[PATCH] D148362: [SLP]Fix cost estimation for buildvectors with extracts and/or constants.

[Dave Green via Phabricator via llvm-commits]

dmgreen added a comment.

Thanks for looking at this. The performance certainly seems to have improved.



================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:6976
       }
-      Cost -= TTI.getVectorInstrCost(*EE, EE->getVectorOperandType(), CostKind,
-                                     Idx);
+      Cost -= TTI.getScalarizationOverhead(
+          EE->getVectorOperandType(),
----------------
Is this to work around the AArch64 cost model?


================
Comment at: llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp:8770
   DenseSet<Value *> UniqueElements;
   // Iterate in reverse order to consider insert elements with the high cost.
+  for (unsigned I = 0, E = VL.size(); I < E; ++I) {
----------------
This comment looks out of date now.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D148362/new/

https://reviews.llvm.org/D148362