[llvm-commits] [PATCH] A patch to add vector SHL/SRL/SRA support for x86

[Eli Friedman]

On Sun, May 8, 2011 at 11:39 AM, Rotem, Nadav <nadav.rotem at intel.com> wrote:
> Hi Eli,
>
> Do you suggest that we remove PerformShiftCombine altogether ?  Can you think of a scenario which would require us to have this in the DAGCombining phase ?
>
> Would you oppose keeping both translations if I refactor the code to use the same utility function ?

I can't think of any particular reason we need to run this in the
DAGCombine phase; the primary issue, the legalization of BUILD_VECTOR,
isn't an issue anymore.

-Eli