<html><head><meta http-equiv="Content-Type" content="text/html charset=windows-1252"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;"><br><div><div>On Mar 26, 2014, at 2:15 PM, Tobias Grosser <<a href="mailto:tobias@grosser.es">tobias@grosser.es</a>> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div style="font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;">On 03/25/2014 03:18 AM, Andrew Trick wrote:<br><blockquote type="cite">Author: atrick<br>Date: Mon Mar 24 21:18:47 2014<br>New Revision: 204699<br><br>URL: <a href="http://llvm.org/viewvc/llvm-project?rev=204699&view=rev">http://llvm.org/viewvc/llvm-project?rev=204699&view=rev</a><br>Log:<br>SLP vectorizer: Don't hoist vector extracts of phis.<br><br>Extracts coming from phis were being hoisted, while all others were<br>sunk to their uses. This was inconsistent and didn't seem to serve a<br>purpose. Changing all extracts to be sunk to uses is a prerequisite<br>for adding block frequency to the SLP vectorizer's cost model.<br><br>I benchmarked the change in isolation (without block frequency). I<br>only saw noise on x86 and some potentially significant improvements on<br>ARM. No major regressions is good enough for me.<br></blockquote><br>Hi Andrew,<br><br>my performance testers spotted a 37% performance regression for fp-convert. In the set of five relevant commits there are only debug info changes, lldb changes and your commit:<br><br><a href="http://llvm.org/perf/db_default/v4/nts/graph?plot.0=34.161.2&highlight_run=23986">http://llvm.org/perf/db_default/v4/nts/graph?plot.0=34.161.2&highlight_run=23986</a><br><br>I am not insisting to investigate this, but wanted to give you at least this information.<br><br>Cheers,<br>Tobias</div></blockquote></div><div><br></div>Thanks. I filed PR19258. When I benchmarked this I saw a slight improvement with fp-convert on sandybridge (2.6s from 2.8s). I don’t see any obvious difference with your builder other than the platform/machine. So I don’t know if this is just a flaky test. It would be great if someone could reproduce and explain what happened.<div><br></div><div><div>-Andy</div><div><br></div></div></body></html>