Skip to content

UPSTREAM PR #21431: vulkan: Tweak Xe2 warptile configuration #1341

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
loci-dev wants to merge 1 commit into
base: main
Choose a base branch
from
Open

UPSTREAM PR #21431: vulkan: Tweak Xe2 warptile configuration #1341

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 9 additions & 8 deletions ggml/src/ggml-vulkan/ggml-vulkan.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3287,6 +3287,13 @@ 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, s_warptile_wm, 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 };
Expand All @@ -3298,17 +3305,11 @@ static void ggml_vk_load_shaders(vk_device& device) {
l_warptile_mmq_int_k = { 256, 128, 128, 32, subgroup_size_16, 64, 1, 4, 2, 1, subgroup_size_16 };
} else if (device->vendor_id == VK_VENDOR_ID_INTEL && device->coopmat_support && device->architecture == INTEL_XE2) {
// Xe2/Xe3 with coopmat enabled - warptile performance tuning
l_warptile = { 512, 128, 128, 16, subgroup_size_8, 32, 2, tm_m, tn_m, tk_m, subgroup_size_8 };
l_warptile = { 512, 128, 64, 16, subgroup_size_8, 16, 2, tm_m, tn_m, tk_m, subgroup_size_8 };
l_wg_denoms = {128, 64, 1};
l_warptile_mmq = { 512, 128, 128, 32, subgroup_size_8, 32, 2, tm_m, tn_m, tk_m, subgroup_size_8 };
}

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;

for (uint32_t i = 0; i < GGML_TYPE_COUNT; ++i) {
ggml_type t = (ggml_type)i;
// Disable medium and large matrix multiplication if not enough shared memory is available
Expand Down
Loading