<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Thu, May 25, 2017 at 7:51 PM, Davide Italiano via llvm-commits <span dir="ltr"><<a href="mailto:llvm-commits@lists.llvm.org" target="_blank">llvm-commits@lists.llvm.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On Thu, May 25, 2017 at 7:47 PM, Rui Ueyama via llvm-commits<br>
<<a href="mailto:llvm-commits@lists.llvm.org">llvm-commits@lists.llvm.org</a>> wrote:<br>
> Here is the table of your benchmark results. It seems they aren't different<br>
> to me.<br>
><br>
> --threads<br>
><br>
> Index w/o Index<br>
> <a href="tel:3.128589132" value="+13128589132">3.128589132</a> <a href="tel:3.128069636" value="+13128069636">3.128069636</a> 1.000<br>
> 3.302902865 3.294943688 1.002<br>
> 4.842033728 4.852418117 0.997<br>
> 5.031806070 5.007721261 1.004<br>
> 5.061450822 5.067567389 0.998<br>
> 5.223625773 5.213661213 1.001<br>
> 2.777111220 2.770632998 1.002<br>
> 2.870435296 2.876185518 0.998<br>
> 2.041417779 2.043296975 0.999<br>
> 2.120939637 2.121668866 0.999<br>
> 1.543028971 1.551262595 0.994<br>
> 1.556195605 1.559701231 0.997<br>
><br>
> --no-threads<br>
><br>
> Index w/o Index<br>
> 4.648823604 4.634666491 1.003<br>
> 4.656633102 4.656090134 1.000<br>
> 7.298663474 7.288234015 1.001<br>
> 7.310040628 7.303389055 1.000<br>
> 7.508681533 7.496336533 1.001<br>
> 7.517153977 7.512991823 1.000<br>
> 4.089264337 4.097596474 0.997<br>
> 4.082382992 4.089621870 0.998<br>
> 2.202735522 2.210372640 0.996<br>
> 2.205530082 2.216226573 0.995<br>
> 3.156375368 3.165926995 0.996<br>
> 3.158119913 3.166431410 0.997<br>
><br>
<br>
</div></div>Honestly, the improvements/regressions seem just noise to me. If these<br>
are numbers patched/unpatched, I'd say this patch is mostly<br>
performance neutral.<br></blockquote><div><br></div><div>IIRC Rafael's setup was using CPU sets to pin to certain cores. Is it possible that this is preventing Rui's parallelization from fully utilizing all the cores?</div><div>(i.e. testing --threads configuration when pinned to just a single core)</div><div><br></div><div>-- Sean Silva</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>
--<br>
Davide<br>
<br>
"There are no solved problems; there are only problems that are more<br>
or less solved" -- Henri Poincare<br>
</font></span><div class="HOEnZb"><div class="h5">______________________________<wbr>_________________<br>
llvm-commits mailing list<br>
<a href="mailto:llvm-commits@lists.llvm.org">llvm-commits@lists.llvm.org</a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits" rel="noreferrer" target="_blank">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/llvm-commits</a><br>
</div></div></blockquote></div><br></div></div>