llvm-project

0a5012f0 - [X86] Allow AVX512 rotate intrinsics to be used in constexpr (#157652)

[X86] Allow AVX512 rotate intrinsics to be used in constexpr (#157652) Now that they wrap the __builtin_elementwise_fshl/fshr builtin intrinsics this is pretty trivial. Another step towards #153152 - just VBMI2 double shifts remaining