On Fri, 30 Sept 2022 at 19:38, Jakub Jelinek wrote:
>
> On Fri, Sep 30, 2022 at 06:21:04PM +, Joseph Myers wrote:
> > On Fri, 30 Sep 2022, Jakub Jelinek via Gcc-patches wrote:
> >
> > > What isn't in the patch but I think we'll need to also change are some
> > > minimal set of __builtin_*bf16
On Fri, Sep 30, 2022 at 06:21:04PM +, Joseph Myers wrote:
> On Fri, 30 Sep 2022, Jakub Jelinek via Gcc-patches wrote:
>
> > What isn't in the patch but I think we'll need to also change are some
> > minimal set of __builtin_*bf16 builtins. Seems for _Float16, GCC provides
> > all the
On Fri, 30 Sep 2022, Jakub Jelinek via Gcc-patches wrote:
> What isn't in the patch but I think we'll need to also change are some
> minimal set of __builtin_*bf16 builtins. Seems for _Float16, GCC provides
> all the __builtin_*f16 (and for C/ObjC even with *f16 names), but there is
> no glibc
On Fri, Sep 30, 2022 at 09:49:08AM -0400, Jason Merrill wrote:
> The comment from Apple on the ABI mangling proposal suggests to me that we
> might want to delay enabling C++ std::bfloat16_t (i.e. defining
> __STDCPP_BFLOAT16_T__) until we have that excess precision support?
I saw that comment.
On 9/29/22 11:55, Jakub Jelinek wrote:
Hi!
Here is more complete patch to add std::bfloat16_t support on
x86, AArch64 and (only partially) on ARM 32-bit. No BFmode optabs
are added by the patch, so for binops/unops it extends to SFmode
first and then truncates back to BFmode.
For
Hi!
Here is more complete patch to add std::bfloat16_t support on
x86, AArch64 and (only partially) on ARM 32-bit. No BFmode optabs
are added by the patch, so for binops/unops it extends to SFmode
first and then truncates back to BFmode.
For {HF,SF,DF,XF,TF}mode -> BFmode conversions libgcc has