[PATCH] D46957: [x86] Lower some trunc + shuffle patterns to vpmov[q|d][b|w]

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added a comment.

Commit the tests with the current codegen so the patch diff shows the improvement?



================
Comment at: lib/Target/X86/X86ISelLowering.cpp:9391
 
+static bool maskContainsSequenceForVPMOV(ArrayRef<int> Mask, bool SwappedOps,
+                                         int delta)
----------------
clang format all of this


================
Comment at: lib/Target/X86/X86ISelLowering.cpp:9395
+  if (SwappedOps) {
+    for (int i = 0; (size_t)i < Mask.size() / delta; ++i) {
+      if ((size_t)Mask[i] != Mask.size() + i * delta)
----------------
Cleanup the for-loop condition to avoid so much evaluation and size_t casts: 
http://llvm.org/docs/CodingStandards.html#don-t-use-default-labels-in-fully-covered-switches-over-enumerations


================
Comment at: lib/Target/X86/X86ISelLowering.cpp:9413
+        return false;
+    }
+  }
----------------
Those for loop braces can be removed?


https://reviews.llvm.org/D46957