Skip to content

Persistent SDPA kernel #608

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
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Persistent SDPA kernel #608

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
252b0d1
Persistent SDPA kernel
wuxun-zhang Nov 4, 2025
93ec5c8
update tile scheduler & add runtime check
wuxun-zhang Nov 4, 2025
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
3 changes: 2 additions & 1 deletion applications/flash_attention_v2/collective/xe_fmha_fwd_mainloop.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -171,6 +171,7 @@ struct FMHAFwdMainloop<XeDefault<Stages>, CausalMask_,
QVCoord blk_qv, // WG tile indices: (Q,V)
int blk_k0, // K block range: [K0,K1)
int blk_k1,
int total_blk, // Total # of K blocks
int thr_id) { // Work-item ID

using namespace sycl::ext::oneapi::this_work_item;
Expand Down Expand Up @@ -289,7 +290,7 @@ struct FMHAFwdMainloop<XeDefault<Stages>, CausalMask_,
prefetch(prefetch_v, pVgV(_,_,_,K));

/* k masking for remainder tiles */
if (check_remainder_k && K == blk_k1 - 1) {
if (check_remainder_k && K == total_blk - 1) {
FragSRow k_rem_mask;
int k = get<0>(tKgK(0,0,0,K,0)) + get_sub_group().get_local_id()[0];
CUTLASS_PRAGMA_UNROLL
Expand Down
Loading
Loading