<div dir="ltr">In terms of code size, for speccpu2006int, the text size change is listed below:<div><br></div><div>
<table cellspacing="0" border="0">
<colgroup span="2" width="98"></colgroup>
<tbody><tr>
<td height="30" align="left"><font face="Liberation Sans">xalan</font></td>
<td align="right"><font face="Liberation Sans">0.00%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">astar</font></td>
<td align="right"><font face="Liberation Sans">0.10%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">bzip2</font></td>
<td align="right"><font face="Liberation Sans">0.00%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">gcc</font></td>
<td align="right"><font face="Liberation Sans">0.00%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">gobmk</font></td>
<td align="right"><font face="Liberation Sans">0.02%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">h265</font></td>
<td align="right"><font face="Liberation Sans">0.03%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">hmmer</font></td>
<td align="right"><font face="Liberation Sans">0.12%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">libquantum</font></td>
<td align="right"><font face="Liberation Sans">0.00%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">mcf</font></td>
<td align="right"><font face="Liberation Sans">0.00%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">omnetpp</font></td>
<td align="right"><font face="Liberation Sans">0.04%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">perlbench</font></td>
<td align="right"><font face="Liberation Sans">0.00%</font></td>
</tr>
<tr>
<td height="30" align="left"><font face="Liberation Sans">sjeng</font></td>
<td align="right"><font face="Liberation Sans">0.00%</font></td>
</tr>
<tr>
<td height="20" align="left"><font face="Liberation Sans">geomean</font></td>
<td align="right"><font face="Liberation Sans">0.03%</font></td>
</tr>
</tbody></table><div><br></div><div>Dehao</div></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Jun 27, 2016 at 7:50 PM, Dehao Chen <span dir="ltr"><<a href="mailto:danielcdh@gmail.com" target="_blank">danielcdh@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">It's sandybridge. Each benchmark is repeated 3 times and get the medium. But by looking at the raw data, the run-to-run variation are pretty large (2%~5% range across 3 iterations). Thus I would suspect there might be some noise in the result.<span class="HOEnZb"><font color="#888888"><div><br></div><div>Dehao</div></font></span></div><div class="HOEnZb"><div class="h5"><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Jun 27, 2016 at 7:19 PM, Xinliang David Li <span dir="ltr"><<a href="mailto:davidxl@google.com" target="_blank">davidxl@google.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">What platform, Ivybridge ?<span><font color="#888888"><div><br></div><div>David</div></font></span></div><div><div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Jun 27, 2016 at 6:28 PM, Dehao Chen <span dir="ltr"><<a href="mailto:danielcdh@gmail.com" target="_blank">danielcdh@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">danielcdh added a comment.<br>
<br>
I just tested the perf impact of this patch for SPECCPU INT2006 benchmarks:<br>
<br>
400.perlbench 2.21%<br>
401.bzip2 -0.46%<br>
403.gcc 1.59%<br>
429.mcf 1.79%<br>
445.gobmk 1.52%<br>
456.hmmer 1.63%<br>
458.sjeng 5.62%<br>
462.libquantum 1.54%<br>
464.h264ref -1.28%<br>
471.omnetpp 0.00%<br>
473.astar 2.35%<br>
483.xalancbmk -0.27%<br>
overall 1.29%<br>
<br>
About the -Oz issue, I agree that there should be no xor inserted at -Oz. But that's a different issue and should be fixed in a separate patch.<br>
<br>
<br>
<a href="http://reviews.llvm.org/D21560" rel="noreferrer" target="_blank">http://reviews.llvm.org/D21560</a><br>
<br>
<br>
<br>
</blockquote></div><br></div>
</div></div></blockquote></div><br></div>
</div></div></blockquote></div><br></div>