<div dir="auto">Woops, I meant to say:<div dir="auto"><br></div><div dir="auto">Binary size isn't the only factor. I'm not too concerned about an increase in size here if the performance increase pulls the weight.</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri., Jun. 14, 2019, 1:00 a.m. Eric Fiselier via Phabricator, <<a href="mailto:reviews@reviews.llvm.org">reviews@reviews.llvm.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">EricWF added a comment.<br>
<br>
Binary size isn't<br>
<br>
<br>
<br>
================<br>
Comment at: benchmarks/algorithms.merge.bench.cpp:1<br>
+#include <array><br>
+#include <string><br>
----------------<br>
Missing license header.<br>
<br>
<br>
================<br>
Comment at: benchmarks/algorithms.merge.bench.cpp:87<br>
+    auto Data = TestType::generateInput(Quantity);<br>
+    using VecT = decltype(Data);<br>
+<br>
----------------<br>
I know this doesn't affect the timing (or shouldn't), but it's useful to split out the prologue of a benchmark so it's easier to inspect the assembly the loop generates.<br>
<br>
Also, if you could provide some snippits or analysis about what changed in the assembly that's giving the performance win.<br>
<br>
<br>
CHANGES SINCE LAST ACTION<br>
  <a href="https://reviews.llvm.org/D63063/new/" rel="noreferrer noreferrer" target="_blank">https://reviews.llvm.org/D63063/new/</a><br>
<br>
<a href="https://reviews.llvm.org/D63063" rel="noreferrer noreferrer" target="_blank">https://reviews.llvm.org/D63063</a><br>
<br>
<br>
<br>
</blockquote></div>