[PATCH] D134982: [X86] Add support for "light" AVX

[Ilya Tokar via Phabricator via llvm-commits]

TokarIP added a comment.

In D134982#4000705 <https://reviews.llvm.org/D134982#4000705>, @RKSimon wrote:

> In D134982#3999687 <https://reviews.llvm.org/D134982#3999687>, @TokarIP wrote:
>
>> In D134982#3999680 <https://reviews.llvm.org/D134982#3999680>, @lebedev.ri wrote:
>>
>>> I think `LightAVX` is a misnomer. If we want to
>>> always utilize full potential of vector load-store unit,
>>> then the Tuning should say as much.
>>
>> I'll probably expand this to other "light" AVX instructions (like vpcmpeq for memcmp intrinsic) in the future.
>> Also we don't want the full width, 512-bit load/stores still cause some frequency drop on skylake.
>
> Do we have a definitive list of what intel considers "light" 256-bit instructions?

Official? - don't think so. I remember seeing some unofficial lists, but can't find them right now. Considering new cpus released since 2016, retesting would make sense anyway.



================
Comment at: llvm/lib/Target/X86/X86.td:619
+def TuningAllowLightAVX
+    : SubtargetFeature<"allow-light-avx", "AllowLightAVX", "true",
+                       "Enable generation of 256 AVX load/stores even if we prefer 128-bit">;
----------------
RKSimon wrote:
> Maybe rename this "allow-light-256-bit"?
Done for the feature name, should I also rename other mentions of light avx?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D134982/new/

https://reviews.llvm.org/D134982