Skip to content

Fix CPU FlashAttention disabled on Linux/aarch64 by detecting L2 cache via cpuinfo #29621

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
Sammy-Dabbas wants to merge 2 commits into
base: main
Choose a base branch
from
+20 −1
Open

Fix CPU FlashAttention disabled on Linux/aarch64 by detecting L2 cache via cpuinfo #29621

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
047ab5c
Fix CPU FlashAttention disabled on Linux/aarch64 by detecting L2 cach…
Sammy-Dabbas Jul 8, 2026
187fa19
Prefer sysconf L2 cache size and fall back to cpuinfo only when it fails
Sammy-Dabbas Jul 10, 2026
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
13 changes: 13 additions & 0 deletions onnxruntime/core/platform/posix/env.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -302,6 +302,19 @@ class PosixEnv : public Env {
}

int GetL2CacheSize() const override {
#ifdef ORT_USE_CPUINFO
Comment thread
tianleiwu marked this conversation as resolved.
Outdated
// Prefer cpuinfo for detection. glibc's aarch64 sysconf backend does not
// implement the cache-size queries, so sysconf(_SC_LEVEL2_CACHE_SIZE)
// returns 0 on Linux/aarch64, which silently disables cache-tiled kernels
// (e.g. CPU FlashAttention in MultiHeadAttention). cpuinfo reads the sizes
// from sysfs cacheinfo and works across architectures.
if (cpuinfo_available_ && cpuinfo_get_l2_caches_count() > 0) {
const auto* l2_cache = cpuinfo_get_l2_cache(0);
if (l2_cache != nullptr && l2_cache->size > 0) {
return static_cast<int>(l2_cache->size);
}
}
Comment on lines +318 to +323
#endif // ORT_USE_CPUINFO
#ifdef _SC_LEVEL2_CACHE_SIZE
return static_cast<int>(sysconf(_SC_LEVEL2_CACHE_SIZE));
#else
Expand Down
Loading