[PATCH] D8688: Update MergedLoadStoreMotion to use MemorySSA

[Daniel Berlin via llvm-commits]

On Mon, Jun 20, 2016 at 12:34 PM, Geoff Berry <gberry at codeaurora.org> wrote:

> gberry added a comment.
>
> Would it make sense to split out the DFS numbering code motion barrier
> code into a separate change?
>
>
I'm happy to commit it separately if anyone cares.
I'll be honest; I don't actually  care enough about the pass to want to
speed it up outside of memoryssa.
While the DFS numbering removes an N^3'ness added to the existing code, the
code is N^3 in so many places it's hard to keep track :)
On the other hand, it's necessary  for memoryssa because memoryssa does not
normally walk the blocks at all

.

> ================
> Comment at: lib/Transforms/Scalar/MergedLoadStoreMotion.cpp:977
> @@ +976,3 @@
> +///
> +/// \brief True when the store is not downwards explosed to block \p End
> +
> ----------------
> Typo: "explosed".
> Also, there is no 'End' parameter as referenced in the comment.
>
>
> http://reviews.llvm.org/D8688
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160620/86190e22/attachment.html>