[PATCH v2] Support Intel AVX10.2 BF16 instructions
Jiang, Haochen
haochen.jiang@intel.com
Thu Nov 14 15:13:54 GMT 2024
> From: Jan Beulich <jbeulich@suse.com>
> Sent: Thursday, November 14, 2024 5:16 PM
>
> On 14.11.2024 06:54, Jiang, Haochen wrote:
> >> From: Jan Beulich <jbeulich@suse.com>
> >> Sent: Thursday, October 31, 2024 4:25 PM
> >>
> >> On 31.10.2024 08:54, Jiang, Haochen wrote:
> >>>
> >
> > The change supposed to be made on mnemonics for BF16 related insts:
> >
> > VSCALEFPBF16 -> VSCALEFNEPBF16:
> > Avoid confusion from legacy PS/PD/PH part.
> >
> > VMINMAXNEPBF16 -> VMINMAXPBF16:
> > The calculation is always exact and no need for NE.
> >
> > VCVTTNEBF162I[,U]BS -> VCVTTBF162I[,U]BS:
> > It is always round toward zero, not an RNE.
>
> I expect you're going to update the patch accordingly then?
Yes, and they will affect the following three patches
(two unsent for now).
For this patch, only VSCALEFPBF16 is in the list.
Thx,
Haochen
More information about the Binutils
mailing list