[PATCH] D47214: [InstCombine] Replace bcmp/bcopy with memcmp/memmove

[Dávid Bolvanský via Phabricator via llvm-commits]

xbolva00 added a comment.

In https://reviews.llvm.org/D47214#1108642, @sanjoy wrote:

> In https://reviews.llvm.org/D47214#1108641, @xbolva00 wrote:
>
> > In https://reviews.llvm.org/D47214#1108622, @bkramer wrote:
> >
> > > why?
> >
> >
> > Anyway these functions call mem* functions, so why not call them directly?
>
>
> There are an infinite number of optimizations we can implement, but each one adds more maintenance burden.  We need to make sure new optimizations "carry their weight".
>
> Do you have example programs or benchmarks that measurably speed up or have smaller code size etc. because of this transform?


I have no benchmarks. But changing them to mem-functions allows us to apply another set of transformations in "optimizeMemcmp/Memmove".

Anyway, If you say no, no problem. I will close this patch.


Repository:
  rL LLVM

https://reviews.llvm.org/D47214