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

[Dinar Temirbulatov via Phabricator via llvm-commits]

dtemirbulatov marked 7 inline comments as done.
dtemirbulatov added inline comments.


================
Comment at: lib/Transforms/Vectorize/SLPVectorizer.cpp:4323
+void BoUpSLP::removeFromScheduling(BlockScheduling *BS,
+                                   ArrayRef<int> Oprations) {
+  bool Removed = false;
----------------
xbolva00 wrote:
> Operations?
Fixed, Thanks.


================
Comment at: test/Transforms/SLPVectorizer/X86/intrinsic.ll:150
 ; CHECK-NEXT:    [[TMP6:%.*]] = call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> [[TMP5]], i1 true)
-; CHECK-NEXT:    [[TMP7:%.*]] = bitcast i32* [[C:%.*]] to <4 x i32>*
-; CHECK-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* [[TMP7]], align 4
+; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <4 x i32> [[TMP6]], i32 0
+; CHECK-NEXT:    [[TMP8:%.*]] = insertelement <4 x i32> undef, i32 [[TMP7]], i32 0
----------------
ABataev wrote:
> This does not look better
Done, Thanks.


================
Comment at: test/Transforms/SLPVectorizer/X86/long_chains.ll:12
 ; CHECK-NEXT:  entry:
-; CHECK-NEXT:    [[TMP0:%.*]] = bitcast i8* [[B:%.*]] to <2 x i8>*
-; CHECK-NEXT:    [[TMP1:%.*]] = load <2 x i8>, <2 x i8>* [[TMP0]], align 1
-; CHECK-NEXT:    [[TMP2:%.*]] = add <2 x i8> <i8 3, i8 3>, [[TMP1]]
-; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x i8> [[TMP2]], i32 0
-; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <2 x i8> undef, i8 [[TMP3]], i32 0
-; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <2 x i8> [[TMP2]], i32 1
-; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <2 x i8> [[TMP4]], i8 [[TMP5]], i32 1
-; CHECK-NEXT:    [[TMP7:%.*]] = sitofp <2 x i8> [[TMP6]] to <2 x double>
+; CHECK-NEXT:    [[TMP0:%.*]] = load i8, i8* [[B:%.*]], align 1
+; CHECK-NEXT:    [[ARRAYIDX1:%.*]] = getelementptr inbounds i8, i8* [[B]], i64 1
----------------
ABataev wrote:
> Again, are you sure this one is better than it was before?
Done, Thanks,


================
Comment at: test/Transforms/SLPVectorizer/X86/slp-throttle.ll:1
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -slp-vectorizer -S -mtriple=x86_64-unknown-linux-gnu -mcpu=bdver2 < %s | FileCheck %s
----------------
ABataev wrote:
> Do you really need such a big test? Could you simplify it?
Done, Thanks.


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

https://reviews.llvm.org/D57779