Revert "[SYCL] fallback mmvq (ggml/9088)" (llama/9579)

This reverts commit 50addec9a532a6518146ab837a85504850627316.
This commit is contained in:
Akarshan Biswas 2024-09-23 08:58:06 +05:30 committed by Georgi Gerganov
parent 13f41af43e
commit c36ddc43c6
2 changed files with 1 additions and 3 deletions

View File

@ -3496,8 +3496,7 @@ static void ggml_sycl_mul_mat(ggml_backend_sycl_context & ctx, const ggml_tensor
bool use_mul_mat_vec_q = ggml_is_quantized(src0->type)
&& src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32
&& src1->ne[1] <= MMVQ_MAX_BATCH_SIZE
&& (ctx.stream()->get_backend() == sycl::backend::ext_oneapi_cuda || src1->ne[1] > MMVQ_MIN_BATCH_SIZE);
&& src1->ne[1] <= MMVQ_MAX_BATCH_SIZE;
bool use_mul_mat_q = ggml_sycl_supports_mmq(src0->type)
&& src1->type == GGML_TYPE_F32 && dst->type == GGML_TYPE_F32;

View File

@ -134,7 +134,6 @@ typedef sycl::float2 dfloat2;
#endif // GGML_SYCL_F16
#define MMVQ_MAX_BATCH_SIZE 8
#define MMVQ_MIN_BATCH_SIZE 4
static const int8_t kvalues_iq4nl[16]={-127, -104, -83, -65, -49, -35, -22, -10, 1, 13, 25, 38, 53, 69, 89, 113};