happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

move intel check to chip specific tuning part

This commit is contained in:
Viraj Wadhwa 2025-12-18 15:45:06 -07:00
parent 6b7f1e867e
commit 0cfa616d61
1 changed files with 8 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
ggml/src/ggml-vulkan/ggml-vulkan.cpp
View File

@ -2838,20 +2838,19 @@ static void ggml_vk_load_shaders(vk_device& device) {
m_warptile_mmqid_int_k = { 128, 64, 64, 32, mul_mat_subgroup_size_16, 32, 1, 2, 2, 1, mul_mat_subgroup_size_16 };
s_warptile_mmqid_int_k = { mul_mat_subgroup_size_32, 32, 32, 32, 32, 32, 1, 2, 1, 1, mul_mat_subgroup_size_16 };
l_mmq_wg_denoms = l_wg_denoms = { 128, 128, 1 };
m_mmq_wg_denoms = m_wg_denoms = { 64, 64, 1 };
s_mmq_wg_denoms = s_wg_denoms = { 32, 32, 1 };
l_align = 128;
m_align = 64;
s_align = 32;
// chip specific tuning
if ((device->architecture == AMD_GCN) && (device->driver_id != vk::DriverId::eAmdProprietary)) {
m_warptile_mmq = m_warptile_mmq_int = { 256, 64, 64, 32, 16, 16, 2, 2, 2, 1, 16 };
m_warptile_mmqid = m_warptile_mmqid_int = { 256, 64, 64, 32, 16, 16, 2, 2, 2, 1, 16 };
}
l_mmq_wg_denoms = l_wg_denoms = {128, 128, 1 };
m_mmq_wg_denoms = m_wg_denoms = { 64, 64, 1 };
s_mmq_wg_denoms = s_wg_denoms = { 32, 32, 1 };
l_align = 128;
m_align = 64;
s_align = 32;
if ((device->vendor_id == VK_VENDOR_ID_INTEL) && (device->driver_id == vk::DriverId::eIntelProprietaryWindows)) {
else if ((device->vendor_id == VK_VENDOR_ID_INTEL) && (device->driver_id == vk::DriverId::eIntelProprietaryWindows)) {
if (device->coopmat_support && device->architecture == INTEL_XE2) {
// Xe2/Xe3 with coopmat enabled - warptile performance tuning
m_warptile = { 512, 128, 128, 16, subgroup_size_8, 32, 2, tm_m, tn_m, tk_m, subgroup_size_8 };