[PATCH] D112108: [InstCombine] Fold `(a & ~b) & ~c` to `a & ~(b | c)`

[Sanjay Patel via Phabricator via llvm-commits]

spatel added a comment.

In D112108#3076632 <https://reviews.llvm.org/D112108#3076632>, @rampitec wrote:

> In D112108#3075918 <https://reviews.llvm.org/D112108#3075918>, @rampitec wrote:
>
>> In D112108#3074952 <https://reviews.llvm.org/D112108#3074952>, @spatel wrote:
>>
>>> This seems like another short-coming of the reassociation pass, but I think it's ok to deal with the minimal case here.
>>> For example (if I'm seeing it correctly), this test still won't change:
>>>
>>>   define i4 @src(i4 %a, i4 %b, i4 %c, i4 %d) {
>>>     %notb = xor i4 %b, -1
>>>     %notc = xor i4 %c, -1
>>>     %and1 = and i4 %a, %notb
>>>     %and2 = and i4 %and1, %d
>>>     %and3 = and i4 %and2, %notc
>>>     ret i4 %and3
>>>   }
>>>
>>> https://alive2.llvm.org/ce/z/_T8ZhP
>>
>> Yes, this test does not change.
>
> Actually it is transformed too. It does not with `opt -instcombine` but does with `opt -reassociate -instcombine` or just `opt -O3`.

Thanks, that's good news. If you want to make sure that combination of transforms doesn't break invisibly, you could add a test like that for -O{1,2,3} runs to /test/Transforms/PhaseOrdering..


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D112108/new/

https://reviews.llvm.org/D112108