public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-12 01:05:43 +02:00
Code Issues Packages Projects Releases Wiki Activity

vulkan: revert mul_mat_id tile selection PR #29182 (llama/29936)

Browse Source 
* Fix Intel prefill regression on MoE models

* Revert the n_per_expert change back to nei1
This commit is contained in:
virajwad
2026-10-06 10:38:22 +03:00
committed by Georgi Gerganov
parent ebb8a1706f
commit 563a9f7ae7
1 changed files with 3 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-vulkan/ggml-vulkan.cpp
+3 -4
View File
@@ -7428,11 +7428,10 @@ static void ggml_vk_mul_mat_id_q_f16(ggml_backend_vk_context * ctx, vk_context&
GGML_ASSERT(mmp_map != nullptr);
const uint32_t n_per_expert = (uint32_t)CEIL_DIV(nei0 * nei1, n_as);
const uint32_t kpad = quantize_y ? 0 : ggml_vk_align_size(ne10, ggml_vk_guess_matmul_pipeline_align_map(ctx, *mmp_map, ne01, n_per_expert, true));
const bool aligned = !quantize_y && ne10 == kpad && ne01 > 8 && n_per_expert > 8;
const uint32_t kpad = quantize_y ? 0 : ggml_vk_align_size(ne10, ggml_vk_guess_matmul_pipeline_align_map(ctx, *mmp_map, ne01, nei1, true));
const bool aligned = !quantize_y && ne10 == kpad && ne01 > 8 && nei1 > 8;
vk_pipeline pipeline = ggml_vk_guess_matmul_pipeline_map(ctx, *mmp_map, ne01, n_per_expert, aligned, true);
vk_pipeline pipeline = ggml_vk_guess_matmul_pipeline_map(ctx, *mmp_map, ne01, nei1, aligned, true);
if (ggml_nbytes(src0) > ctx->device->properties.limits.maxStorageBufferRange) {
pipeline = ggml_vk_get_64b_indexing_pipeline(ctx, pipeline);