Skip to content

[GPU] SDPA micro: use prefill config on integrated GPUs #39316

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

Merged
e-ddykim merged 1 commit into from
Nov 2, 2026
Merged

[GPU] SDPA micro: use prefill config on integrated GPUs #39316

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
10 changes: 6 additions & 4 deletions src/plugins/intel_gpu/src/graph/impls/ocl_v2/sdpa/sdpa_gen_micro.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -723,6 +723,11 @@ sdpa_config_t* choose_config_xehpc(int head_size, int seq, bool thin_q, bool qua
if (seq <= 0 && is_pa) {
return is_prefill ? &xehpc_h128 : &xehpc_h128_pa;
}
// On an integrated part a multi-token call whose key count is only known at run time is a
// prefill over a long context; do not size it as if it had no keys.
if (seq <= 0 && !thin_q && is_integrated) {
seq = std::numeric_limits<int>::max();
}
if (quantized) {
if (thin_q) {
if (is_integrated) {
Expand All @@ -749,11 +754,8 @@ sdpa_config_t* choose_config_xehpc(int head_size, int seq, bool thin_q, bool qua
}
return &xehpc_q_h128;
}
if (is_integrated) {
return &xehpc_q_h128_2nd_integrated;
}
if (thin_q) {
return &xehpc_h128_2nd;
return is_integrated ? &xehpc_q_h128_2nd_integrated : &xehpc_h128_2nd;
}
if (seq <= 32) {
return &xehpc_h128_s32;
Expand Down
Loading