[PATCH] D12046: AMDGPU: Use DFS to avoid second loop over function

[Daniel Berlin via llvm-commits]

The comment means space efficient, in terms of set reuse of existing sets.
Leaving it is likely to just confuse future readers
:)

On Fri, Aug 14, 2015, 8:45 PM Matt Arsenault <Matthew.Arsenault at amd.com>
wrote:

> arsenm added a comment.
>
> In http://reviews.llvm.org/D12046#225002, @dberlin wrote:
>
> > Do you actually use Visited outside of the iterator (It's not visible
> >  in this patch if you do)?
> >  If not, you should just use depth_first, not depth_first_ext.
>
>
> No. I thought might have a use for it, but this doesn't use it yet. The
> comment on it also says "This iterator stores the 'visited' set in an
> external set, which allows it to be more efficient, and allows external
> clients to use the set for other purposes." so I figured if it's faster to
> just leave it
>
>
> http://reviews.llvm.org/D12046
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150821/fbfedc17/attachment.html>