[PATCH] D150388: [CodeGen]Allow targets to use target specific COPY instructions for live range splitting

Matt Arsenault via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Jul 17 17:01:34 PDT 2023


arsenm added a comment.

In D150388#4508264 <https://reviews.llvm.org/D150388#4508264>, @arsenm wrote:

> In D150388#4506932 <https://reviews.llvm.org/D150388#4506932>, @vitalybuka wrote:
>
>> In D150388#4504387 <https://reviews.llvm.org/D150388#4504387>, @arsenm wrote:
>>
>>> In D150388#4504380 <https://reviews.llvm.org/D150388#4504380>, @vitalybuka wrote:
>>>
>>>> It's still regression. Unless someone has a patch ready to land, we need to revert this.
>>>
>>> I'd ask that you provide a reproducer, and perform the experiment I mentioned. If we can just disable the expanded isCopyInstr identification, it will be a lot less painful
>>
>> I am running creduce for ubsan case whole weekend, still quite large. I can try the experiment if you can provide a draft patch.
>
>
>
> In D150388#4506932 <https://reviews.llvm.org/D150388#4506932>, @vitalybuka wrote:
>
>> In D150388#4504387 <https://reviews.llvm.org/D150388#4504387>, @arsenm wrote:
>>
>>> In D150388#4504380 <https://reviews.llvm.org/D150388#4504380>, @vitalybuka wrote:
>>>
>>>> It's still regression. Unless someone has a patch ready to land, we need to revert this.
>>>
>>> I'd ask that you provide a reproducer, and perform the experiment I mentioned. If we can just disable the expanded isCopyInstr identification, it will be a lot less painful
>>
>> I am running creduce for ubsan case whole weekend, still quite large. I can try the experiment if you can provide a draft patch.
>
> https://reviews.llvm.org/differential/diff/541268/ is the draft
>
> There's something wrong elsewhere even if this fixes it, but I can't see what else would change anything

I just spotted a bug, isCopyOfBundle isn't using the proper result from isCopyInstr. Copy bundles are niche enough that I'd be surprised if they help your case (x86 I presume?)


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D150388/new/

https://reviews.llvm.org/D150388



More information about the llvm-commits mailing list