vulkan : fix out-of-bounds access in argmax kernel (llama/15342)

ggml-ci
This commit is contained in:
Georgi Gerganov
2025-08-15 17:16:36 +03:00
parent 1a92fde1b6
commit 1e8d692365
2 changed files with 14 additions and 5 deletions

View File

@@ -8392,7 +8392,7 @@ static void ggml_vk_sum_rows(ggml_backend_vk_context * ctx, vk_context& subctx,
} }
static void ggml_vk_argmax(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, ggml_tensor * dst, bool dryrun = false) { static void ggml_vk_argmax(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, ggml_tensor * dst, bool dryrun = false) {
ggml_vk_op_f32<vk_op_push_constants>(ctx, subctx, src0, nullptr, nullptr, dst, GGML_OP_ARGMAX, { (uint32_t)src0->ne[0], 0, 0.0f, 0.0f }, dryrun); ggml_vk_op_f32<vk_op_push_constants>(ctx, subctx, src0, nullptr, nullptr, dst, GGML_OP_ARGMAX, { (uint32_t)src0->ne[0], (uint32_t)src0->ne[1], 0.0f, 0.0f }, dryrun);
} }
static void ggml_vk_count_equal(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, bool dryrun = false) { static void ggml_vk_count_equal(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, bool dryrun = false) {

View File

@@ -5,6 +5,8 @@
#extension GL_EXT_control_flow_attributes : enable #extension GL_EXT_control_flow_attributes : enable
#define FLT_MAX 3.402823466e+38F
layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in; layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
layout (binding = 0) readonly buffer A {A_TYPE data_a[];}; layout (binding = 0) readonly buffer A {A_TYPE data_a[];};
@@ -19,19 +21,26 @@ void main() {
const uint row = gl_WorkGroupID.z * 262144 + gl_WorkGroupID.y * 512 + gl_WorkGroupID.x; const uint row = gl_WorkGroupID.z * 262144 + gl_WorkGroupID.y * 512 + gl_WorkGroupID.x;
const uint col = gl_LocalInvocationID.x; const uint col = gl_LocalInvocationID.x;
if (col >= p.KX) { if (row >= p.KY) {
return; return;
} }
A_TYPE amax = data_a[row*p.KX + col];
tmp[col] = col; A_TYPE amax = -FLT_MAX;
uint acol = col;
if (col < p.KX) {
amax = data_a[row*p.KX + col];
}
for (uint i = col + BLOCK_SIZE; i < p.KX; i += BLOCK_SIZE) { for (uint i = col + BLOCK_SIZE; i < p.KX; i += BLOCK_SIZE) {
A_TYPE val = data_a[row*p.KX + i]; A_TYPE val = data_a[row*p.KX + i];
if (val > amax) { if (val > amax) {
amax = val; amax = val;
tmp[col] = i; acol = i;
} }
} }
tmp[col] = acol;
tmpmax[col] = amax; tmpmax[col] = amax;
barrier(); barrier();