == 64
|
|
|
|
const float d = (float) *((device const half *)(sc - 1));
|
|
|
|
sumf[row] += d * ((sum[0] + delta1) * (2*((sc[0] >> (8*(ib%2)+0)) & 0xf) + 1) +
|
|
|
|
(sum[1] + delta2) * (2*((sc[0] >> (8*(ib%2)+4)) & 0xf) + 1));
|
|
|
|
#else
|
|
|
|
sumf[row] += (float)scale.f16 * ((sum[0] + delta1) * (2*((sc[ib/2] >> (6*(ib%2)+0)) & 7) + 1) +
|
|
|
|
(sum[1] + delta2) * (2*((sc[ib/2] >> (6*(ib%2)+3)) & 7) + 1));
|
|
|
|
#endif
|
|
|
|
|
|
|
|
sc += nb*sizeof(block_iq1_m)/2;
|
|
|
|
qs += nb*sizeof(block_iq1_m);
|
|
|
|
qh += nb*sizeof(block_iq1_m);
|
|
|
|
}
|
|
|
|
|
|
|
|
y4 += 32 * 32;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < N_DST; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2024-02-21 15:19:39 +01:00
|
|
|
|
|
|
|
void kernel_mul_mv_iq4_nl_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values_i8,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-02-21 15:19:39 +01:00
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
threadgroup float * shared_values = (threadgroup float *)shared_values_i8;
|
2024-02-21 15:19:39 +01:00
|
|
|
const int nb = ne00/QK4_NL;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
const int first_row = (r0 * 2 + sgitg) * 2;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
device const block_iq4_nl * x = (device const block_iq4_nl *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
|
|
|
const int ix = tiisg/2; // 0...15
|
|
|
|
const int it = tiisg%2; // 0 or 1
|
|
|
|
|
|
|
|
shared_values[tiisg] = kvalues_iq4nl_f[tiisg%16];
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
float4 yl[4];
|
|
|
|
float sumf[2]={0.f}, all_sum;
|
|
|
|
|
|
|
|
device const float * yb = y + ix * QK4_NL + it * 8;
|
|
|
|
|
|
|
|
uint32_t aux32[2];
|
|
|
|
thread const uint8_t * q8 = (thread const uint8_t *)aux32;
|
|
|
|
|
|
|
|
float4 qf1, qf2;
|
|
|
|
|
|
|
|
for (int ib = ix; ib < nb; ib += 16) {
|
|
|
|
|
|
|
|
device const float4 * y4 = (device const float4 *)yb;
|
|
|
|
yl[0] = y4[0]; yl[1] = y4[4]; yl[2] = y4[1]; yl[3] = y4[5];
|
|
|
|
|
|
|
|
for (int row = 0; row < 2; ++row) {
|
|
|
|
|
|
|
|
device const block_iq4_nl & xb = x[row*nb + ib];
|
|
|
|
device const uint16_t * q4 = (device const uint16_t *)(xb.qs + 8*it);
|
|
|
|
|
|
|
|
float4 acc1 = {0.f}, acc2 = {0.f};
|
|
|
|
|
|
|
|
aux32[0] = q4[0] | (q4[1] << 16);
|
|
|
|
aux32[1] = (aux32[0] >> 4) & 0x0f0f0f0f;
|
|
|
|
aux32[0] &= 0x0f0f0f0f;
|
|
|
|
qf1 = {shared_values[q8[0]], shared_values[q8[1]], shared_values[q8[2]], shared_values[q8[3]]};
|
|
|
|
qf2 = {shared_values[q8[4]], shared_values[q8[5]], shared_values[q8[6]], shared_values[q8[7]]};
|
|
|
|
acc1 += yl[0] * qf1;
|
|
|
|
acc2 += yl[1] * qf2;
|
|
|
|
|
|
|
|
aux32[0] = q4[2] | (q4[3] << 16);
|
|
|
|
aux32[1] = (aux32[0] >> 4) & 0x0f0f0f0f;
|
|
|
|
aux32[0] &= 0x0f0f0f0f;
|
|
|
|
qf1 = {shared_values[q8[0]], shared_values[q8[1]], shared_values[q8[2]], shared_values[q8[3]]};
|
|
|
|
qf2 = {shared_values[q8[4]], shared_values[q8[5]], shared_values[q8[6]], shared_values[q8[7]]};
|
|
|
|
acc1 += yl[2] * qf1;
|
|
|
|
acc2 += yl[3] * qf2;
|
|
|
|
|
|
|
|
acc1 += acc2;
|
|
|
|
|
|
|
|
sumf[row] += (float)xb.d * (acc1[0] + acc1[1] + acc1[2] + acc1[3]);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
yb += 16 * QK4_NL;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < 2; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-02-28 09:37:02 +01:00
|
|
|
#if QK_K != 64
|
2024-02-27 15:34:24 +01:00
|
|
|
void kernel_mul_mv_iq4_xs_f32_impl(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values_i8,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
threadgroup float * shared_values = (threadgroup float *)shared_values_i8;
|
2024-02-27 15:34:24 +01:00
|
|
|
const int nb = ne00/QK_K;
|
|
|
|
const int r0 = tgpig.x;
|
|
|
|
const int r1 = tgpig.y;
|
|
|
|
const int im = tgpig.z;
|
|
|
|
const int first_row = (r0 * 2 + sgitg) * 2;
|
|
|
|
const int ib_row = first_row * nb;
|
|
|
|
|
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
const uint offset0 = (i12/r2)*(nb*ne01) + (i13/r3)*(nb*ne01*ne02);
|
|
|
|
device const block_iq4_xs * x = (device const block_iq4_xs *) src0 + ib_row + offset0;
|
|
|
|
device const float * y = (device const float *) src1 + r1*ne10 + im*ne00*ne1;
|
|
|
|
|
|
|
|
const int ix = tiisg/16; // 0 or 1
|
|
|
|
const int it = tiisg%16; // 0...15
|
|
|
|
const int ib = it/2;
|
|
|
|
const int il = it%2;
|
|
|
|
|
|
|
|
shared_values[tiisg] = kvalues_iq4nl_f[tiisg%16];
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
float4 yl[4];
|
|
|
|
float sumf[2]={0.f}, all_sum;
|
|
|
|
|
|
|
|
device const float * yb = y + ix * QK_K + ib * 32 + il * 8;
|
|
|
|
|
|
|
|
uint32_t aux32[2];
|
|
|
|
thread const uint8_t * q8 = (thread const uint8_t *)aux32;
|
|
|
|
|
|
|
|
float4 qf1, qf2;
|
|
|
|
|
|
|
|
for (int ibl = ix; ibl < nb; ibl += 2) {
|
|
|
|
|
|
|
|
device const float4 * y4 = (device const float4 *)yb;
|
|
|
|
yl[0] = y4[0]; yl[1] = y4[4]; yl[2] = y4[1]; yl[3] = y4[5];
|
|
|
|
|
|
|
|
for (int row = 0; row < 2; ++row) {
|
|
|
|
|
|
|
|
device const block_iq4_xs & xb = x[row*nb + ibl];
|
|
|
|
device const uint32_t * q4 = (device const uint32_t *)(xb.qs + 16*ib + 8*il);
|
|
|
|
|
|
|
|
float4 acc1 = {0.f}, acc2 = {0.f};
|
|
|
|
|
|
|
|
aux32[0] = q4[0] & 0x0f0f0f0f;
|
|
|
|
aux32[1] = (q4[0] >> 4) & 0x0f0f0f0f;
|
|
|
|
qf1 = {shared_values[q8[0]], shared_values[q8[1]], shared_values[q8[2]], shared_values[q8[3]]};
|
|
|
|
qf2 = {shared_values[q8[4]], shared_values[q8[5]], shared_values[q8[6]], shared_values[q8[7]]};
|
|
|
|
acc1 += yl[0] * qf1;
|
|
|
|
acc2 += yl[1] * qf2;
|
|
|
|
|
|
|
|
aux32[0] = q4[1] & 0x0f0f0f0f;
|
|
|
|
aux32[1] = (q4[1] >> 4) & 0x0f0f0f0f;
|
|
|
|
qf1 = {shared_values[q8[0]], shared_values[q8[1]], shared_values[q8[2]], shared_values[q8[3]]};
|
|
|
|
qf2 = {shared_values[q8[4]], shared_values[q8[5]], shared_values[q8[6]], shared_values[q8[7]]};
|
|
|
|
acc1 += yl[2] * qf1;
|
|
|
|
acc2 += yl[3] * qf2;
|
|
|
|
|
|
|
|
acc1 += acc2;
|
|
|
|
|
|
|
|
const int ls = (((xb.scales_l[ib/2] >> 4*(ib%2)) & 0xf) | (((xb.scales_h >> 2*ib) & 3) << 4)) - 32;
|
|
|
|
sumf[row] += (float)xb.d * ls * (acc1[0] + acc1[1] + acc1[2] + acc1[3]);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
yb += 2 * QK_K;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (int row = 0; row < 2; ++row) {
|
|
|
|
all_sum = simd_sum(sumf[row]);
|
|
|
|
if (tiisg == 0) {
|
|
|
|
dst[r1*ne0 + im*ne0*ne1 + first_row + row] = all_sum;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2024-02-28 09:37:02 +01:00
|
|
|
#endif
|
2024-02-27 15:34:24 +01:00
|
|
|
|
2024-02-18 17:16:55 +01:00
|
|
|
[[host_name("kernel_mul_mv_iq1_s_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq1_s_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
kernel_mul_mv_iq1_s_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg);
|
2024-02-18 17:16:55 +01:00
|
|
|
}
|
|
|
|
|
2024-03-27 17:55:10 +01:00
|
|
|
[[host_name("kernel_mul_mv_iq1_m_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq1_m_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
kernel_mul_mv_iq1_m_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, nullptr, tgpig, tiisg, sgitg);
|
2024-03-27 17:55:10 +01:00
|
|
|
}
|
|
|
|
|
2024-02-21 15:19:39 +01:00
|
|
|
[[host_name("kernel_mul_mv_iq4_nl_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq4_nl_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2024-04-12 18:13:20 +02:00
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
2024-02-21 15:19:39 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
|
|
|
kernel_mul_mv_iq4_nl_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
|
|
|
}
|
2024-01-30 14:14:12 +01:00
|
|
|
|
2024-02-27 15:34:24 +01:00
|
|
|
[[host_name("kernel_mul_mv_iq4_xs_f32")]]
|
|
|
|
kernel void kernel_mul_mv_iq4_xs_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
2024-04-12 18:13:20 +02:00
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
2024-02-27 15:34:24 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
2024-02-28 09:37:02 +01:00
|
|
|
#if QK_K == 64
|
|
|
|
kernel_mul_mv_iq4_nl_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
|
|
|
#else
|
2024-02-27 15:34:24 +01:00
|
|
|
kernel_mul_mv_iq4_xs_f32_impl(src0, src1, dst, ne00, ne01, ne02, ne10, ne12, ne0, ne1, r2, r3, shared_values, tgpig, tiisg, sgitg);
|
2024-02-28 09:37:02 +01:00
|
|
|
#endif
|
2024-02-27 15:34:24 +01:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
//============================= templates and their specializations =============================
|
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
// NOTE: this is not dequantizing - we are simply fitting the template
|
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_f32(device const float4x4 * src, short il, thread type4x4 & reg) {
|
|
|
|
float4x4 temp = *(((device float4x4 *)src));
|
|
|
|
for (int i = 0; i < 16; i++){
|
|
|
|
reg[i/4][i%4] = temp[i/4][i%4];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_f16(device const half4x4 * src, short il, thread type4x4 & reg) {
|
|
|
|
half4x4 temp = *(((device half4x4 *)src));
|
|
|
|
for (int i = 0; i < 16; i++){
|
|
|
|
reg[i/4][i%4] = temp[i/4][i%4];
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q4_0(device const block_q4_0 *xb, short il, thread type4x4 & reg) {
|
|
|
|
device const uint16_t * qs = ((device const uint16_t *)xb + 1);
|
2023-09-15 11:18:18 +02:00
|
|
|
const float d1 = il ? (xb->d / 16.h) : xb->d;
|
|
|
|
const float d2 = d1 / 256.f;
|
|
|
|
const float md = -8.h * xb->d;
|
2023-09-05 12:54:40 +02:00
|
|
|
const ushort mask0 = il ? 0x00F0 : 0x000F;
|
2023-09-15 11:18:18 +02:00
|
|
|
const ushort mask1 = mask0 << 8;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
for (int i=0;i<8;i++) {
|
2023-09-15 11:18:18 +02:00
|
|
|
reg[i/2][2*(i%2)+0] = d1 * (qs[i] & mask0) + md;
|
|
|
|
reg[i/2][2*(i%2)+1] = d2 * (qs[i] & mask1) + md;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q4_1(device const block_q4_1 *xb, short il, thread type4x4 & reg) {
|
|
|
|
device const uint16_t * qs = ((device const uint16_t *)xb + 2);
|
2023-09-15 11:18:18 +02:00
|
|
|
const float d1 = il ? (xb->d / 16.h) : xb->d;
|
|
|
|
const float d2 = d1 / 256.f;
|
|
|
|
const float m = xb->m;
|
2023-09-05 12:54:40 +02:00
|
|
|
const ushort mask0 = il ? 0x00F0 : 0x000F;
|
2023-09-15 11:18:18 +02:00
|
|
|
const ushort mask1 = mask0 << 8;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
for (int i=0;i<8;i++) {
|
2023-09-15 11:18:18 +02:00
|
|
|
reg[i/2][2*(i%2)+0] = ((qs[i] & mask0) * d1) + m;
|
|
|
|
reg[i/2][2*(i%2)+1] = ((qs[i] & mask1) * d2) + m;
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-11-03 20:35:05 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q5_0(device const block_q5_0 *xb, short il, thread type4x4 & reg) {
|
|
|
|
device const uint16_t * qs = ((device const uint16_t *)xb + 3);
|
|
|
|
const float d = xb->d;
|
|
|
|
const float md = -16.h * xb->d;
|
|
|
|
const ushort mask = il ? 0x00F0 : 0x000F;
|
|
|
|
|
|
|
|
const uint32_t qh = *((device const uint32_t *)xb->qh);
|
|
|
|
|
|
|
|
const int x_mv = il ? 4 : 0;
|
|
|
|
|
|
|
|
const int gh_mv = il ? 12 : 0;
|
|
|
|
const int gh_bk = il ? 0 : 4;
|
|
|
|
|
|
|
|
for (int i = 0; i < 8; i++) {
|
|
|
|
// extract the 5-th bits for x0 and x1
|
|
|
|
const uint8_t xh_0 = ((qh >> (gh_mv + 2*i )) << gh_bk) & 0x10;
|
|
|
|
const uint8_t xh_1 = ((qh >> (gh_mv + 2*i+1)) << gh_bk) & 0x10;
|
|
|
|
|
|
|
|
// combine the 4-bits from qs with the 5th bit
|
|
|
|
const int32_t x0 = ((((qs[i] ) & mask) >> x_mv) | xh_0);
|
|
|
|
const int32_t x1 = ((((qs[i] >> 8) & mask) >> x_mv) | xh_1);
|
|
|
|
|
|
|
|
reg[i/2][2*(i%2)+0] = d * x0 + md;
|
|
|
|
reg[i/2][2*(i%2)+1] = d * x1 + md;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q5_1(device const block_q5_1 *xb, short il, thread type4x4 & reg) {
|
|
|
|
device const uint16_t * qs = ((device const uint16_t *)xb + 4);
|
|
|
|
const float d = xb->d;
|
|
|
|
const float m = xb->m;
|
|
|
|
const ushort mask = il ? 0x00F0 : 0x000F;
|
|
|
|
|
|
|
|
const uint32_t qh = *((device const uint32_t *)xb->qh);
|
|
|
|
|
|
|
|
const int x_mv = il ? 4 : 0;
|
|
|
|
|
|
|
|
const int gh_mv = il ? 12 : 0;
|
|
|
|
const int gh_bk = il ? 0 : 4;
|
|
|
|
|
|
|
|
for (int i = 0; i < 8; i++) {
|
|
|
|
// extract the 5-th bits for x0 and x1
|
|
|
|
const uint8_t xh_0 = ((qh >> (gh_mv + 2*i )) << gh_bk) & 0x10;
|
|
|
|
const uint8_t xh_1 = ((qh >> (gh_mv + 2*i+1)) << gh_bk) & 0x10;
|
|
|
|
|
|
|
|
// combine the 4-bits from qs with the 5th bit
|
|
|
|
const int32_t x0 = ((((qs[i] ) & mask) >> x_mv) | xh_0);
|
|
|
|
const int32_t x1 = ((((qs[i] >> 8) & mask) >> x_mv) | xh_1);
|
|
|
|
|
|
|
|
reg[i/2][2*(i%2)+0] = d * x0 + m;
|
|
|
|
reg[i/2][2*(i%2)+1] = d * x1 + m;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q8_0(device const block_q8_0 *xb, short il, thread type4x4 & reg) {
|
|
|
|
device const int8_t * qs = ((device const int8_t *)xb->qs);
|
|
|
|
const half d = xb->d;
|
|
|
|
|
2024-01-02 09:57:44 +01:00
|
|
|
for (int i = 0; i < 16; i++) {
|
2023-09-05 12:54:40 +02:00
|
|
|
reg[i/4][i%4] = (qs[i + 16*il] * d);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q2_K(device const block_q2_K *xb, short il, thread type4x4 & reg) {
|
2023-12-13 20:55:03 +01:00
|
|
|
const float d = xb->d;
|
|
|
|
const float min = xb->dmin;
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint8_t * q = (device const uint8_t *)xb->qs;
|
2023-12-13 20:55:03 +01:00
|
|
|
float dl, ml;
|
2023-09-05 12:54:40 +02:00
|
|
|
uint8_t sc = xb->scales[il];
|
|
|
|
|
|
|
|
#if QK_K == 256
|
|
|
|
q = q + 32*(il/8) + 16*(il&1);
|
|
|
|
il = (il/2)%4;
|
|
|
|
#endif
|
|
|
|
half coef = il>1 ? (il>2 ? 1/64.h : 1/16.h) : (il>0 ? 1/4.h : 1.h);
|
|
|
|
uchar mask = il>1 ? (il>2 ? 192 : 48) : (il>0 ? 12 : 3);
|
|
|
|
dl = d * (sc & 0xF) * coef, ml = min * (sc >> 4);
|
|
|
|
for (int i = 0; i < 16; ++i) {
|
|
|
|
reg[i/4][i%4] = dl * (q[i] & mask) - ml;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q3_K(device const block_q3_K *xb, short il, thread type4x4 & reg) {
|
2023-09-15 11:18:18 +02:00
|
|
|
const half d_all = xb->d;
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint8_t * q = (device const uint8_t *)xb->qs;
|
|
|
|
device const uint8_t * h = (device const uint8_t *)xb->hmask;
|
|
|
|
device const int8_t * scales = (device const int8_t *)xb->scales;
|
|
|
|
|
|
|
|
#if QK_K == 256
|
|
|
|
q = q + 32 * (il/8) + 16 * (il&1);
|
|
|
|
h = h + 16 * (il&1);
|
|
|
|
uint8_t m = 1 << (il/2);
|
|
|
|
uint16_t kmask1 = (il/4)>1 ? ((il/4)>2 ? 192 : 48) : \
|
|
|
|
((il/4)>0 ? 12 : 3);
|
|
|
|
uint16_t kmask2 = il/8 ? 0xF0 : 0x0F;
|
|
|
|
uint16_t scale_2 = scales[il%8], scale_1 = scales[8 + il%4];
|
2023-09-15 11:18:18 +02:00
|
|
|
int16_t dl_int = (il/4)&1 ? (scale_2&kmask2) | ((scale_1&kmask1) << 2)
|
|
|
|
: (scale_2&kmask2) | ((scale_1&kmask1) << 4);
|
2024-01-09 18:37:08 +01:00
|
|
|
float dl = il<8 ? d_all * (dl_int - 32.f) : d_all * (dl_int / 16.f - 32.f);
|
|
|
|
const float ml = 4.f * dl;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
il = (il/2) & 3;
|
|
|
|
const half coef = il>1 ? (il>2 ? 1/64.h : 1/16.h) : (il>0 ? 1/4.h : 1.h);
|
|
|
|
const uint8_t mask = il>1 ? (il>2 ? 192 : 48) : (il>0 ? 12 : 3);
|
|
|
|
dl *= coef;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
for (int i = 0; i < 16; ++i) {
|
2023-09-15 11:18:18 +02:00
|
|
|
reg[i/4][i%4] = dl * (q[i] & mask) - (h[i] & m ? 0 : ml);
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
#else
|
|
|
|
float kcoef = il&1 ? 1.f/16.f : 1.f;
|
|
|
|
uint16_t kmask = il&1 ? 0xF0 : 0x0F;
|
|
|
|
float dl = d_all * ((scales[il/2] & kmask) * kcoef - 8);
|
|
|
|
float coef = il>1 ? (il>2 ? 1/64.h : 1/16.h) : (il>0 ? 1/4.h : 1.h);
|
|
|
|
uint8_t mask = il>1 ? (il>2 ? 192 : 48) : (il>0 ? 12 : 3);
|
|
|
|
uint8_t m = 1<<(il*2);
|
|
|
|
for (int i = 0; i < 16; ++i) {
|
|
|
|
reg[i/4][i%4] = coef * dl * ((q[i] & mask) - ((h[i%8] & (m * (1 + i/8))) ? 0 : 4.f/coef));
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
static inline uchar2 get_scale_min_k4_just2(int j, int k, device const uchar * q) {
|
|
|
|
return j < 4 ? uchar2{uchar(q[j+0+k] & 63), uchar(q[j+4+k] & 63)}
|
|
|
|
: uchar2{uchar((q[j+4+k] & 0xF) | ((q[j-4+k] & 0xc0) >> 2)), uchar((q[j+4+k] >> 4) | ((q[j-0+k] & 0xc0) >> 2))};
|
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q4_K(device const block_q4_K *xb, short il, thread type4x4 & reg) {
|
2023-09-15 11:18:18 +02:00
|
|
|
device const uchar * q = xb->qs;
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
#if QK_K == 256
|
|
|
|
short is = (il/4) * 2;
|
|
|
|
q = q + (il/4) * 32 + 16 * (il&1);
|
2023-09-15 11:18:18 +02:00
|
|
|
il = il & 3;
|
|
|
|
const uchar2 sc = get_scale_min_k4_just2(is, il/2, xb->scales);
|
2023-12-13 20:55:03 +01:00
|
|
|
const float d = il < 2 ? xb->d : xb->d / 16.h;
|
|
|
|
const float min = xb->dmin;
|
|
|
|
const float dl = d * sc[0];
|
|
|
|
const float ml = min * sc[1];
|
2023-09-05 12:54:40 +02:00
|
|
|
#else
|
2024-02-18 20:39:58 +01:00
|
|
|
(void) get_scale_min_k4_just2;
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
q = q + 16 * (il&1);
|
|
|
|
device const uint8_t * s = xb->scales;
|
|
|
|
device const half2 * dh = (device const half2 *)xb->d;
|
|
|
|
const float2 d = (float2)dh[0];
|
|
|
|
const float dl = il<2 ? d[0] * (s[0]&0xF) : d[0] * (s[1]&0xF)/16.h;
|
2023-09-15 11:18:18 +02:00
|
|
|
const float ml = il<2 ? d[1] * (s[0]>>4) : d[1] * (s[1]>>4);
|
2023-09-05 12:54:40 +02:00
|
|
|
#endif
|
|
|
|
const ushort mask = il<2 ? 0x0F : 0xF0;
|
|
|
|
for (int i = 0; i < 16; ++i) {
|
|
|
|
reg[i/4][i%4] = dl * (q[i] & mask) - ml;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q5_K(device const block_q5_K *xb, short il, thread type4x4 & reg) {
|
|
|
|
device const uint8_t * q = xb->qs;
|
|
|
|
device const uint8_t * qh = xb->qh;
|
|
|
|
|
|
|
|
#if QK_K == 256
|
|
|
|
short is = (il/4) * 2;
|
|
|
|
q = q + 32 * (il/4) + 16 * (il&1);
|
|
|
|
qh = qh + 16 * (il&1);
|
|
|
|
uint8_t ul = 1 << (il/2);
|
2023-09-15 11:18:18 +02:00
|
|
|
il = il & 3;
|
|
|
|
const uchar2 sc = get_scale_min_k4_just2(is, il/2, xb->scales);
|
2024-01-09 18:37:08 +01:00
|
|
|
const float d = il < 2 ? xb->d : xb->d / 16.f;
|
2023-12-13 20:55:03 +01:00
|
|
|
const float min = xb->dmin;
|
|
|
|
const float dl = d * sc[0];
|
|
|
|
const float ml = min * sc[1];
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
const ushort mask = il<2 ? 0x0F : 0xF0;
|
|
|
|
const float qh_val = il<2 ? 16.f : 256.f;
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = 0; i < 16; ++i) {
|
|
|
|
reg[i/4][i%4] = dl * ((q[i] & mask) + (qh[i] & ul ? qh_val : 0)) - ml;
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
q = q + 16 * (il&1);
|
|
|
|
device const int8_t * s = xb->scales;
|
|
|
|
const float dl = xb->d * s[il];
|
|
|
|
uint8_t m = 1<<(il*2);
|
|
|
|
const float coef = il<2 ? 1.f : 1.f/16.f;
|
|
|
|
const ushort mask = il<2 ? 0x0F : 0xF0;
|
|
|
|
for (int i = 0; i < 16; ++i) {
|
|
|
|
reg[i/4][i%4] = coef * dl * ((q[i] & mask) - (qh[i%8] & (m*(1+i/8)) ? 0.f : 16.f/coef));
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_q6_K(device const block_q6_K *xb, short il, thread type4x4 & reg) {
|
2023-09-15 11:18:18 +02:00
|
|
|
const half d_all = xb->d;
|
2023-09-05 12:54:40 +02:00
|
|
|
device const uint8_t * ql = (device const uint8_t *)xb->ql;
|
|
|
|
device const uint8_t * qh = (device const uint8_t *)xb->qh;
|
|
|
|
device const int8_t * scales = (device const int8_t *)xb->scales;
|
|
|
|
|
|
|
|
#if QK_K == 256
|
|
|
|
ql = ql + 64*(il/8) + 32*((il/2)&1) + 16*(il&1);
|
|
|
|
qh = qh + 32*(il/8) + 16*(il&1);
|
2024-01-09 18:37:08 +01:00
|
|
|
float sc = scales[(il%2) + 2 * ((il/2))];
|
2023-09-15 11:18:18 +02:00
|
|
|
il = (il/2) & 3;
|
2023-09-05 12:54:40 +02:00
|
|
|
#else
|
|
|
|
ql = ql + 16 * (il&1);
|
2024-01-09 18:37:08 +01:00
|
|
|
float sc = scales[il];
|
2023-09-05 12:54:40 +02:00
|
|
|
#endif
|
2023-09-15 11:18:18 +02:00
|
|
|
const uint16_t kmask1 = il>1 ? (il>2 ? 192 : 48) : (il>0 ? 12 : 3);
|
|
|
|
const uint16_t kmask2 = il>1 ? 0xF0 : 0x0F;
|
2024-01-09 18:37:08 +01:00
|
|
|
const float coef = il>1 ? 1.f/16.f : 1.f;
|
|
|
|
const float ml = d_all * sc * 32.f;
|
|
|
|
const float dl = d_all * sc * coef;
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = 0; i < 16; ++i) {
|
2023-09-15 11:18:18 +02:00
|
|
|
const half q = il&1 ? ((ql[i] & kmask2) | ((qh[i] & kmask1) << 2))
|
|
|
|
: ((ql[i] & kmask2) | ((qh[i] & kmask1) << 4));
|
|
|
|
reg[i/4][i%4] = dl * q - ml;
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
|
|
|
|
2024-01-08 16:02:32 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq2_xxs(device const block_iq2_xxs * xb, short il, thread type4x4 & reg) {
|
|
|
|
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
|
|
|
|
const float d = xb->d;
|
|
|
|
const int ib32 = il/2;
|
|
|
|
il = il%2;
|
|
|
|
// il = 0 or 1. il = 0 processes the first 16 quants in a block of 32, il = 1 the second 16
|
|
|
|
// each block of 32 needs 2 uint32_t's for the quants & scale, so 4 uint16_t's.
|
|
|
|
device const uint16_t * q2 = xb->qs + 4*ib32;
|
|
|
|
const uint32_t aux32_g = q2[0] | (q2[1] << 16);
|
|
|
|
const uint32_t aux32_s = q2[2] | (q2[3] << 16);
|
|
|
|
thread const uint8_t * aux8 = (thread const uint8_t *)&aux32_g;
|
|
|
|
const float dl = d * (0.5f + (aux32_s >> 28)) * 0.25f;
|
2024-01-11 20:39:39 +01:00
|
|
|
constant uint8_t * grid = (constant uint8_t *)(iq2xxs_grid + aux8[2*il+0]);
|
2024-01-08 16:02:32 +01:00
|
|
|
uint8_t signs = ksigns_iq2xs[(aux32_s >> 14*il) & 127];
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
reg[i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f);
|
|
|
|
}
|
2024-01-11 20:39:39 +01:00
|
|
|
grid = (constant uint8_t *)(iq2xxs_grid + aux8[2*il+1]);
|
2024-01-08 16:02:32 +01:00
|
|
|
signs = ksigns_iq2xs[(aux32_s >> (14*il+7)) & 127];
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
reg[2+i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-01-11 20:39:39 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq2_xs(device const block_iq2_xs * xb, short il, thread type4x4 & reg) {
|
|
|
|
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
|
|
|
|
const float d = xb->d;
|
|
|
|
const int ib32 = il/2;
|
|
|
|
il = il%2;
|
|
|
|
// il = 0 or 1. il = 0 processes the first 16 quants in a block of 32, il = 1 the second 16
|
|
|
|
device const uint16_t * q2 = xb->qs + 4*ib32;
|
|
|
|
const float dl = d * (0.5f + ((xb->scales[ib32] >> 4*il) & 0xf)) * 0.25f;
|
|
|
|
constant uint8_t * grid = (constant uint8_t *)(iq2xs_grid + (q2[2*il+0] & 511));
|
|
|
|
uint8_t signs = ksigns_iq2xs[q2[2*il+0] >> 9];
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
reg[i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
grid = (constant uint8_t *)(iq2xs_grid + (q2[2*il+1] & 511));
|
|
|
|
signs = ksigns_iq2xs[q2[2*il+1] >> 9];
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
reg[2+i/4][i%4] = dl * grid[i] * (signs & kmask_iq2xs[i] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-01-30 14:14:12 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq3_xxs(device const block_iq3_xxs * xb, short il, thread type4x4 & reg) {
|
|
|
|
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
|
|
|
|
const float d = xb->d;
|
|
|
|
const int ib32 = il/2;
|
|
|
|
il = il%2;
|
|
|
|
// il = 0 or 1. il = 0 processes the first 16 quants in a block of 32, il = 1 the second 16
|
|
|
|
device const uint8_t * q3 = xb->qs + 8*ib32;
|
|
|
|
device const uint16_t * gas = (device const uint16_t *)(xb->qs + QK_K/4) + 2*ib32;
|
|
|
|
const uint32_t aux32 = gas[0] | (gas[1] << 16);
|
|
|
|
const float dl = d * (0.5f + (aux32 >> 28)) * 0.5f;
|
|
|
|
constant uint8_t * grid1 = (constant uint8_t *)(iq3xxs_grid + q3[4*il+0]);
|
|
|
|
constant uint8_t * grid2 = (constant uint8_t *)(iq3xxs_grid + q3[4*il+1]);
|
|
|
|
uint8_t signs = ksigns_iq2xs[(aux32 >> 14*il) & 127];
|
|
|
|
for (int i = 0; i < 4; ++i) {
|
|
|
|
reg[0][i] = dl * grid1[i] * (signs & kmask_iq2xs[i+0] ? -1.f : 1.f);
|
|
|
|
reg[1][i] = dl * grid2[i] * (signs & kmask_iq2xs[i+4] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
grid1 = (constant uint8_t *)(iq3xxs_grid + q3[4*il+2]);
|
|
|
|
grid2 = (constant uint8_t *)(iq3xxs_grid + q3[4*il+3]);
|
|
|
|
signs = ksigns_iq2xs[(aux32 >> (14*il+7)) & 127];
|
|
|
|
for (int i = 0; i < 4; ++i) {
|
|
|
|
reg[2][i] = dl * grid1[i] * (signs & kmask_iq2xs[i+0] ? -1.f : 1.f);
|
|
|
|
reg[3][i] = dl * grid2[i] * (signs & kmask_iq2xs[i+4] ? -1.f : 1.f);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-02-24 15:23:52 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq3_s(device const block_iq3_s * xb, short il, thread type4x4 & reg) {
|
|
|
|
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
|
|
|
|
const float d = xb->d;
|
|
|
|
const int ib32 = il/2;
|
|
|
|
il = il%2;
|
|
|
|
// il = 0 or 1. il = 0 processes the first 16 quants in a block of 32, il = 1 the second 16
|
|
|
|
device const uint8_t * qs = xb->qs + 8*ib32;
|
|
|
|
device const uint8_t * signs = xb->signs + 4*ib32 + 2*il;
|
|
|
|
const uint8_t qh = xb->qh[ib32] >> 4*il;
|
2024-03-02 16:00:51 +01:00
|
|
|
const float dl = d * (1 + 2*((xb->scales[ib32/2] >> 4*(ib32%2)) & 0xf));
|
|
|
|
constant uint8_t * grid1 = (constant uint8_t *)(iq3s_grid + (qs[4*il+0] | ((qh << 8) & 256)));
|
|
|
|
constant uint8_t * grid2 = (constant uint8_t *)(iq3s_grid + (qs[4*il+1] | ((qh << 7) & 256)));
|
2024-02-24 15:23:52 +01:00
|
|
|
for (int i = 0; i < 4; ++i) {
|
|
|
|
reg[0][i] = dl * grid1[i] * select(1, -1, signs[0] & kmask_iq2xs[i+0]);
|
|
|
|
reg[1][i] = dl * grid2[i] * select(1, -1, signs[0] & kmask_iq2xs[i+4]);
|
|
|
|
}
|
2024-03-02 16:00:51 +01:00
|
|
|
grid1 = (constant uint8_t *)(iq3s_grid + (qs[4*il+2] | ((qh << 6) & 256)));
|
|
|
|
grid2 = (constant uint8_t *)(iq3s_grid + (qs[4*il+3] | ((qh << 5) & 256)));
|
2024-02-24 15:23:52 +01:00
|
|
|
for (int i = 0; i < 4; ++i) {
|
|
|
|
reg[2][i] = dl * grid1[i] * select(1, -1, signs[1] & kmask_iq2xs[i+0]);
|
|
|
|
reg[3][i] = dl * grid2[i] * select(1, -1, signs[1] & kmask_iq2xs[i+4]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-02-26 17:28:38 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq2_s(device const block_iq2_s * xb, short il, thread type4x4 & reg) {
|
|
|
|
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
|
|
|
|
const float d = xb->d;
|
|
|
|
const int ib32 = il/2;
|
|
|
|
il = il%2;
|
|
|
|
// il = 0 or 1. il = 0 processes the first 16 quants in a block of 32, il = 1 the second 16
|
|
|
|
device const uint8_t * qs = xb->qs + 4*ib32 + 2*il;
|
|
|
|
device const uint8_t * signs = qs + QK_K/8;
|
|
|
|
const uint8_t qh = xb->qh[ib32] >> 4*il;
|
|
|
|
const float dl = d * (0.5f + ((xb->scales[ib32] >> 4*il) & 0xf)) * 0.25f;
|
|
|
|
constant uint8_t * grid1 = (constant uint8_t *)(iq2s_grid + (qs[0] | ((qh << 8) & 0x300)));
|
|
|
|
constant uint8_t * grid2 = (constant uint8_t *)(iq2s_grid + (qs[1] | ((qh << 6) & 0x300)));
|
|
|
|
for (int i = 0; i < 8; ++i) {
|
|
|
|
reg[i/4+0][i%4] = dl * grid1[i] * select(1, -1, signs[0] & kmask_iq2xs[i]);
|
|
|
|
reg[i/4+2][i%4] = dl * grid2[i] * select(1, -1, signs[1] & kmask_iq2xs[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-02-18 17:16:55 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq1_s(device const block_iq1_s * xb, short il, thread type4x4 & reg) {
|
|
|
|
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
|
2024-03-11 07:51:49 +01:00
|
|
|
const int ib32 = il/2;
|
|
|
|
il = il%2;
|
2024-02-18 17:16:55 +01:00
|
|
|
const float d = xb->d;
|
2024-03-11 07:51:49 +01:00
|
|
|
device const uint8_t * qs = xb->qs + 4*ib32 + 2*il;
|
|
|
|
device const uint16_t * qh = xb->qh;
|
2024-03-11 16:53:15 +01:00
|
|
|
const float dl = d * (2*((qh[ib32] >> 12) & 7) + 1);
|
|
|
|
const float ml = dl * (qh[ib32] & 0x8000 ? -1 - IQ1S_DELTA : -1 + IQ1S_DELTA);
|
|
|
|
const uint16_t h = qh[ib32] >> 6*il;
|
|
|
|
constant uint8_t * grid1 = (constant uint8_t *)(iq1s_grid_gpu + (qs[0] | ((h << 8) & 0x700)));
|
|
|
|
constant uint8_t * grid2 = (constant uint8_t *)(iq1s_grid_gpu + (qs[1] | ((h << 5) & 0x700)));
|
2024-03-11 07:51:49 +01:00
|
|
|
for (int i = 0; i < 4; ++i) {
|
2024-03-11 16:53:15 +01:00
|
|
|
reg[0][i] = dl * (grid1[i] & 0xf) + ml;
|
|
|
|
reg[1][i] = dl * (grid1[i] >> 4) + ml;
|
|
|
|
reg[2][i] = dl * (grid2[i] & 0xf) + ml;
|
|
|
|
reg[3][i] = dl * (grid2[i] >> 4) + ml;
|
2024-02-18 17:16:55 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-03-27 17:55:10 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq1_m(device const block_iq1_m * xb, short il, thread type4x4 & reg) {
|
|
|
|
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
|
|
|
|
const int ib32 = il/2;
|
|
|
|
il = il%2;
|
|
|
|
device const uint16_t * sc = (device const uint16_t *)xb->scales;
|
|
|
|
#if QK_K == 64
|
|
|
|
const float d = xb->d;
|
|
|
|
#else
|
|
|
|
iq1m_scale_t scale;
|
|
|
|
scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000);
|
|
|
|
const float d = scale.f16;
|
|
|
|
#endif
|
|
|
|
device const uint8_t * qs = xb->qs + 4*ib32 + 2*il;
|
|
|
|
device const uint8_t * qh = xb->qh + 2*ib32 + il;
|
|
|
|
#if QK_K == 64
|
|
|
|
const float dl = d * (2*((sc[ib32/2] >> (8*(ib32%2)+4*il)) & 0xf) + 1);
|
|
|
|
#else
|
|
|
|
const float dl = d * (2*((sc[ib32/2] >> (6*(ib32%2)+3*il)) & 7) + 1);
|
|
|
|
#endif
|
|
|
|
const float ml1 = dl * (qh[0] & 0x08 ? -1 - IQ1M_DELTA : -1 + IQ1M_DELTA);
|
|
|
|
const float ml2 = dl * (qh[0] & 0x80 ? -1 - IQ1M_DELTA : -1 + IQ1M_DELTA);
|
|
|
|
constant uint8_t * grid1 = (constant uint8_t *)(iq1s_grid_gpu + (qs[0] | ((qh[0] << 8) & 0x700)));
|
|
|
|
constant uint8_t * grid2 = (constant uint8_t *)(iq1s_grid_gpu + (qs[1] | ((qh[0] << 4) & 0x700)));
|
|
|
|
for (int i = 0; i < 4; ++i) {
|
|
|
|
reg[0][i] = dl * (grid1[i] & 0xf) + ml1;
|
|
|
|
reg[1][i] = dl * (grid1[i] >> 4) + ml1;
|
|
|
|
reg[2][i] = dl * (grid2[i] & 0xf) + ml2;
|
|
|
|
reg[3][i] = dl * (grid2[i] >> 4) + ml2;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-02-21 15:19:39 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq4_nl(device const block_iq4_nl * xb, short il, thread type4x4 & reg) {
|
|
|
|
device const uint16_t * q4 = (device const uint16_t *)xb->qs;
|
|
|
|
const float d = xb->d;
|
|
|
|
uint32_t aux32;
|
|
|
|
thread const uint8_t * q8 = (thread const uint8_t *)&aux32;
|
|
|
|
for (int i = 0; i < 4; ++i) {
|
|
|
|
aux32 = ((q4[2*i] | (q4[2*i+1] << 16)) >> 4*il) & 0x0f0f0f0f;
|
|
|
|
reg[i][0] = d * kvalues_iq4nl_f[q8[0]];
|
|
|
|
reg[i][1] = d * kvalues_iq4nl_f[q8[1]];
|
|
|
|
reg[i][2] = d * kvalues_iq4nl_f[q8[2]];
|
|
|
|
reg[i][3] = d * kvalues_iq4nl_f[q8[3]];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-02-27 15:34:24 +01:00
|
|
|
template <typename type4x4>
|
|
|
|
void dequantize_iq4_xs(device const block_iq4_xs * xb, short il, thread type4x4 & reg) {
|
2024-02-28 09:37:02 +01:00
|
|
|
#if QK_K == 64
|
|
|
|
dequantize_iq4_nl(xb, il, reg);
|
|
|
|
#else
|
2024-02-27 15:34:24 +01:00
|
|
|
// il is 0...15 for QK_K = 256 => index of block of 32 is il/2
|
|
|
|
const int ib32 = il/2;
|
|
|
|
il = il%2;
|
|
|
|
// il = 0 or 1. il = 0 processes the first 16 quants in a block of 32, il = 1 the second 16
|
|
|
|
device const uint32_t * q4 = (device const uint32_t *)xb->qs + 4*ib32;
|
|
|
|
const int ls = ((xb->scales_l[ib32/2] >> 4*(ib32%2)) & 0xf) | (((xb->scales_h >> 2*ib32) & 3) << 4);
|
|
|
|
const float d = (float)xb->d * (ls - 32);
|
|
|
|
uint32_t aux32;
|
|
|
|
thread const uint8_t * q8 = (thread const uint8_t *)&aux32;
|
|
|
|
for (int i = 0; i < 4; ++i) {
|
|
|
|
aux32 = (q4[i] >> 4*il) & 0x0f0f0f0f;
|
|
|
|
reg[i][0] = d * kvalues_iq4nl_f[q8[0]];
|
|
|
|
reg[i][1] = d * kvalues_iq4nl_f[q8[1]];
|
|
|
|
reg[i][2] = d * kvalues_iq4nl_f[q8[2]];
|
|
|
|
reg[i][3] = d * kvalues_iq4nl_f[q8[3]];
|
|
|
|
}
|
2024-02-28 09:37:02 +01:00
|
|
|
#endif
|
2024-02-27 15:34:24 +01:00
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
template<typename block_q, short nl, void (*dequantize_func)(device const block_q *, short, thread float4x4 &)>
|
|
|
|
kernel void kernel_get_rows(
|
|
|
|
device const void * src0,
|
2023-12-13 20:55:03 +01:00
|
|
|
device const char * src1,
|
2023-09-05 12:54:40 +02:00
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant uint64_t & nb01,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
2023-09-05 12:54:40 +02:00
|
|
|
constant uint64_t & nb1,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant uint64_t & nb2,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
2023-09-05 12:54:40 +02:00
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tptg [[threads_per_threadgroup]]) {
|
|
|
|
//const int64_t i = tgpig;
|
|
|
|
//const int64_t r = ((device int32_t *) src1)[i];
|
|
|
|
|
|
|
|
const int64_t i10 = tgpig.x;
|
|
|
|
const int64_t i11 = tgpig.y;
|
2023-06-25 14:40:30 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
const int64_t r = ((device int32_t *) ((device char *) src1 + i11*nb11 + i10*nb10))[0];
|
|
|
|
|
|
|
|
const int64_t i02 = i11;
|
|
|
|
|
|
|
|
for (int64_t ind = tiitg; ind < ne00/16; ind += tptg.x) {
|
2023-09-05 12:54:40 +02:00
|
|
|
float4x4 temp;
|
|
|
|
dequantize_func(
|
2023-12-13 20:55:03 +01:00
|
|
|
((device const block_q *) ((device char *) src0 + r*nb01 + i02*nb02)) + ind/nl, ind%nl, temp);
|
|
|
|
*(((device float4x4 *) ((device char *) dst + i11*nb2 + i10*nb1)) + ind) = temp;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_get_rows_f32(
|
|
|
|
device const void * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint3 tptg [[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i10 = tgpig.x;
|
|
|
|
const int64_t i11 = tgpig.y;
|
|
|
|
|
|
|
|
const int64_t r = ((device int32_t *) ((device char *) src1 + i11*nb11 + i10*nb10))[0];
|
|
|
|
|
|
|
|
const int64_t i02 = i11;
|
|
|
|
|
|
|
|
for (int ind = tiitg; ind < ne00; ind += tptg.x) {
|
|
|
|
((device float *) ((device char *) dst + i11*nb2 + i10*nb1))[ind] =
|
|
|
|
((device float *) ((device char *) src0 + r*nb01 + i02*nb02))[ind];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel void kernel_get_rows_f16(
|
|
|
|
device const void * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint3 tptg [[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i10 = tgpig.x;
|
|
|
|
const int64_t i11 = tgpig.y;
|
|
|
|
|
|
|
|
const int64_t r = ((device int32_t *) ((device char *) src1 + i11*nb11 + i10*nb10))[0];
|
|
|
|
|
|
|
|
const int64_t i02 = i11;
|
|
|
|
|
|
|
|
for (int ind = tiitg; ind < ne00; ind += tptg.x) {
|
|
|
|
((device float *) ((device char *) dst + i11*nb2 + i10*nb1))[ind] =
|
|
|
|
((device half *) ((device char *) src0 + r*nb01 + i02*nb02))[ind];
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-06-25 14:40:30 +02:00
|
|
|
}
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2024-01-03 10:35:46 +01:00
|
|
|
kernel void kernel_get_rows_i32(
|
|
|
|
device const void * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device int32_t * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb1,
|
|
|
|
constant uint64_t & nb2,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint3 tptg [[threads_per_threadgroup]]) {
|
|
|
|
const int64_t i10 = tgpig.x;
|
|
|
|
const int64_t i11 = tgpig.y;
|
|
|
|
|
|
|
|
const int64_t r = ((device int32_t *) ((device char *) src1 + i11*nb11 + i10*nb10))[0];
|
|
|
|
|
|
|
|
const int64_t i02 = i11;
|
|
|
|
|
|
|
|
for (int ind = tiitg; ind < ne00; ind += tptg.x) {
|
|
|
|
((device int32_t *) ((device char *) dst + i11*nb2 + i10*nb1))[ind] =
|
|
|
|
((device int32_t *) ((device char *) src0 + r*nb01 + i02*nb02))[ind];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
#define BLOCK_SIZE_M 64 // 8 simdgroup matrices from matrix A
|
2023-11-03 20:35:05 +01:00
|
|
|
#define BLOCK_SIZE_N 32 // 4 simdgroup matrices from matrix B
|
2023-09-05 12:54:40 +02:00
|
|
|
#define BLOCK_SIZE_K 32
|
|
|
|
#define THREAD_MAT_M 4 // each thread take 4 simdgroup matrices from matrix A
|
|
|
|
#define THREAD_MAT_N 2 // each thread take 2 simdgroup matrices from matrix B
|
|
|
|
#define THREAD_PER_BLOCK 128
|
|
|
|
#define THREAD_PER_ROW 2 // 2 thread for each row in matrix A to load numbers
|
|
|
|
#define THREAD_PER_COL 4 // 4 thread for each row in matrix B to load numbers
|
|
|
|
#define SG_MAT_SIZE 64 // simdgroup matrix is of shape 8x8
|
|
|
|
#define SG_MAT_ROW 8
|
|
|
|
|
|
|
|
// each block_q contains 16*nl weights
|
|
|
|
template<typename block_q, short nl, void (*dequantize_func)(device const block_q *, short, thread half4x4 &)>
|
2023-12-07 21:27:19 +01:00
|
|
|
void kernel_mul_mm_impl(device const uchar * src0,
|
|
|
|
device const uchar * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne02,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne12,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
threadgroup uchar * shared_memory [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2023-09-15 11:18:18 +02:00
|
|
|
|
|
|
|
threadgroup half * sa = (threadgroup half *)(shared_memory);
|
2023-09-05 12:54:40 +02:00
|
|
|
threadgroup float * sb = (threadgroup float *)(shared_memory + 4096);
|
|
|
|
|
|
|
|
const uint r0 = tgpig.y;
|
|
|
|
const uint r1 = tgpig.x;
|
|
|
|
const uint im = tgpig.z;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
// if this block is of 64x32 shape or smaller
|
|
|
|
short n_rows = (ne0 - r0 * BLOCK_SIZE_M < BLOCK_SIZE_M) ? (ne0 - r0 * BLOCK_SIZE_M) : BLOCK_SIZE_M;
|
|
|
|
short n_cols = (ne1 - r1 * BLOCK_SIZE_N < BLOCK_SIZE_N) ? (ne1 - r1 * BLOCK_SIZE_N) : BLOCK_SIZE_N;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
// a thread shouldn't load data outside of the matrix
|
|
|
|
short thread_row = ((short)tiitg/THREAD_PER_ROW) < n_rows ? ((short)tiitg/THREAD_PER_ROW) : n_rows - 1;
|
|
|
|
short thread_col = ((short)tiitg/THREAD_PER_COL) < n_cols ? ((short)tiitg/THREAD_PER_COL) : n_cols - 1;
|
|
|
|
|
2023-09-15 11:18:18 +02:00
|
|
|
simdgroup_half8x8 ma[4];
|
2023-09-05 12:54:40 +02:00
|
|
|
simdgroup_float8x8 mb[2];
|
|
|
|
simdgroup_float8x8 c_res[8];
|
|
|
|
for (int i = 0; i < 8; i++){
|
|
|
|
c_res[i] = make_filled_simdgroup_matrix<float, 8>(0.f);
|
|
|
|
}
|
|
|
|
|
|
|
|
short il = (tiitg % THREAD_PER_ROW);
|
2023-09-15 11:18:18 +02:00
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
const uint i12 = im%ne12;
|
|
|
|
const uint i13 = im/ne12;
|
|
|
|
|
|
|
|
uint offset0 = (i12/r2)*nb02 + (i13/r3)*(nb02*ne02);
|
2023-09-15 11:18:18 +02:00
|
|
|
ushort offset1 = il/nl;
|
|
|
|
|
|
|
|
device const block_q * x = (device const block_q *)(src0 + (r0 * BLOCK_SIZE_M + thread_row) * nb01 + offset0) + offset1;
|
|
|
|
device const float * y = (device const float *)(src1
|
|
|
|
+ nb12 * im
|
|
|
|
+ nb11 * (r1 * BLOCK_SIZE_N + thread_col)
|
|
|
|
+ nb10 * (BLOCK_SIZE_K / THREAD_PER_COL * (tiitg % THREAD_PER_COL)));
|
2023-09-05 12:54:40 +02:00
|
|
|
|
|
|
|
for (int loop_k = 0; loop_k < ne00; loop_k += BLOCK_SIZE_K) {
|
2023-11-03 20:35:05 +01:00
|
|
|
// load data and store to threadgroup memory
|
2023-09-05 12:54:40 +02:00
|
|
|
half4x4 temp_a;
|
|
|
|
dequantize_func(x, il, temp_a);
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
#pragma unroll(16)
|
|
|
|
for (int i = 0; i < 16; i++) {
|
|
|
|
*(sa + SG_MAT_SIZE * ((tiitg / THREAD_PER_ROW / 8) \
|
2023-11-03 20:35:05 +01:00
|
|
|
+ (tiitg % THREAD_PER_ROW) * 16 + (i / 8) * 8) \
|
|
|
|
+ (tiitg / THREAD_PER_ROW) % 8 + (i & 7) * 8) = temp_a[i/4][i%4];
|
2023-09-05 12:54:40 +02:00
|
|
|
}
|
2023-11-03 20:35:05 +01:00
|
|
|
|
|
|
|
*(threadgroup float2x4 *)(sb + (tiitg % THREAD_PER_COL) * 8 * 32 + 8 * (tiitg / THREAD_PER_COL)) = *((device float2x4 *)y);
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
il = (il + 2 < nl) ? il + 2 : il % 2;
|
|
|
|
x = (il < 2) ? x + (2+nl-1)/nl : x;
|
|
|
|
y += BLOCK_SIZE_K;
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
|
|
|
// load matrices from threadgroup memory and conduct outer products
|
2023-09-05 12:54:40 +02:00
|
|
|
threadgroup half * lsma = (sa + THREAD_MAT_M * SG_MAT_SIZE * (sgitg % 2));
|
|
|
|
threadgroup float * lsmb = (sb + THREAD_MAT_N * SG_MAT_SIZE * (sgitg / 2));
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
#pragma unroll(4)
|
|
|
|
for (int ik = 0; ik < BLOCK_SIZE_K / 8; ik++) {
|
|
|
|
#pragma unroll(4)
|
|
|
|
for (int i = 0; i < 4; i++) {
|
|
|
|
simdgroup_load(ma[i],lsma + SG_MAT_SIZE * i);
|
|
|
|
}
|
|
|
|
simdgroup_barrier(mem_flags::mem_none);
|
|
|
|
#pragma unroll(2)
|
|
|
|
for (int i = 0; i < 2; i++) {
|
|
|
|
simdgroup_load(mb[i],lsmb + SG_MAT_SIZE * i);
|
|
|
|
}
|
|
|
|
|
|
|
|
lsma += BLOCK_SIZE_M / SG_MAT_ROW * SG_MAT_SIZE;
|
|
|
|
lsmb += BLOCK_SIZE_N / SG_MAT_ROW * SG_MAT_SIZE;
|
2023-11-03 20:35:05 +01:00
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
#pragma unroll(8)
|
|
|
|
for (int i = 0; i < 8; i++){
|
|
|
|
simdgroup_multiply_accumulate(c_res[i], mb[i/4], ma[i%4], c_res[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if ((r0 + 1) * BLOCK_SIZE_M <= ne0 && (r1 + 1) * BLOCK_SIZE_N <= ne1) {
|
2023-11-03 20:35:05 +01:00
|
|
|
device float * C = dst + (BLOCK_SIZE_M * r0 + 32 * (sgitg & 1)) \
|
|
|
|
+ (BLOCK_SIZE_N * r1 + 16 * (sgitg >> 1)) * ne0 + im*ne1*ne0;
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = 0; i < 8; i++) {
|
|
|
|
simdgroup_store(c_res[i], C + 8 * (i%4) + 8 * ne0 * (i/4), ne0);
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
// block is smaller than 64x32, we should avoid writing data outside of the matrix
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
threadgroup float * temp_str = ((threadgroup float *)shared_memory) \
|
2023-09-05 12:54:40 +02:00
|
|
|
+ 32 * (sgitg&1) + (16 * (sgitg>>1)) * BLOCK_SIZE_M;
|
|
|
|
for (int i = 0; i < 8; i++) {
|
|
|
|
simdgroup_store(c_res[i], temp_str + 8 * (i%4) + 8 * BLOCK_SIZE_M * (i/4), BLOCK_SIZE_M);
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
2023-11-03 20:35:05 +01:00
|
|
|
|
|
|
|
device float * C = dst + (BLOCK_SIZE_M * r0) + (BLOCK_SIZE_N * r1) * ne0 + im*ne1*ne0;
|
|
|
|
if (sgitg == 0) {
|
2023-09-05 12:54:40 +02:00
|
|
|
for (int i = 0; i < n_rows; i++) {
|
2023-11-03 20:35:05 +01:00
|
|
|
for (int j = tiitg; j < n_cols; j += BLOCK_SIZE_N) {
|
2023-09-05 12:54:40 +02:00
|
|
|
*(C + i + j * ne0) = *(temp_str + i + j * BLOCK_SIZE_M);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
// same as kernel_mul_mm_impl, but src1 and dst are accessed via indices stored in rowids
|
2024-01-02 20:07:47 +01:00
|
|
|
template<typename block_q, short nl, void (*dequantize_func)(device const block_q *, short, thread half4x4 &)>
|
|
|
|
void kernel_mul_mm_id_impl(
|
|
|
|
device const uchar * src0,
|
|
|
|
device const uchar * src1,
|
2024-04-18 15:18:48 +02:00
|
|
|
threadgroup ushort2 * rowids,
|
2024-01-02 20:07:47 +01:00
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
2024-04-18 15:18:48 +02:00
|
|
|
constant int64_t & ne11,
|
2024-01-02 20:07:47 +01:00
|
|
|
constant int64_t & ne12,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
int64_t ne1,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne0ne1,
|
2024-01-02 20:07:47 +01:00
|
|
|
threadgroup uchar * shared_memory,
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
|
|
|
threadgroup half * sa = (threadgroup half *)(shared_memory);
|
|
|
|
threadgroup float * sb = (threadgroup float *)(shared_memory + 4096);
|
|
|
|
|
|
|
|
const uint r0 = tgpig.y;
|
|
|
|
const uint r1 = tgpig.x;
|
|
|
|
|
|
|
|
if (r1 * BLOCK_SIZE_N >= ne1) return;
|
|
|
|
|
|
|
|
// if this block is of 64x32 shape or smaller
|
|
|
|
short n_rows = (ne0 - r0 * BLOCK_SIZE_M < BLOCK_SIZE_M) ? (ne0 - r0 * BLOCK_SIZE_M) : BLOCK_SIZE_M;
|
|
|
|
short n_cols = (ne1 - r1 * BLOCK_SIZE_N < BLOCK_SIZE_N) ? (ne1 - r1 * BLOCK_SIZE_N) : BLOCK_SIZE_N;
|
|
|
|
|
|
|
|
// a thread shouldn't load data outside of the matrix
|
|
|
|
short thread_row = ((short)tiitg/THREAD_PER_ROW) < n_rows ? ((short)tiitg/THREAD_PER_ROW) : n_rows - 1;
|
|
|
|
short thread_col = ((short)tiitg/THREAD_PER_COL) < n_cols ? ((short)tiitg/THREAD_PER_COL) : n_cols - 1;
|
|
|
|
|
|
|
|
simdgroup_half8x8 ma[4];
|
|
|
|
simdgroup_float8x8 mb[2];
|
|
|
|
simdgroup_float8x8 c_res[8];
|
|
|
|
for (int i = 0; i < 8; i++){
|
|
|
|
c_res[i] = make_filled_simdgroup_matrix<float, 8>(0.f);
|
|
|
|
}
|
|
|
|
short il = (tiitg % THREAD_PER_ROW);
|
|
|
|
|
|
|
|
ushort offset1 = il/nl;
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
threadgroup const auto & id = rowids[r1 * BLOCK_SIZE_N + thread_col];
|
|
|
|
|
|
|
|
device const block_q * x = (device const block_q *)(src0 + (r0 * BLOCK_SIZE_M + thread_row) * nb01) + offset1;
|
2024-01-02 20:07:47 +01:00
|
|
|
device const float * y = (device const float *)(src1
|
2024-04-18 15:18:48 +02:00
|
|
|
+ nb12 * id[1]
|
|
|
|
+ nb11 * (id[0] % ne11)
|
2024-01-02 20:07:47 +01:00
|
|
|
+ nb10 * (BLOCK_SIZE_K / THREAD_PER_COL * (tiitg % THREAD_PER_COL)));
|
|
|
|
|
|
|
|
for (int loop_k = 0; loop_k < ne00; loop_k += BLOCK_SIZE_K) {
|
|
|
|
// load data and store to threadgroup memory
|
|
|
|
half4x4 temp_a;
|
|
|
|
dequantize_func(x, il, temp_a);
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
for (int i = 0; i < 16; i++) {
|
|
|
|
*(sa + SG_MAT_SIZE * ((tiitg / THREAD_PER_ROW / 8) \
|
|
|
|
+ (tiitg % THREAD_PER_ROW) * 16 + (i / 8) * 8) \
|
|
|
|
+ (tiitg / THREAD_PER_ROW) % 8 + (i & 7) * 8) = temp_a[i/4][i%4];
|
|
|
|
}
|
|
|
|
|
|
|
|
*(threadgroup float2x4 *)(sb + (tiitg % THREAD_PER_COL) * 8 * 32 + 8 * (tiitg / THREAD_PER_COL)) = *((device float2x4 *)y);
|
|
|
|
|
|
|
|
il = (il + 2 < nl) ? il + 2 : il % 2;
|
|
|
|
x = (il < 2) ? x + (2+nl-1)/nl : x;
|
|
|
|
y += BLOCK_SIZE_K;
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
|
|
|
// load matrices from threadgroup memory and conduct outer products
|
|
|
|
threadgroup half * lsma = (sa + THREAD_MAT_M * SG_MAT_SIZE * (sgitg % 2));
|
|
|
|
threadgroup float * lsmb = (sb + THREAD_MAT_N * SG_MAT_SIZE * (sgitg / 2));
|
|
|
|
|
|
|
|
for (int ik = 0; ik < BLOCK_SIZE_K / 8; ik++) {
|
|
|
|
for (int i = 0; i < 4; i++) {
|
2024-04-18 15:18:48 +02:00
|
|
|
simdgroup_load(ma[i], lsma + SG_MAT_SIZE * i);
|
2024-01-02 20:07:47 +01:00
|
|
|
}
|
|
|
|
simdgroup_barrier(mem_flags::mem_none);
|
|
|
|
for (int i = 0; i < 2; i++) {
|
2024-04-18 15:18:48 +02:00
|
|
|
simdgroup_load(mb[i], lsmb + SG_MAT_SIZE * i);
|
2024-01-02 20:07:47 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
lsma += BLOCK_SIZE_M / SG_MAT_ROW * SG_MAT_SIZE;
|
|
|
|
lsmb += BLOCK_SIZE_N / SG_MAT_ROW * SG_MAT_SIZE;
|
|
|
|
|
|
|
|
for (int i = 0; i < 8; i++){
|
|
|
|
simdgroup_multiply_accumulate(c_res[i], mb[i/4], ma[i%4], c_res[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
{
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
threadgroup float * temp_str = ((threadgroup float *)shared_memory) \
|
|
|
|
+ 32 * (sgitg&1) + (16 * (sgitg>>1)) * BLOCK_SIZE_M;
|
|
|
|
for (int i = 0; i < 8; i++) {
|
|
|
|
simdgroup_store(c_res[i], temp_str + 8 * (i%4) + 8 * BLOCK_SIZE_M * (i/4), BLOCK_SIZE_M);
|
|
|
|
}
|
|
|
|
|
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
device float * C = dst + (BLOCK_SIZE_M * r0);
|
2024-01-02 20:07:47 +01:00
|
|
|
if (sgitg == 0) {
|
2024-04-18 15:18:48 +02:00
|
|
|
for (int j = tiitg; j < n_cols; j += BLOCK_SIZE_N) {
|
|
|
|
threadgroup const auto & jid = rowids[r1 * BLOCK_SIZE_N + j];
|
|
|
|
int joff = jid[0] * ne0 + jid[1] * ne0ne1;
|
|
|
|
for (int i = 0; i < n_rows; i++) {
|
|
|
|
*(C + i + joff) = *(temp_str + i + j * BLOCK_SIZE_M);
|
2024-01-02 20:07:47 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
template<typename block_q, short nl, void (*dequantize_func)(device const block_q *, short, thread half4x4 &)>
|
|
|
|
kernel void kernel_mul_mm(device const uchar * src0,
|
|
|
|
device const uchar * src1,
|
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne02,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne12,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
|
|
|
constant uint & r2,
|
|
|
|
constant uint & r3,
|
|
|
|
threadgroup uchar * shared_memory [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
kernel_mul_mm_impl<block_q, nl, dequantize_func>(
|
|
|
|
src0,
|
|
|
|
src1,
|
|
|
|
dst,
|
|
|
|
ne00,
|
|
|
|
ne02,
|
|
|
|
nb01,
|
|
|
|
nb02,
|
|
|
|
ne12,
|
|
|
|
nb10,
|
|
|
|
nb11,
|
|
|
|
nb12,
|
|
|
|
ne0,
|
|
|
|
ne1,
|
|
|
|
r2,
|
|
|
|
r3,
|
|
|
|
shared_memory,
|
|
|
|
tgpig,
|
|
|
|
tiitg,
|
|
|
|
sgitg);
|
|
|
|
}
|
|
|
|
|
|
|
|
template<typename block_q, short nl, void (*dequantize_func)(device const block_q *, short, thread half4x4 &)>
|
|
|
|
kernel void kernel_mul_mm_id(
|
2024-04-03 15:07:05 +02:00
|
|
|
device const uchar * src0s,
|
2023-12-07 21:27:19 +01:00
|
|
|
device const uchar * src1,
|
2024-01-02 20:07:47 +01:00
|
|
|
device float * dst,
|
2024-04-03 15:07:05 +02:00
|
|
|
device const uchar * ids,
|
2024-04-18 15:18:48 +02:00
|
|
|
constant int64_t & nei0,
|
|
|
|
constant int64_t & nei1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nbi1,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne02,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
2024-04-18 15:18:48 +02:00
|
|
|
constant int64_t & ne11,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne12,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant int64_t & ne13,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb1,
|
2023-12-07 21:27:19 +01:00
|
|
|
threadgroup uchar * shared_memory [[threadgroup(0)]],
|
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
const int32_t i02 = tgpig.z;
|
|
|
|
tgpig.z = 0;
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
device const uchar * src0 = src0s + i02*nb02;
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
// row indices
|
|
|
|
threadgroup ushort2 * rowids = (threadgroup ushort2 *)(shared_memory + 8192);
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
// TODO: parallelize this loop
|
2024-03-10 22:12:48 +01:00
|
|
|
int64_t _ne1 = 0;
|
2024-04-18 15:18:48 +02:00
|
|
|
for (ushort ii1 = 0; ii1 < nei1; ii1++) {
|
|
|
|
for (ushort ii0 = 0; ii0 < nei0; ii0++) {
|
|
|
|
int32_t id = ((device int32_t *) (ids + ii1*nbi1))[ii0];
|
|
|
|
if (id == i02) {
|
|
|
|
//if (tiitg == 0) {
|
|
|
|
rowids[_ne1] = ushort2(ii0, ii1);
|
|
|
|
//}
|
|
|
|
_ne1++;
|
|
|
|
}
|
2024-01-02 20:07:47 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
threadgroup_barrier(mem_flags::mem_threadgroup);
|
|
|
|
|
2024-01-02 20:07:47 +01:00
|
|
|
kernel_mul_mm_id_impl<block_q, nl, dequantize_func>(
|
2024-04-03 15:07:05 +02:00
|
|
|
src0,
|
2024-01-02 20:07:47 +01:00
|
|
|
src1,
|
2024-04-18 15:18:48 +02:00
|
|
|
rowids,
|
2024-01-02 20:07:47 +01:00
|
|
|
dst,
|
2023-12-07 21:27:19 +01:00
|
|
|
ne00,
|
|
|
|
ne02,
|
|
|
|
nb01,
|
|
|
|
nb02,
|
2024-04-18 15:18:48 +02:00
|
|
|
ne11,
|
2023-12-07 21:27:19 +01:00
|
|
|
ne12,
|
|
|
|
nb10,
|
|
|
|
nb11,
|
|
|
|
nb12,
|
|
|
|
ne0,
|
2024-01-02 20:07:47 +01:00
|
|
|
_ne1,
|
2024-04-18 15:18:48 +02:00
|
|
|
ne0*ne1,
|
2023-12-07 21:27:19 +01:00
|
|
|
shared_memory,
|
|
|
|
tgpig,
|
|
|
|
tiitg,
|
|
|
|
sgitg);
|
|
|
|
}
|
|
|
|
|
2023-09-05 12:54:40 +02:00
|
|
|
#if QK_K == 256
|
|
|
|
#define QK_NL 16
|
|
|
|
#else
|
|
|
|
#define QK_NL 4
|
|
|
|
#endif
|
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
//
|
|
|
|
// get rows
|
|
|
|
//
|
|
|
|
|
2023-12-07 21:27:19 +01:00
|
|
|
typedef void (get_rows_t)(
|
|
|
|
device const void * src0,
|
2023-12-13 20:55:03 +01:00
|
|
|
device const char * src1,
|
2023-12-07 21:27:19 +01:00
|
|
|
device float * dst,
|
|
|
|
constant int64_t & ne00,
|
|
|
|
constant uint64_t & nb01,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
2023-12-07 21:27:19 +01:00
|
|
|
constant uint64_t & nb1,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant uint64_t & nb2,
|
|
|
|
uint3, uint, uint3);
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
//template [[host_name("kernel_get_rows_f32")]] kernel get_rows_t kernel_get_rows<float4x4, 1, dequantize_f32>;
|
|
|
|
//template [[host_name("kernel_get_rows_f16")]] kernel get_rows_t kernel_get_rows<half4x4, 1, dequantize_f16>;
|
2023-09-05 12:54:40 +02:00
|
|
|
template [[host_name("kernel_get_rows_q4_0")]] kernel get_rows_t kernel_get_rows<block_q4_0, 2, dequantize_q4_0>;
|
|
|
|
template [[host_name("kernel_get_rows_q4_1")]] kernel get_rows_t kernel_get_rows<block_q4_1, 2, dequantize_q4_1>;
|
2023-11-03 20:35:05 +01:00
|
|
|
template [[host_name("kernel_get_rows_q5_0")]] kernel get_rows_t kernel_get_rows<block_q5_0, 2, dequantize_q5_0>;
|
|
|
|
template [[host_name("kernel_get_rows_q5_1")]] kernel get_rows_t kernel_get_rows<block_q5_1, 2, dequantize_q5_1>;
|
2023-09-05 12:54:40 +02:00
|
|
|
template [[host_name("kernel_get_rows_q8_0")]] kernel get_rows_t kernel_get_rows<block_q8_0, 2, dequantize_q8_0>;
|
|
|
|
template [[host_name("kernel_get_rows_q2_K")]] kernel get_rows_t kernel_get_rows<block_q2_K, QK_NL, dequantize_q2_K>;
|
|
|
|
template [[host_name("kernel_get_rows_q3_K")]] kernel get_rows_t kernel_get_rows<block_q3_K, QK_NL, dequantize_q3_K>;
|
|
|
|
template [[host_name("kernel_get_rows_q4_K")]] kernel get_rows_t kernel_get_rows<block_q4_K, QK_NL, dequantize_q4_K>;
|
|
|
|
template [[host_name("kernel_get_rows_q5_K")]] kernel get_rows_t kernel_get_rows<block_q5_K, QK_NL, dequantize_q5_K>;
|
|
|
|
template [[host_name("kernel_get_rows_q6_K")]] kernel get_rows_t kernel_get_rows<block_q6_K, QK_NL, dequantize_q6_K>;
|
2024-01-08 16:02:32 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq2_xxs")]] kernel get_rows_t kernel_get_rows<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>;
|
2024-01-11 20:39:39 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq2_xs")]] kernel get_rows_t kernel_get_rows<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
|
2024-01-30 14:14:12 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq3_xxs")]] kernel get_rows_t kernel_get_rows<block_iq3_xxs, QK_NL, dequantize_iq3_xxs>;
|
2024-02-24 15:23:52 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq3_s")]] kernel get_rows_t kernel_get_rows<block_iq3_s, QK_NL, dequantize_iq3_s>;
|
2024-02-26 17:28:38 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq2_s")]] kernel get_rows_t kernel_get_rows<block_iq2_s, QK_NL, dequantize_iq2_s>;
|
2024-02-18 17:16:55 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq1_s")]] kernel get_rows_t kernel_get_rows<block_iq1_s, QK_NL, dequantize_iq1_s>;
|
2024-03-27 17:55:10 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq1_m")]] kernel get_rows_t kernel_get_rows<block_iq1_m, QK_NL, dequantize_iq1_m>;
|
2024-02-27 15:34:24 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq4_nl")]] kernel get_rows_t kernel_get_rows<block_iq4_nl, 2, dequantize_iq4_nl>;
|
2024-02-28 09:37:02 +01:00
|
|
|
#if QK_K == 64
|
|
|
|
template [[host_name("kernel_get_rows_iq4_xs")]] kernel get_rows_t kernel_get_rows<block_iq4_xs, 2, dequantize_iq4_xs>;
|
|
|
|
#else
|
2024-02-27 15:34:24 +01:00
|
|
|
template [[host_name("kernel_get_rows_iq4_xs")]] kernel get_rows_t kernel_get_rows<block_iq4_xs, QK_NL, dequantize_iq4_xs>;
|
2024-02-28 09:37:02 +01:00
|
|
|
#endif
|
2023-09-05 12:54:40 +02:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
//
|
|
|
|
// matrix-matrix multiplication
|
|
|
|
//
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
typedef decltype(kernel_mul_mm<float4x4, 1, dequantize_f32>) mat_mm_t;
|
2023-09-15 11:18:18 +02:00
|
|
|
|
2024-03-27 17:55:10 +01:00
|
|
|
template [[host_name("kernel_mul_mm_f32_f32")]] kernel mat_mm_t kernel_mul_mm<float4x4, 1, dequantize_f32>;
|
|
|
|
template [[host_name("kernel_mul_mm_f16_f32")]] kernel mat_mm_t kernel_mul_mm<half4x4, 1, dequantize_f16>;
|
|
|
|
template [[host_name("kernel_mul_mm_q4_0_f32")]] kernel mat_mm_t kernel_mul_mm<block_q4_0, 2, dequantize_q4_0>;
|
|
|
|
template [[host_name("kernel_mul_mm_q4_1_f32")]] kernel mat_mm_t kernel_mul_mm<block_q4_1, 2, dequantize_q4_1>;
|
|
|
|
template [[host_name("kernel_mul_mm_q5_0_f32")]] kernel mat_mm_t kernel_mul_mm<block_q5_0, 2, dequantize_q5_0>;
|
|
|
|
template [[host_name("kernel_mul_mm_q5_1_f32")]] kernel mat_mm_t kernel_mul_mm<block_q5_1, 2, dequantize_q5_1>;
|
|
|
|
template [[host_name("kernel_mul_mm_q8_0_f32")]] kernel mat_mm_t kernel_mul_mm<block_q8_0, 2, dequantize_q8_0>;
|
|
|
|
template [[host_name("kernel_mul_mm_q2_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q2_K, QK_NL, dequantize_q2_K>;
|
|
|
|
template [[host_name("kernel_mul_mm_q3_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q3_K, QK_NL, dequantize_q3_K>;
|
|
|
|
template [[host_name("kernel_mul_mm_q4_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q4_K, QK_NL, dequantize_q4_K>;
|
|
|
|
template [[host_name("kernel_mul_mm_q5_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q5_K, QK_NL, dequantize_q5_K>;
|
|
|
|
template [[host_name("kernel_mul_mm_q6_K_f32")]] kernel mat_mm_t kernel_mul_mm<block_q6_K, QK_NL, dequantize_q6_K>;
|
2024-01-08 16:02:32 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq2_xxs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>;
|
2024-01-11 20:39:39 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq2_xs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
|
2024-01-30 14:14:12 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq3_xxs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq3_xxs, QK_NL, dequantize_iq3_xxs>;
|
2024-02-24 15:23:52 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq3_s_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq3_s, QK_NL, dequantize_iq3_s>;
|
2024-02-26 17:28:38 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq2_s_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq2_s, QK_NL, dequantize_iq2_s>;
|
2024-02-18 17:16:55 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq1_s_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq1_s, QK_NL, dequantize_iq1_s>;
|
2024-03-27 17:55:10 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq1_m_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq1_m, QK_NL, dequantize_iq1_m>;
|
2024-02-27 15:34:24 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq4_nl_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq4_nl, 2, dequantize_iq4_nl>;
|
2024-02-28 09:37:02 +01:00
|
|
|
#if QK_K == 64
|
|
|
|
template [[host_name("kernel_mul_mm_iq4_xs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq4_nl, 2, dequantize_iq4_xs>;
|
|
|
|
#else
|
2024-02-27 15:34:24 +01:00
|
|
|
template [[host_name("kernel_mul_mm_iq4_xs_f32")]] kernel mat_mm_t kernel_mul_mm<block_iq4_xs, QK_NL, dequantize_iq4_xs>;
|
2024-02-28 09:37:02 +01:00
|
|
|
#endif
|
2023-12-07 21:27:19 +01:00
|
|
|
|
2023-12-13 20:55:03 +01:00
|
|
|
//
|
|
|
|
// indirect matrix-matrix multiplication
|
|
|
|
//
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
typedef decltype(kernel_mul_mm_id<float4x4, 1, dequantize_f32>) mat_mm_id_t;
|
2023-12-07 21:27:19 +01:00
|
|
|
|
2024-03-27 17:55:10 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_f32_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<float4x4, 1, dequantize_f32>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_f16_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<half4x4, 1, dequantize_f16>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q4_0_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q4_0, 2, dequantize_q4_0>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q4_1_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q4_1, 2, dequantize_q4_1>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q5_0_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q5_0, 2, dequantize_q5_0>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q5_1_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q5_1, 2, dequantize_q5_1>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q8_0_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q8_0, 2, dequantize_q8_0>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q2_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q2_K, QK_NL, dequantize_q2_K>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q3_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q3_K, QK_NL, dequantize_q3_K>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q4_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q4_K, QK_NL, dequantize_q4_K>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q5_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q5_K, QK_NL, dequantize_q5_K>;
|
|
|
|
template [[host_name("kernel_mul_mm_id_q6_K_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_q6_K, QK_NL, dequantize_q6_K>;
|
2024-01-08 16:02:32 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq2_xxs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq2_xxs, QK_NL, dequantize_iq2_xxs>;
|
2024-01-11 20:39:39 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq2_xs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq2_xs, QK_NL, dequantize_iq2_xs>;
|
2024-01-30 14:14:12 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq3_xxs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq3_xxs, QK_NL, dequantize_iq3_xxs>;
|
2024-02-24 15:23:52 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq3_s_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq3_s, QK_NL, dequantize_iq3_s>;
|
2024-02-26 17:28:38 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq2_s_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq2_s, QK_NL, dequantize_iq2_s>;
|
2024-02-18 17:16:55 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq1_s_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq1_s, QK_NL, dequantize_iq1_s>;
|
2024-03-27 17:55:10 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq1_m_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq1_m, QK_NL, dequantize_iq1_m>;
|
2024-02-27 15:34:24 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq4_nl_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq4_nl, 2, dequantize_iq4_nl>;
|
2024-02-28 09:37:02 +01:00
|
|
|
#if QK_K == 64
|
|
|
|
template [[host_name("kernel_mul_mm_id_iq4_xs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq4_xs, 2, dequantize_iq4_xs>;
|
|
|
|
#else
|
2024-02-27 15:34:24 +01:00
|
|
|
template [[host_name("kernel_mul_mm_id_iq4_xs_f32")]] kernel mat_mm_id_t kernel_mul_mm_id<block_iq4_xs, QK_NL, dequantize_iq4_xs>;
|
2024-02-28 09:37:02 +01:00
|
|
|
#endif
|
2023-12-13 20:55:03 +01:00
|
|
|
|
|
|
|
//
|
|
|
|
// matrix-vector multiplication
|
|
|
|
//
|
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
typedef void (kernel_mul_mv_impl_t)(
|
|
|
|
device const char * src0,
|
|
|
|
device const char * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
uint64_t nb00,
|
|
|
|
uint64_t nb01,
|
|
|
|
uint64_t nb02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne11,
|
|
|
|
int64_t ne12,
|
|
|
|
uint64_t nb10,
|
|
|
|
uint64_t nb11,
|
|
|
|
uint64_t nb12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg);
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
typedef void (kernel_mul_mv2_impl_t)(
|
|
|
|
device const void * src0,
|
|
|
|
device const float * src1,
|
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg);
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
template<kernel_mul_mv_impl_t impl_fn>
|
|
|
|
void mmv_fn(
|
|
|
|
device const char * src0,
|
2023-12-13 20:55:03 +01:00
|
|
|
device const char * src1,
|
2024-01-02 20:07:47 +01:00
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
uint64_t nb00,
|
|
|
|
uint64_t nb01,
|
|
|
|
uint64_t nb02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne11,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne13,
|
|
|
|
uint64_t nb10,
|
|
|
|
uint64_t nb11,
|
|
|
|
uint64_t nb12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint64_t nb1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiitg,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-04-12 18:13:20 +02:00
|
|
|
impl_fn(src0,src1,dst,ne00,ne01,ne02,nb00,nb01,nb02,ne10,ne11,ne12,nb10,nb11,nb12,ne0,ne1,r2,r3,tgpig,tiisg);
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
template<kernel_mul_mv2_impl_t impl_fn>
|
|
|
|
void mmv_fn(
|
|
|
|
device const char * src0,
|
2023-12-13 20:55:03 +01:00
|
|
|
device const char * src1,
|
2024-01-02 20:07:47 +01:00
|
|
|
device float * dst,
|
2024-04-18 15:18:48 +02:00
|
|
|
int64_t ne00,
|
|
|
|
int64_t ne01,
|
|
|
|
int64_t ne02,
|
|
|
|
uint64_t nb00,
|
|
|
|
uint64_t nb01,
|
|
|
|
uint64_t nb02,
|
|
|
|
int64_t ne10,
|
|
|
|
int64_t ne11,
|
|
|
|
int64_t ne12,
|
|
|
|
int64_t ne13,
|
|
|
|
uint64_t nb10,
|
|
|
|
uint64_t nb11,
|
|
|
|
uint64_t nb12,
|
|
|
|
int64_t ne0,
|
|
|
|
int64_t ne1,
|
|
|
|
uint64_t nb1,
|
|
|
|
uint r2,
|
|
|
|
uint r3,
|
|
|
|
threadgroup int8_t * shared_values,
|
|
|
|
uint3 tgpig,
|
|
|
|
uint tiitg,
|
|
|
|
uint tiisg,
|
|
|
|
uint sgitg) {
|
2024-04-12 18:13:20 +02:00
|
|
|
impl_fn(src0,(const device float *)src1,dst,ne00,ne01,ne02,ne10,ne12,ne0,ne1,r2,r3,shared_values,tgpig,tiisg,sgitg);
|
2023-12-13 20:55:03 +01:00
|
|
|
}
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
typedef decltype(mmv_fn<kernel_mul_mv_f32_f32_impl>) mul_mv_impl_fn_t;
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
template<mul_mv_impl_fn_t impl_fn>
|
|
|
|
kernel void kernel_mul_mv_id(
|
2024-04-03 15:07:05 +02:00
|
|
|
device const char * src0s,
|
2023-12-13 20:55:03 +01:00
|
|
|
device const char * src1,
|
2024-01-02 20:07:47 +01:00
|
|
|
device float * dst,
|
2024-04-03 15:07:05 +02:00
|
|
|
device const char * ids,
|
2024-04-18 15:18:48 +02:00
|
|
|
constant int64_t & nei0,
|
|
|
|
constant int64_t & nei1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nbi1,
|
2023-12-13 20:55:03 +01:00
|
|
|
constant int64_t & ne00,
|
|
|
|
constant int64_t & ne01,
|
|
|
|
constant int64_t & ne02,
|
|
|
|
constant uint64_t & nb00,
|
|
|
|
constant uint64_t & nb01,
|
|
|
|
constant uint64_t & nb02,
|
|
|
|
constant int64_t & ne10,
|
|
|
|
constant int64_t & ne11,
|
|
|
|
constant int64_t & ne12,
|
|
|
|
constant int64_t & ne13,
|
|
|
|
constant uint64_t & nb10,
|
|
|
|
constant uint64_t & nb11,
|
|
|
|
constant uint64_t & nb12,
|
|
|
|
constant int64_t & ne0,
|
|
|
|
constant int64_t & ne1,
|
2024-01-02 09:57:44 +01:00
|
|
|
constant uint64_t & nb1,
|
2024-04-12 18:13:20 +02:00
|
|
|
threadgroup int8_t * shared_values [[threadgroup(0)]],
|
2023-12-13 20:55:03 +01:00
|
|
|
uint3 tgpig[[threadgroup_position_in_grid]],
|
|
|
|
uint tiitg[[thread_index_in_threadgroup]],
|
|
|
|
uint tiisg[[thread_index_in_simdgroup]],
|
|
|
|
uint sgitg[[simdgroup_index_in_threadgroup]]) {
|
2024-04-18 15:18:48 +02:00
|
|
|
const int iid1 = tgpig.z/nei0;
|
|
|
|
const int idx = tgpig.z%nei0;
|
|
|
|
|
|
|
|
tgpig.z = 0;
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
const int32_t i02 = ((device const int32_t *) (ids + iid1*nbi1))[idx];
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
const int64_t i11 = idx % ne11;
|
|
|
|
const int64_t i12 = iid1;
|
|
|
|
|
|
|
|
const int64_t i1 = idx;
|
|
|
|
const int64_t i2 = i12;
|
|
|
|
|
|
|
|
device const char * src0_cur = src0s + i02*nb02;
|
|
|
|
device const char * src1_cur = src1 + i11*nb11 + i12*nb12;
|
|
|
|
device float * dst_cur = dst + i1*ne0 + i2*ne1*ne0;
|
2023-12-13 20:55:03 +01:00
|
|
|
|
2024-04-12 18:13:20 +02:00
|
|
|
impl_fn(
|
2024-04-18 15:18:48 +02:00
|
|
|
/* src0 */ src0_cur,
|
|
|
|
/* src1 */ src1_cur,
|
|
|
|
/* dst */ dst_cur,
|
|
|
|
/* ne00 */ ne00,
|
|
|
|
/* ne01 */ ne01,
|
|
|
|
/* ne02 */ 1,//ne02,
|
|
|
|
/* nb00 */ nb00,
|
|
|
|
/* nb01 */ nb01,
|
|
|
|
/* nb02 */ nb02,
|
|
|
|
/* ne10 */ ne10,
|
|
|
|
/* ne11 */ 1,//ne11,
|
|
|
|
/* ne12 */ 1,//ne12,
|
|
|
|
/* ne13 */ 1,//ne13,
|
|
|
|
/* nb10 */ nb10,
|
|
|
|
/* nb11 */ nb11,
|
|
|
|
/* nb12 */ nb12,
|
|
|
|
/* ne0 */ ne0,
|
|
|
|
/* ne1 */ 1,//ne1,
|
|
|
|
/* nb1 */ nb1,
|
|
|
|
/* r2 */ 1,
|
|
|
|
/* r3 */ 1,
|
2024-04-12 18:13:20 +02:00
|
|
|
shared_values,
|
2023-12-13 20:55:03 +01:00
|
|
|
tgpig,
|
2024-04-12 18:13:20 +02:00
|
|
|
tiitg,
|
2023-12-13 20:55:03 +01:00
|
|
|
tiisg,
|
|
|
|
sgitg);
|
|
|
|
}
|
|
|
|
|
2024-04-18 15:18:48 +02:00
|
|
|
typedef decltype(kernel_mul_mv_id<mmv_fn<kernel_mul_mv_f32_f32_impl>>) kernel_mul_mv_id_t;
|
2024-04-12 18:13:20 +02:00
|
|
|
|
|
|
|
template [[host_name("kernel_mul_mv_id_f32_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_f32_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_f16_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_f16_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q8_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_q8_0_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q4_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<mul_vec_q_n_f32_impl<block_q4_0, N_DST, N_SIMDGROUP, N_SIMDWIDTH>>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q4_1_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<mul_vec_q_n_f32_impl<block_q4_1, N_DST, N_SIMDGROUP, N_SIMDWIDTH>>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q5_0_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<mul_vec_q_n_f32_impl<block_q5_0, N_DST, N_SIMDGROUP, N_SIMDWIDTH>>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q5_1_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<mul_vec_q_n_f32_impl<block_q5_1, N_DST, N_SIMDGROUP, N_SIMDWIDTH>>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q2_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_q2_K_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q3_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_q3_K_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q4_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_q4_K_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q5_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_q5_K_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_q6_K_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_q6_K_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq1_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq1_s_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq1_m_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq1_m_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq2_xxs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq2_xxs_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq2_xs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq2_xs_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq3_xxs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq3_xxs_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq3_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq3_s_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq2_s_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq2_s_f32_impl>>;
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq4_nl_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq4_nl_f32_impl>>;
|
|
|
|
#if QK_K != 64
|
|
|
|
template [[host_name("kernel_mul_mv_id_iq4_xs_f32")]] kernel kernel_mul_mv_id_t kernel_mul_mv_id<mmv_fn<kernel_mul_mv_iq4_xs_f32_impl>>;
|
|
|
|
#endif
|
2023-12-13 20:55:03 +01:00
|
|
|
|