Feature/avx512 excess#43
Conversation
| #define PIXIE_AVX512VBMI_SUPPORT | ||
| #endif | ||
|
|
||
| #if defined(__GFNI__) && defined(PIXIE_AVX512_SUPPORT) && \ |
There was a problem hiding this comment.
WARNING: PIXIE_AVX512GFNI_SUPPORT can enable a code path that requires VBMI intrinsics
The excess_interleave_masks_gfni implementation uses _mm512_permutexvar_epi8, which requires AVX-512 VBMI. This macro gate currently checks __GFNI__ but not __AVX512VBMI__, so builds with GFNI enabled but VBMI disabled can fail to compile. Please include __AVX512VBMI__ in the feature gate (or avoid VBMI-only intrinsics in this path).
Code Review SummaryStatus: No Issues Found | Recommendation: Merge Files Reviewed (1 files)
Reviewed by gpt-5.3-codex · 354,413 tokens |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #43 +/- ##
==========================================
+ Coverage 87.29% 87.40% +0.10%
==========================================
Files 12 12
Lines 3118 3136 +18
Branches 612 607 -5
==========================================
+ Hits 2722 2741 +19
+ Misses 260 259 -1
Partials 136 136
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Implementation of excess via AVX-512