[PATCH] D57779: [SLP] Add support for throttling.

[Alexey Bataev via Phabricator via llvm-commits]

ABataev added inline comments.


================
Comment at: test/Transforms/SLPVectorizer/X86/unreachable.ll:24
 ; CHECK-NEXT:    [[T10:%.*]] = load i32, i32* [[T9]], align 4
+; CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> undef, i32 [[T4]], i32 0
+; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[T6]], i32 1
----------------
Still the problem with the cost model! I did not look at your patch closely but seems to me you're throwing away the cost of leaf gather nodes from the vectorization tree. Also, seems to me, you don't take into account that for the reduced trees you also need to build the leaf gather nodes and take their cost into account.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D57779/new/

https://reviews.llvm.org/D57779