[PATCH] [Builtins] Implement half-precision conversions.
Stephen Canon
scanon at apple.com
Tue May 12 03:03:29 PDT 2015
Looks reasonable to me. Someone else will need to answer your question about gnu_*_ieee aliases.
– Steve
> On May 11, 2015, at 6:39 PM, Ahmed Bougacha <ahmed.bougacha at gmail.com> wrote:
>
> Hi scanon,
>
> Mostly uninteresting, except:
>
> - in __extendXfYf2, when checking if the number is normal, the current code relies on the unsignedness of src_rep_t, which is fine. Except that when sizeof(src_rep_t) < sizeof(int), the result gets promoted to int, the signedness of which breaks the comparison. I added an explicit cast to avoid that; it shouldn't affect other types.
>
> - we can't just pass __fp16 around (you even get a clang error), so I made uint16_t explicit.
>
> - is there a good way to define aliases (for the gnu_*_ieee names) ? Using the attribute, on darwin, clang complains about only weak aliases being supported, so I went with the simple def.
>
>
> I also need to finish writing unit tests in compiler-rt proper (I paused when I realized lit+cmake isn't supported). In the meantime, opinions welcome.
>
> http://reviews.llvm.org/D9693
>
> Files:
> lib/builtins/CMakeLists.txt
> lib/builtins/extendhfsf2.c
> lib/builtins/fp_extend.h
> lib/builtins/fp_extend_impl.inc
> lib/builtins/fp_trunc.h
> lib/builtins/truncdfhf2.c
> lib/builtins/truncsfhf2.c
>
> EMAIL PREFERENCES
> http://reviews.llvm.org/settings/panel/emailpreferences/
> <D9693.25541.patch>
More information about the llvm-commits
mailing list