[PATCH] D13996: Fix SLPVectorizer commutativity reordering

[Mehdi AMINI via llvm-commits]

joker.eph added inline comments.

================
Comment at: test/Transforms/SLPVectorizer/X86/commutativity.ll:1
@@ +1,2 @@
+; RUN: opt -slp-vectorizer %s -S | FileCheck %s
+
----------------
mzolotukhin wrote:
> Should we do `opt < %s` instead of `opt %s` to avoid overwriting the input file?
I don't think it would ever overwrite the input file unless you force it to do so explicitly. But Duncan had a point: the ModuleID would be patch sensitive without the redirection.

================
Comment at: test/Transforms/SLPVectorizer/X86/commutativity.ll:59-60
@@ +58,4 @@
+
+; Check that we correctly detects a splat/broadcast by levaraging the
+; commutativity propriety of `xor`.
+
----------------
mzolotukhin wrote:
> Just noticed several typos here (and in the same comment in the previous test):
> s/detects/detect/
> s/levaraging/leveraging/
> s/propriety/property/
Thanks, will update.


http://reviews.llvm.org/D13996