[PATCH] D74185: Revert the revert of vectorization commits

[Alexey Bataev via Phabricator via llvm-commits]

ABataev added a comment.

In D74185#1864077 <https://reviews.llvm.org/D74185#1864077>, @george.karpenkov wrote:

> @ABataev  Please check out the runnable reproducer above. It clearly demonstrates a miscompile: if the pass is correct, same IR input should give identical output (barring undefined behavior, but we use sanitizers to check for that).


Runnable reproducer requires some prerequisites. It would be good if you could provide a simpler one.

> 
> 
>> Also, reverting is not productive, better to try to find a proper fix.
> 
> Miscompiles should be reverted just based on incorrect behavior.

I'm not sure that this patch is the real cause of the problem you have. Maybe, it triggers some other issues in the backend. But it very hard to say currently if it is so or not.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74185/new/

https://reviews.llvm.org/D74185