ExternalVendorCode/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/whisper.cpp.git synced 2025-04-27 06:20:07 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Vulkan: Tune Vulkan mmq int dot shader for performance (llama/12767)

Browse Source
This commit is contained in:
0cc4m 2025-04-05 18:04:03 +02:00 committed by Georgi Gerganov
parent 8f8ede1b12
commit 8add58aa5e
1 changed files with 4 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
ggml/src/ggml-vulkan/ggml-vulkan.cpp
View File

@ -1736,19 +1736,9 @@ static void ggml_vk_load_shaders(vk_device& device) {
m_warptile_mmq = { 128, 64, 64, 32, subgroup_size_8, 32, 2, tm_m, tn_m, tk_m, subgroup_size_8 }; m_warptile_mmq = { 128, 64, 64, 32, subgroup_size_8, 32, 2, tm_m, tn_m, tk_m, subgroup_size_8 };
s_warptile_mmq = { subgroup_size_32, 32, 32, 32, 32, 32, 2, tm_s, tn_s, tk_s, subgroup_size_8 }; s_warptile_mmq = { subgroup_size_32, 32, 32, 32, 32, 32, 2, tm_s, tn_s, tk_s, subgroup_size_8 };
const uint32_t tm_int_l = device->coopmat_int_support ? device->coopmat_int_m : 4; l_warptile_mmq_int = { 128, 128, 128, 32, subgroup_size_8 * 2, 64, 2, 4, 4, 1, subgroup_size_8 };
const uint32_t tm_int_m = device->coopmat_int_support ? device->coopmat_int_m : 4; m_warptile_mmq_int = { 128, 64, 64, 32, subgroup_size_8, 32, 2, 2, 2, 1, subgroup_size_8 };
const uint32_t tm_int_s = device->coopmat_int_support ? device->coopmat_int_m : 2; s_warptile_mmq_int = { subgroup_size_32, 32, 32, 32, 32, 32, 2, 2, 1, 1, subgroup_size_8 };
const uint32_t tn_int_l = device->coopmat_int_support ? device->coopmat_int_n : 4;
const uint32_t tn_int_m = device->coopmat_int_support ? device->coopmat_int_n : 2;
const uint32_t tn_int_s = device->coopmat_int_support ? device->coopmat_int_n : 2;
const uint32_t tk_int_l = device->coopmat_int_support ? device->coopmat_int_k : 1;
const uint32_t tk_int_m = device->coopmat_int_support ? device->coopmat_int_k : 1;
const uint32_t tk_int_s = device->coopmat_int_support ? device->coopmat_int_k : 1;
l_warptile_mmq_int = { 128, 128, 128, 32, subgroup_size_8 * 2, 64, 2, tm_int_l, tn_int_l, tk_int_l, subgroup_size_8 };
m_warptile_mmq_int = { 128, 64, 64, 32, subgroup_size_8, 32, 2, tm_int_m, tn_int_m, tk_int_m, subgroup_size_8 };
s_warptile_mmq_int = { subgroup_size_32, 32, 32, 32, 32, 32, 2, tm_int_s, tn_int_s, tk_int_s, subgroup_size_8 };
l_mmq_wg_denoms = l_wg_denoms = {128, 128, 1 }; l_mmq_wg_denoms = l_wg_denoms = {128, 128, 1 };
m_mmq_wg_denoms = m_wg_denoms = { 64, 64, 1 }; m_mmq_wg_denoms = m_wg_denoms = { 64, 64, 1 };
@ -7922,7 +7912,7 @@ static void ggml_vk_preallocate_buffers(ggml_backend_vk_context * ctx) {
128, 49, 49, 128, 49, 49,
4096, 49, 4096, 4096, 49, 4096,
}; };
const size_t num_it = 1; const size_t num_it = 100;
ggml_vk_test_dequant_matmul(ctx, 4096, 512, 4096, 2, num_it, 1, 0, GGML_TYPE_Q4_0); ggml_vk_test_dequant_matmul(ctx, 4096, 512, 4096, 2, num_it, 1, 0, GGML_TYPE_Q4_0);
ggml_vk_test_dequant_matmul(ctx, 4096, 512, 4096, 2, num_it, 1, 1, GGML_TYPE_Q4_0); ggml_vk_test_dequant_matmul(ctx, 4096, 512, 4096, 2, num_it, 1, 1, GGML_TYPE_Q4_0);