AMD General

Hi

> -----Original Message-----
> From: Jiang, Haochen <[email protected]>
> Sent: Monday, July 6, 2026 8:42 AM
> To: Kumar, Venkataramanan <[email protected]>; gcc-
> [email protected]
> Cc: Liu, Hongtao <[email protected]>; [email protected]; Sharma,
> Dipesh <[email protected]>
> Subject: RE: [PATCH 1/7] Initial support for ACEv1
>
> Caution: This message originated from an External Source. Use proper caution
> when opening attachments, clicking links, or responding.
>
>
> > From: Kumar, Venkataramanan <[email protected]>
> > Sent: Monday, July 6, 2026 1:19 AM
> >
> > > From: Haochen Jiang <[email protected]>
> > > Sent: Thursday, July 2, 2026 12:19 PM
> > >
> > > +  /* Get Advanced Features at level 0x1d (eax = 0x1d).
> > > +     ACE check must be put after AVX10 check to get AVX10 features.
> > > +     TODO: Change the condition after AVX10V1AUX is added.  */  if
> >
> > As per the Specification
> > To detect full ACE v1 support, software should verify:
> > 1. (AVX10.1 and AVX10_V1_AUX) or AVX10.2 2. AVX10_V2_AUX 3. ACE 4.
> > ACE_VSN >= 1 5. XCR0[20,18:17] = 0b111 (XSAVE state enabled for tile +
> > BSR) 6. XCR0[7:5] = 0b111 (AVX-512 state) 7. CR4.OSXSAVE = 1
> >
> > Here the check says AVX10.2 is mandatory.
> >   Can u comment on that TODO: clearly saying version >= 2 to be
> > replaced with ((version >= 1 && avx10v1ux_set) || (version >= 2)).
>
> The comment now is exactly explaining the same issue as this condition?

I did not understand the comment.  But fine to leave it as it is .

>
> >
> > > + (version >= 2 && avx10v2aux_set && ace_set && max_cpuid_level >=
> > 0x1d)
> > > +    {
> > > +      __cpuid_count (0x1d, 0, eax, ebx, ecx, edx);
> > > +      if (eax == 2)
> > > +       {
> > > +         __cpuid_count (0x1d, 2, eax, ebx, ecx, edx);
> > > +         version = eax & 0xff;
> > > +         switch (version)
> > > +           {
> > > +           case 1:
> > > +             set_feature (FEATURE_ACEV1);
> > > +             break;
> > > +           default:
> >
> > Can we error out instead of defaulting to FEATURE_ACEV1 ??
>
> We should not error out this since this will lead to new HW running on old
> GCCs error out since they will have newer versions and drop to default.

Ok.

>
> > > diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi index
> > > 2f927603f38..32595a934a8 100644
> > > --- a/gcc/doc/invoke.texi
> > > +++ b/gcc/doc/invoke.texi
> > > @@ -1551,7 +1551,7 @@ See RS/6000 and PowerPC Options.
> > >  -mavxneconvert  -mcmpccxadd  -mamx-fp16  -mprefetchi  -mraoint  -
> > mamx-
> > > complex  -mavxvnniint16  -msm3  -msha512  -msm4  -mapxf  -musermsr
> > > -
> > > mavx10.1  -mavx10.2  -mamx-avx512  -mmovrs  -mamx-movrs --
> > mavx512bmm  -
> > > mavx10v2aux
> > > +-mavx512bmm  -mavx10v2aux  -macev1
> > Only -macev1 is extra right?? .
>
> Yes. As mentioned in an upcoming letter of the cover letter, the patch is and
> must based on AVX10V2AUX due to feature imply. That is why avx10v2aux is
> already there.

Ok.

Regards,
Venkat.
>
> Thx,
> Haochen

Reply via email to