[llvm-dev] [VSXFMAMutate] OldFMAReg may be wrongly rewritten

[Tim Shen via llvm-dev]

On Tue, Mar 22, 2016 at 5:13 PM Eric Christopher wrote:

> I think we can probably go ahead and throw this up on Phabricator for
> review. I'd probably bring in Matthias for review as well.
>
> Thanks!
>
> -eric
>

(Follow up on the discuss in IRC)

I tried to bridge PPC backend and commuteInstruction, not sure if
correctly, but here I got some non-optimal results: in 12 differences, 8 of
them are improvements, but 3 of them are regressions.

If I implemented commuteInstruction correctly, and current two-address
isn't picking up all of them, I guess my pass is still worth trying (it
catches all these cases), as long as it's ultimately "melt into"
two-address ?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20160325/dcc19e11/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bridge-no-commit.diff
Type: application/octet-stream
Size: 10238 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20160325/dcc19e11/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: result.diff
Type: application/octet-stream
Size: 2298 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20160325/dcc19e11/attachment-0001.obj>