[llvm-commits] [PATCH] Promote all possible memory to registers

[Chris Lattner]

I don't think this will help.  From the description, this optimization falls out of GVN using memdep.  Splitting scalar and memory GVN apart from each other isn't practical: each unblocks important cse's in the other.

-Chris

On Oct 24, 2010, at 2:41 PM, Duncan Sands <baldrick at free.fr> wrote:

> Hi Chris,
> 
>> I haven't looked at the patch yet, but how is this different than the standard GVN pass?
> 
> while GVN surely does this too (what doesn't it do? :) ), maybe it is good to
> factorize this functionality out of GVN, to shrink and simplify it.
> 
> ciao,
> 
> Duncan.
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits