pengfei added inline comments.
================ Comment at: llvm/lib/Support/X86TargetParser.cpp:531 constexpr FeatureBitset ImpliedFeaturesSSSE3 = FeatureSSE3; constexpr FeatureBitset ImpliedFeaturesSSE4_1 = FeatureSSSE3; constexpr FeatureBitset ImpliedFeaturesSSE4_2 = FeatureSSE4_1; ---------------- tianqing wrote: > pengfei wrote: > > hjl.tools wrote: > > > pengfei wrote: > > > > Can we let `ImpliedFeaturesSSE4_1 = FeatureSSSE3 | FeaturesCRC32` so > > > > that we don't need to add `crc32` on sse4.1 and above? > > > SSE4.1 implies CRC32. But CRC32 shouldn't imply SSE4.1. > > Yes. The constexpr here means `FeaturesSSE4_1` implies both `FeatureSSSE3` > > and `FeaturesCRC32`. > CRC32 was added in SSE4.2. > > In LLVM this implication relationship is bidirectional, that is: > > * -msse4.2 implies -mcrc32 > * -mcrc32 doesn't implies -msse4.2. > * -mno-sse4.2 doesn't implies -mno-crc32. > * But -mno-crc32 also implies -mno-sse4.2. Sorry, I mistook SSE4.1 with SSE4.2. I meant to `constexpr FeatureBitset ImpliedFeaturesSSE4_2 = FeatureSSE4_1 | FeaturesCRC32;` then. I see you make "-msse4.2 implies -mcrc32" by FE. Changing here should make it implies in backend, so that you don't need to explicitly add crc32 in LLVM tests. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D105462/new/ https://reviews.llvm.org/D105462 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
