<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Apr 6, 2015 at 11:26 AM, Sanjoy Das <span dir="ltr"><<a href="mailto:sanjoy@playingwithpointers.com" target="_blank">sanjoy@playingwithpointers.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">> I wouldn't be surprised if InstSimplify were relying on (1) to implement<br>
> some of its optimizations:<br>
> <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Analysis/InstructionSimplify.cpp?revision=233938&view=markup#l2298" target="_blank">http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Analysis/InstructionSimplify.cpp?revision=233938&view=markup#l2298</a><br>
<br>
</span>Yup!  It looks like LLVM really implements (1) and (2) is just a<br>
misleading anecdote.  So I'll change my vote from "remove (1)" to<br>
"remove (2)".  :)<br></blockquote><div><br></div><div>I wrote those optimizations with an understanding that (1) was the one true way.  Would you be apposed to switching to (2)? I have examples where implementing (1) creates real regressions where (2) does not.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<span class="HOEnZb"><font color="#888888"><br>
-- Sanjoy<br>
</font></span></blockquote></div><br></div></div>