Rewrite i386 backend as hand-written assembly

Replaces the C fallback (borrowed from PowerPC) with dedicated i386 assembly for all nine kernels, matching the amd64/arm64 convention. Fixes two real bugs found and verified on physical Pentium 4 hardware: scalar FP silently running on x87 instead of SSE2, and fm_int_math calling __udivdi3 on every iteration of the highest-weighted test because the C compiler didn't know its 0xdeadbeef divisor fits in 32 bits. Real-hardware SINGLECORE/MULTICORE scores go from 425/448 to 539/581.
2026-07-18 04:35:03 -05:00
parent 136b417cf8
commit 357e6f0d6d
6 changed files with 1370 additions and 38 deletions
@@ -170,4 +170,5 @@ jobs:
            release/SHA256SUMS \
            --repo "$GITHUB_REPOSITORY" \
            --title "Release $RELEASE_TAG" \
+            --notes "⚠️ **i386 (Pentium 4) support is new in this release and still being validated on real hardware.** i386 benchmark scores should not be considered accurate yet." \
            --generate-notes