[llvm] [AMDGPU] Change default loop alignment (PR #155343)

[via llvm-commits]

hjagasiaAMD wrote:

> > > You also need to factor the idea that an inner loop may be nested 10 levels deep, which is not unusual for MIOpen. Then the inner one tends to be small.
> > 
> > 
> > Can you pls elaborate on this. All small loops up to 3 cache lines (192 bytes) are aligned, nested or not. Innermost Loops larger than 3 cache lines are also aligned.
> 
> When you have a deep loop nest and align the inner loop it automatically grows the size of all outer loops.


Yes true, we are already doing that though for loops between 1-3 cache lines

https://github.com/llvm/llvm-project/pull/155343