[PATCH 09/10] Support Intel AMX-FP16

[Jiang, Haochen]

> -----Original Message-----
> From: Jan Beulich <jbeulich@suse.com>
> Sent: Tuesday, October 18, 2022 5:24 PM
> To: Cui, Lili <lili.cui@intel.com>
> Cc: hjl.tools@gmail.com; binutils@sourceware.org; Jiang, Haochen
> <haochen.jiang@intel.com>
> Subject: Re: [PATCH 09/10] Support Intel AMX-FP16
> 
> On 18.10.2022 11:01, Cui, Lili wrote:
> >>> +  { "CPU_AMX_FP16_FLAGS",
> >>> +    "CpuAMX_FP16" },
> >>>    { "CPU_IAMCU_FLAGS",
> >>>      "Cpu186|Cpu286|Cpu386|Cpu486|Cpu586|CpuIAMCU" },
> >>>    { "CPU_ADX_FLAGS",
> >>
> >> Can you please insert next to the other similar AMX entries? Seeing the
> flaw
> >> here, I'll be making a patch to address the lack of CPU_AMX_TILE_FLAGS
> in
> >> the similar pre-existing entries. When you move the insertion, it'll be
> easier
> >> to keep things in sync.
> > Done.
> >
> >>> -    "CpuAMX_TILE|CpuAMX_INT8|CpuAMX_BF16" },
> >>> +    "CpuAMX_TILE|CpuAMX_INT8|CpuAMX_BF16|CpuAMX_FP16" },
> >>>    { "CPU_ANY_MSRLIST_FLAGS",
> >>>      "CpuMSRLIST" },
> >>> +  { "CPU_ANY_AMX_FP16_FLAGS",
> >>> +    "CpuAMX_FP16" },
> >>>  };
> >>
> >> Same here then.
> > Done.
> 
> I guess my comment here was a little misleading (I'm sorry for that), in
> that in addition I was expecting you to consider the comment regarding
> the need for the *_ANY_* constants that I did give for several of the
> patches in this series. I think the question applies here as well: Are
> there dependent features known to appear? If not, there's no need for the
> extra constant.
> 
> > --- a/gas/config/tc-i386.c
> > +++ b/gas/config/tc-i386.c
> > @@ -1101,6 +1101,7 @@ static const arch_entry cpu_arch[] =
> >    SUBARCH (raoint, RAOINT, ANY_RAOINT, false),
> >    SUBARCH (wrmsrns, WRMSRNS, ANY_WRMSRNS, false),
> >    SUBARCH (msrlist, MSRLIST, ANY_MSRLIST, false),
> > +  SUBARCH (amx_fp16, AMX_FP16, ANY_AMX_FP16, false),
> >  };
> 
> This might also benefit from moving up, but I'm not going to insist.
> 
> > --- a/gas/doc/c-i386.texi
> > +++ b/gas/doc/c-i386.texi
> > @@ -220,6 +220,7 @@ accept various extension mnemonics.  For example,
> >  @code{raoint},
> >  @code{wrmsrns},
> >  @code{msrlist},
> > +@code{amx_fp16},
> >  @code{noavx512f},
> >  @code{noavx512cd},
> >  @code{noavx512er},
> > @@ -247,6 +248,7 @@ accept various extension mnemonics.  For example,
> >  @code{noraoint},
> >  @code{nowrmsrns},
> >  @code{nomsrlist},
> > +@code{noamx_fp16},
> >  @code{noenqcmd},
> >  @code{noserialize},
> >  @code{notsxldtrk},
> 
> There are no @code{no...} entries here anymore, as of earlier today. Hence
> no new ones should appear (and you need to re-base in any event).

I will do the final rebase for all the patches in the end to solve texi problem.
Since there are several people are involved in this series of patches, I just keep
the branch freeze at Friday's main trunk (the day we release all the patches),
which will help keep the branch stable under everyone's change. But if the patch
need main trunk fix out of texi, I will remind them and do the rebase before we
send out the patch. I suppose there has been patches related to table folding and
AMX prereq. 

BRs,
Haochen
> 
> Jan