<div dir="ltr">Hi,<div><br></div><div>I agree with Chandler's comment in the review thread. I'm actively working on a patch at the moment (it has a couple of correctness issues still to resolve) to do this - identify operations that could be truncated and performed in a narrower type, and update the cost model to understand this.</div><div><br></div><div>As far as I can see, what I have is a superset of your patch.</div><div><br></div><div>Cheers,</div><div><br></div><div>James</div></div><br><div class="gmail_quote"><div dir="ltr">On Wed, 23 Sep 2015 at 12:51 <a href="mailto:hfinkel@anl.gov">hfinkel@anl.gov</a> via llvm-commits <<a href="mailto:llvm-commits@lists.llvm.org">llvm-commits@lists.llvm.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">hfinkel added a comment.<br>
<br>
Have you run LLVM's test suite with this turned on? Are there any significant performance changes? [I'm happy for this to go in, given that it's disabled by default, even if there are regressions to fix, but I'd like to know where we stand].<br>
<br>
.<br>
<br>
<br>
================<br>
Comment at: lib/Transforms/Vectorize/LoopVectorize.cpp:4585<br>
@@ +4584,3 @@<br>
+ // Collect all viable vectorization factors.<br>
+ SmallVector<unsigned, 4> VFs;<br>
+ unsigned NewMaxVectorSize = WidestRegister / SmallestType;<br>
----------------<br>
I'd make this 8 instead of 4 (we might have 7 VF for 8-bit integers in AVX-512, for example).<br>
<br>
<br>
<a href="http://reviews.llvm.org/D8943" rel="noreferrer" target="_blank">http://reviews.llvm.org/D8943</a><br>
<br>
<br>
<br>
_______________________________________________<br>
llvm-commits mailing list<br>
<a href="mailto:llvm-commits@lists.llvm.org" target="_blank">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/mailman/listinfo/llvm-commits</a><br>
</blockquote></div>