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

[Alexey Bataev via Phabricator via llvm-commits]

ABataev added inline comments.


================
Comment at: test/Transforms/SLPVectorizer/X86/extractcost.ll:10
 ; CHECK-NEXT:  entry:
-; CHECK-NEXT:    [[TMP0:%.*]] = insertelement <4 x i32> undef, i32 [[N:%.*]], i32 0
-; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32> [[TMP0]], i32 [[N]], i32 1
-; CHECK-NEXT:    [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 [[N]], i32 2
-; CHECK-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[N]], i32 3
-; CHECK-NEXT:    [[TMP4:%.*]] = mul nsw <4 x i32> [[TMP3]], <i32 5, i32 9, i32 3, i32 10>
-; CHECK-NEXT:    [[TMP5:%.*]] = shl <4 x i32> [[TMP3]], <i32 5, i32 9, i32 3, i32 10>
-; CHECK-NEXT:    [[TMP6:%.*]] = shufflevector <4 x i32> [[TMP4]], <4 x i32> [[TMP5]], <4 x i32> <i32 0, i32 1, i32 6, i32 3>
-; CHECK-NEXT:    [[TMP7:%.*]] = add nsw <4 x i32> <i32 9, i32 9, i32 9, i32 9>, [[TMP6]]
-; CHECK-NEXT:    [[TMP8:%.*]] = bitcast i32* [[A:%.*]] to <4 x i32>*
-; CHECK-NEXT:    store <4 x i32> [[TMP7]], <4 x i32>* [[TMP8]], align 4
-; CHECK-NEXT:    [[TMP9:%.*]] = extractelement <4 x i32> [[TMP7]], i32 0
-; CHECK-NEXT:    [[EXTERNALUSE1:%.*]] = add nsw i32 [[TMP9]], [[M:%.*]]
-; CHECK-NEXT:    [[EXTERNALUSE2:%.*]] = mul nsw i32 [[TMP9]], [[M]]
+; CHECK-NEXT:    [[MUL:%.*]] = mul nsw i32 [[N:%.*]], 5
+; CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[MUL]], 9
----------------
Are you sure that your vectorization result is better than it was before?


================
Comment at: test/Transforms/SLPVectorizer/X86/resched.ll:77
 ; CHECK-NEXT:    [[ARRAYIDX_I_I7_15_I_I:%.*]] = getelementptr inbounds %"struct.std::array", %"struct.std::array"* undef, i64 0, i32 0, i64 15
-; CHECK-NEXT:    [[TMP45:%.*]] = bitcast i8* [[TMP0]] to <16 x i8>*
-; CHECK-NEXT:    store <16 x i8> [[TMP44]], <16 x i8>* [[TMP45]], align 1
+; CHECK-NEXT:    [[TMP17:%.*]] = insertelement <16 x i8> undef, i8 [[CONV_I_I1199]], i32 0
+; CHECK-NEXT:    [[TMP18:%.*]] = insertelement <16 x i8> [[TMP17]], i8 [[CONV_1_I_I]], i32 1
----------------
Same question: are you sure it is better?


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

https://reviews.llvm.org/D57779