Re: [Xen-devel] [PATCH v4] x86/cpuid: AVX-512 Feature Detection

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 05.07.16 at 04:31, <luwei.kang@xxxxxxxxx> wrote:

First of all - please don't top post.

> What about remove the dependency between AVX2 and AVX512F ( AVX2: [AVX512F], 
> ) ?

Yes, that's what I think we want, but we need Andrew's agreement
here.

> From: Jan Beulich [mailto:JBeulich@xxxxxxxx] 
> Sent: Friday, July 1, 2016 3:56 PM
>>>> On 30.06.16 at 07:50, <luwei.kang@xxxxxxxxx> wrote:
>> --- a/xen/tools/gen-cpuid.py
>> +++ b/xen/tools/gen-cpuid.py
>> @@ -243,6 +243,17 @@ def crunch_numbers(state):
>>          # AMD K6-2+ and K6-III processors shipped with 3DNow+, beyond the
>>          # standard 3DNow in the earlier K6 processors.
>>          _3DNOW: [_3DNOWEXT],
>> +
>> +        # AVX2 is an extension to AVX, providing mainly new integer 
>> instructions.
>> +        # In principle, AVX512 only takes YMM register state as a 
>> prerequisite
>> +        # and many AVX2 instructions are extended by AVX512F to 512-bit 
>> forms.
>> +        AVX2: [AVX512F],
> 
> I think this was meant to be "but" or "yet" instead of "and". And then I'm 
> not particularly happy about the asymmetry with AVX (which is not dependent 
> on SSE or any of its successors, despite extending many of them to 256-bit 
> forms). Plus e.g. FMA gets extended too, but isn't being made a dependency. 
> And quite opposite to that, at least some AVX2 instructions get extended only 
> by e.g. AVX512BW.
> Nor are any of the extended forms listed to require other than the
> AVX512* bits checked in the doc.
> 
> Jan




_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel