SME2 SIMD Loops: loop_216 is ★ under its gemm tolerance; remove loop_032's baseline-sme2
#6
by awu12 - opened
loop_216 passes every workload on M4 under the gemm tolerance it now has (code repo PRs #111, #115), so its SME2 cell is ★.
loop_032 has no SME2 code upstream: under -mcpu=apple-m4 its source falls back to NEON, which streaming mode rejects on M4 (no FEAT_SME_FA64), so the published baseline-sme2 can never run there. This removes it and its sme2-kernel tag; its SME2 cell becomes —.
SME2 column for SIMD Loops: 7 ★ / 16 ✓ / 0 ? / 8 ✗ / 16 —; 31 baseline-sme2 (24 streaming SVE, 7 ZA).
awu12 changed pull request status to closed