[PATCH] D130964: [X86][BF16] Enable __bf16 for x86 targets.
Benjamin Kramer via Phabricator via cfe-commits
cfe-commits at lists.llvm.org
Tue Aug 2 13:04:58 PDT 2022
bkramer added a comment.
In D130964#3694473 <https://reviews.llvm.org/D130964#3694473>, @rjmccall wrote:
> How are you actually implementing `__bf16` on these targets? There isn't even hardware support for conversions.
`bf16` -> `float` is really just a bit shift. The other direction gets lowered to a libcall, compiler-rt has a conversion function with proper rounding. I added some support to make the backend promote all other arithmetic to float, but I think that's only enabled on x86 so far.
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
More information about the cfe-commits