[llvm] 7aadfc5 - [VectorCombine] Add tests for shuffle binops patterns. NFC
David Green via llvm-commits
llvm-commits at lists.llvm.org
Wed May 4 07:08:10 PDT 2022
Author: David Green
Date: 2022-05-04T15:07:47+01:00
New Revision: 7aadfc509997c1a9078c6adf0ec01a002fceb8c4
URL: https://github.com/llvm/llvm-project/commit/7aadfc509997c1a9078c6adf0ec01a002fceb8c4
DIFF: https://github.com/llvm/llvm-project/commit/7aadfc509997c1a9078c6adf0ec01a002fceb8c4.diff
LOG: [VectorCombine] Add tests for shuffle binops patterns. NFC
Added:
llvm/test/Transforms/VectorCombine/AArch64/select-shuffle.ll
Modified:
Removed:
################################################################################
diff --git a/llvm/test/Transforms/VectorCombine/AArch64/select-shuffle.ll b/llvm/test/Transforms/VectorCombine/AArch64/select-shuffle.ll
new file mode 100644
index 0000000000000..28fac19b465ef
--- /dev/null
+++ b/llvm/test/Transforms/VectorCombine/AArch64/select-shuffle.ll
@@ -0,0 +1,826 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -passes='vector-combine' -S %s | FileCheck %s
+
+target triple = "aarch64-none-eabi"
+
+define <16 x i32> @test1(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test1(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 21, i32 22, i32 7, i32 24, i32 25, i32 10, i32 27, i32 28, i32 13, i32 30, i32 31>
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 21, i32 22, i32 7, i32 24, i32 25, i32 10, i32 27, i32 28, i32 13, i32 30, i32 31>
+ ret <16 x i32> %s3
+}
+
+define i32 @test1_reduce(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test1_reduce(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 21, i32 22, i32 7, i32 24, i32 25, i32 10, i32 27, i32 28, i32 13, i32 30, i32 31>
+; CHECK-NEXT: [[R:%.*]] = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> [[S3]])
+; CHECK-NEXT: ret i32 [[R]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 21, i32 22, i32 7, i32 24, i32 25, i32 10, i32 27, i32 28, i32 13, i32 30, i32 31>
+ %r = call i32 @llvm.vector.reduce.add(<16 x i32> %s3)
+ ret i32 %r
+}
+declare i32 @llvm.vector.reduce.add(<16 x i32>)
+
+define <16 x i32> @test1more(<16 x i32> %a1, <16 x i32> %a2) {
+; CHECK-LABEL: @test1more(
+; CHECK-NEXT: [[X:%.*]] = add nsw <16 x i32> [[A1:%.*]], [[A2:%.*]]
+; CHECK-NEXT: [[Y:%.*]] = sub nsw <16 x i32> [[A1]], [[A2]]
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X]], <16 x i32> [[Y]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 21, i32 22, i32 7, i32 24, i32 25, i32 10, i32 27, i32 28, i32 13, i32 30, i32 31>
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %x = add nsw <16 x i32> %a1, %a2
+ %y = sub nsw <16 x i32> %a1, %a2
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 21, i32 22, i32 7, i32 24, i32 25, i32 10, i32 27, i32 28, i32 13, i32 30, i32 31>
+ ret <16 x i32> %s3
+}
+
+define <16 x i32> @test1mores(<16 x i32> %a1, <16 x i32> %a2) {
+; CHECK-LABEL: @test1mores(
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A1:%.*]], <16 x i32> [[A2:%.*]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[S4:%.*]] = shufflevector <16 x i32> [[A2]], <16 x i32> [[A1]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: [[X:%.*]] = add nsw <16 x i32> [[S3]], [[S4]]
+; CHECK-NEXT: [[Y:%.*]] = sub nsw <16 x i32> [[S3]], [[S4]]
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X]], <16 x i32> [[Y]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S5:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 21, i32 22, i32 7, i32 24, i32 25, i32 10, i32 27, i32 28, i32 13, i32 30, i32 31>
+; CHECK-NEXT: ret <16 x i32> [[S5]]
+;
+ %s3 = shufflevector <16 x i32> %a1, <16 x i32> %a2, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %s4 = shufflevector <16 x i32> %a2, <16 x i32> %a1, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ %x = add nsw <16 x i32> %s3, %s4
+ %y = sub nsw <16 x i32> %s3, %s4
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s5 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 21, i32 22, i32 7, i32 24, i32 25, i32 10, i32 27, i32 28, i32 13, i32 30, i32 31>
+ ret <16 x i32> %s5
+}
+
+
+
+define <16 x i32> @test2_1(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test2_1(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+ ret <16 x i32> %s3
+}
+
+define <16 x i32> @test2_1_ins(<16 x i32> %x1, <16 x i32> %x2) {
+; CHECK-LABEL: @test2_1_ins(
+; CHECK-NEXT: [[X:%.*]] = add nsw <16 x i32> [[X1:%.*]], [[X2:%.*]]
+; CHECK-NEXT: [[Y:%.*]] = sub nsw <16 x i32> [[X1]], [[X2]]
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X]], <16 x i32> [[Y]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %x = add nsw <16 x i32> %x1, %x2
+ %y = sub nsw <16 x i32> %x1, %x2
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+ ret <16 x i32> %s3
+}
+
+define <16 x i32> @test2_2(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test2_2(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+ ret <16 x i32> %s3
+}
+
+define <16 x i32> @test2_12(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test2_12(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+; CHECK-NEXT: [[S4:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+; CHECK-NEXT: [[C:%.*]] = mul <16 x i32> [[S3]], [[S4]]
+; CHECK-NEXT: ret <16 x i32> [[C]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 20, i32 21, i32 6, i32 23, i32 24, i32 9, i32 26, i32 27, i32 12, i32 29, i32 30, i32 15>
+ %s4 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 31, i32 28, i32 25, i32 22, i32 5, i32 4, i32 19, i32 8, i32 7, i32 18, i32 11, i32 10, i32 17, i32 14, i32 13, i32 16>
+ %c = mul <16 x i32> %s3, %s4
+ ret <16 x i32> %c
+}
+
+define <16 x i32> @test3_1(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test3_1(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 23, i32 27, i32 31, i32 6, i32 10, i32 14, i32 21, i32 25, i32 29, i32 4, i32 8, i32 12>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 18, i32 3, i32 16, i32 1, i32 5, i32 9, i32 13, i32 20, i32 24, i32 28, i32 7, i32 11, i32 15, i32 22, i32 26, i32 30>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 23, i32 27, i32 31, i32 6, i32 10, i32 14, i32 21, i32 25, i32 29, i32 4, i32 8, i32 12>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 18, i32 3, i32 16, i32 1, i32 5, i32 9, i32 13, i32 20, i32 24, i32 28, i32 7, i32 11, i32 15, i32 22, i32 26, i32 30>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ ret <16 x i32> %s3
+}
+
+define <16 x i32> @test3_2(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test3_2(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 23, i32 27, i32 31, i32 6, i32 10, i32 14, i32 21, i32 25, i32 29, i32 4, i32 8, i32 12>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 18, i32 3, i32 16, i32 1, i32 5, i32 9, i32 13, i32 20, i32 24, i32 28, i32 7, i32 11, i32 15, i32 22, i32 26, i32 30>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 23, i32 27, i32 31, i32 6, i32 10, i32 14, i32 21, i32 25, i32 29, i32 4, i32 8, i32 12>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 18, i32 3, i32 16, i32 1, i32 5, i32 9, i32 13, i32 20, i32 24, i32 28, i32 7, i32 11, i32 15, i32 22, i32 26, i32 30>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ ret <16 x i32> %s3
+}
+
+define <16 x i32> @test3_12(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test3_12(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 23, i32 27, i32 31, i32 6, i32 10, i32 14, i32 21, i32 25, i32 29, i32 4, i32 8, i32 12>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 18, i32 3, i32 16, i32 1, i32 5, i32 9, i32 13, i32 20, i32 24, i32 28, i32 7, i32 11, i32 15, i32 22, i32 26, i32 30>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[S4:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: [[C:%.*]] = add <16 x i32> [[S3]], [[S4]]
+; CHECK-NEXT: ret <16 x i32> [[C]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 23, i32 27, i32 31, i32 6, i32 10, i32 14, i32 21, i32 25, i32 29, i32 4, i32 8, i32 12>
+ %s2 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 18, i32 3, i32 16, i32 1, i32 5, i32 9, i32 13, i32 20, i32 24, i32 28, i32 7, i32 11, i32 15, i32 22, i32 26, i32 30>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %s4 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ %c = add <16 x i32> %s3, %s4
+ ret <16 x i32> %c
+}
+
+
+define <16 x i32> @test23(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test23(
+; CHECK-NEXT: [[S10:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[S20:%.*]] = shufflevector <16 x i32> [[Y]], <16 x i32> [[X]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: [[A0:%.*]] = add nsw <16 x i32> [[S10]], [[S20]]
+; CHECK-NEXT: [[B0:%.*]] = sub nsw <16 x i32> [[S10]], [[S20]]
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[A0]], <16 x i32> [[B0]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 23, i32 27, i32 31, i32 6, i32 10, i32 14, i32 21, i32 25, i32 29, i32 4, i32 8, i32 12>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[B0]], <16 x i32> [[A0]], <16 x i32> <i32 18, i32 3, i32 16, i32 1, i32 5, i32 9, i32 13, i32 20, i32 24, i32 28, i32 7, i32 11, i32 15, i32 22, i32 26, i32 30>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[S4:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+; CHECK-NEXT: [[C:%.*]] = add <16 x i32> [[S3]], [[S4]]
+; CHECK-NEXT: ret <16 x i32> [[C]]
+;
+ %s10 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %s20 = shufflevector <16 x i32> %y, <16 x i32> %x, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ %a0 = add nsw <16 x i32> %s10, %s20
+ %b0 = sub nsw <16 x i32> %s10, %s20
+ %s1 = shufflevector <16 x i32> %a0, <16 x i32> %b0, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 23, i32 27, i32 31, i32 6, i32 10, i32 14, i32 21, i32 25, i32 29, i32 4, i32 8, i32 12>
+ %s2 = shufflevector <16 x i32> %b0, <16 x i32> %a0, <16 x i32> <i32 18, i32 3, i32 16, i32 1, i32 5, i32 9, i32 13, i32 20, i32 24, i32 28, i32 7, i32 11, i32 15, i32 22, i32 26, i32 30>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %s4 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 29, i32 26, i32 7, i32 4, i32 3, i32 6, i32 5, i32 2, i32 9, i32 8, i32 17, i32 28, i32 27, i32 16, i32 31, i32 30>
+ %c = add <16 x i32> %s3, %s4
+ ret <16 x i32> %c
+}
+
+define <16 x i32> @testgood(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @testgood(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[X]], <16 x i32> [[Y]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ %s2 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ ret <16 x i32> %s3
+}
+
+declare void @use(<16 x i32>)
+define <16 x i32> @test_extrashuffleuse(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test_extrashuffleuse(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[X]], <16 x i32> [[Y]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: call void @use(<16 x i32> [[S1]])
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ %s2 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ call void @use(<16 x i32> %s1)
+ ret <16 x i32> %s3
+}
+
+define <16 x i32> @test_extraopuse(<16 x i32> %x, <16 x i32> %y) {
+; CHECK-LABEL: @test_extraopuse(
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[X:%.*]], <16 x i32> [[Y:%.*]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[X]], <16 x i32> [[Y]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: [[A:%.*]] = add nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[B:%.*]] = sub nsw <16 x i32> [[S1]], [[S2]]
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[A]], <16 x i32> [[B]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+; CHECK-NEXT: call void @use(<16 x i32> [[B]])
+; CHECK-NEXT: ret <16 x i32> [[S3]]
+;
+ %s1 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ %s2 = shufflevector <16 x i32> %x, <16 x i32> %y, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ %a = add nsw <16 x i32> %s1, %s2
+ %b = sub nsw <16 x i32> %s1, %s2
+ %s3 = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 4, i32 21, i32 6, i32 23, i32 8, i32 25, i32 10, i32 27, i32 12, i32 29, i32 14, i32 31>
+ call void @use(<16 x i32> %b)
+ ret <16 x i32> %s3
+}
+
+define <4 x i32> @test_recurse(<4 x i32> %l0, <4 x i32> %l1, <4 x i32> %l3) {
+; CHECK-LABEL: @test_recurse(
+; CHECK-NEXT: [[S0:%.*]] = shufflevector <4 x i32> [[L3:%.*]], <4 x i32> [[L1:%.*]], <4 x i32> <i32 0, i32 2, i32 1, i32 5>
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <4 x i32> [[L1]], <4 x i32> [[L3]], <4 x i32> <i32 6, i32 1, i32 3, i32 0>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <4 x i32> [[L0:%.*]], <4 x i32> [[L1]], <4 x i32> <i32 4, i32 2, i32 6, i32 2>
+; CHECK-NEXT: [[ADD:%.*]] = add <4 x i32> [[S2]], [[S2]]
+; CHECK-NEXT: [[SUB:%.*]] = sub <4 x i32> [[S1]], [[S0]]
+; CHECK-NEXT: [[T0:%.*]] = shufflevector <4 x i32> [[ADD]], <4 x i32> [[ADD]], <4 x i32> <i32 2, i32 0, i32 0, i32 3>
+; CHECK-NEXT: ret <4 x i32> [[T0]]
+;
+ %s0 = shufflevector <4 x i32> %l3, <4 x i32> %l1, <4 x i32> <i32 0, i32 2, i32 1, i32 5>
+ %s1 = shufflevector <4 x i32> %l1, <4 x i32> %l3, <4 x i32> <i32 6, i32 1, i32 3, i32 0>
+ %s2 = shufflevector <4 x i32> %l0, <4 x i32> %l1, <4 x i32> <i32 4, i32 2, i32 6, i32 2>
+ %add = add <4 x i32> %s2, %s2
+ %sub = sub <4 x i32> %s1, %s0
+ %t0 = shufflevector <4 x i32> %add, <4 x i32> %add, <4 x i32> <i32 2, i32 0, i32 0, i32 3>
+ ret <4 x i32> %t0
+}
+
+define void @test_31(ptr %src, ptr %dst) {
+; CHECK-LABEL: @test_31(
+; CHECK-NEXT: [[G0:%.*]] = getelementptr <4 x i32>, ptr [[SRC:%.*]], i32 0
+; CHECK-NEXT: [[L0:%.*]] = load <4 x i32>, ptr [[G0]], align 16
+; CHECK-NEXT: [[S0:%.*]] = shufflevector <4 x i32> [[L0]], <4 x i32> [[L0]], <4 x i32> <i32 0, i32 2, i32 1, i32 3>
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <4 x i32> [[L0]], <4 x i32> [[L0]], <4 x i32> <i32 0, i32 0, i32 2, i32 3>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <4 x i32> [[L0]], <4 x i32> [[L0]], <4 x i32> <i32 3, i32 1, i32 1, i32 0>
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <4 x i32> [[L0]], <4 x i32> [[L0]], <4 x i32> <i32 3, i32 3, i32 5, i32 6>
+; CHECK-NEXT: [[ADD:%.*]] = add <4 x i32> [[S1]], [[S1]]
+; CHECK-NEXT: [[SUB:%.*]] = sub <4 x i32> [[S3]], [[S0]]
+; CHECK-NEXT: [[T0:%.*]] = shufflevector <4 x i32> [[ADD]], <4 x i32> [[ADD]], <4 x i32> <i32 0, i32 2, i32 5, i32 2>
+; CHECK-NEXT: [[H0:%.*]] = getelementptr <4 x i32>, ptr [[DST:%.*]], i32 0
+; CHECK-NEXT: store <4 x i32> [[T0]], ptr [[H0]], align 16
+; CHECK-NEXT: [[T1:%.*]] = shufflevector <4 x i32> [[ADD]], <4 x i32> [[ADD]], <4 x i32> <i32 6, i32 0, i32 1, i32 7>
+; CHECK-NEXT: [[H1:%.*]] = getelementptr <4 x i32>, ptr [[DST]], i32 1
+; CHECK-NEXT: store <4 x i32> [[T1]], ptr [[H1]], align 16
+; CHECK-NEXT: ret void
+;
+ %g0 = getelementptr <4 x i32>, ptr %src, i32 0
+ %l0 = load <4 x i32>, ptr %g0, align 16
+ %s0 = shufflevector <4 x i32> %l0, <4 x i32> %l0, <4 x i32> <i32 0, i32 2, i32 1, i32 3>
+ %s1 = shufflevector <4 x i32> %l0, <4 x i32> %l0, <4 x i32> <i32 0, i32 0, i32 2, i32 3>
+ %s2 = shufflevector <4 x i32> %l0, <4 x i32> %l0, <4 x i32> <i32 3, i32 1, i32 1, i32 0>
+ %s3 = shufflevector <4 x i32> %l0, <4 x i32> %l0, <4 x i32> <i32 3, i32 3, i32 5, i32 6>
+ %add = add <4 x i32> %s1, %s1
+ %sub = sub <4 x i32> %s3, %s0
+ %t0 = shufflevector <4 x i32> %add, <4 x i32> %add, <4 x i32> <i32 0, i32 2, i32 5, i32 2>
+ %h0 = getelementptr <4 x i32>, ptr %dst, i32 0
+ store <4 x i32> %t0, ptr %h0, align 16
+ %t1 = shufflevector <4 x i32> %add, <4 x i32> %add, <4 x i32> <i32 6, i32 0, i32 1, i32 7>
+ %h1 = getelementptr <4 x i32>, ptr %dst, i32 1
+ store <4 x i32> %t1, ptr %h1, align 16
+ ret void
+}
+
+define <16 x i32> @test_1651256324(<16 x i32> %l0, <16 x i32> %l1, <16 x i32> %l6, <16 x i32> %l7) {
+; CHECK-LABEL: @test_1651256324(
+; CHECK-NEXT: [[S0:%.*]] = shufflevector <16 x i32> [[L0:%.*]], <16 x i32> [[L6:%.*]], <16 x i32> <i32 1, i32 20, i32 15, i32 3, i32 1, i32 10, i32 17, i32 25, i32 29, i32 23, i32 20, i32 10, i32 0, i32 20, i32 30, i32 30>
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <16 x i32> [[L1:%.*]], <16 x i32> [[L1]], <16 x i32> <i32 11, i32 22, i32 1, i32 7, i32 20, i32 0, i32 2, i32 24, i32 28, i32 10, i32 31, i32 12, i32 22, i32 5, i32 11, i32 4>
+; CHECK-NEXT: [[S2:%.*]] = shufflevector <16 x i32> [[L7:%.*]], <16 x i32> [[L7]], <16 x i32> <i32 23, i32 20, i32 29, i32 25, i32 14, i32 21, i32 11, i32 9, i32 2, i32 7, i32 5, i32 15, i32 24, i32 30, i32 26, i32 5>
+; CHECK-NEXT: [[S3:%.*]] = shufflevector <16 x i32> [[L6]], <16 x i32> [[L7]], <16 x i32> <i32 29, i32 19, i32 15, i32 30, i32 13, i32 0, i32 30, i32 23, i32 26, i32 3, i32 15, i32 24, i32 29, i32 8, i32 4, i32 0>
+; CHECK-NEXT: [[ADD:%.*]] = add <16 x i32> [[S1]], [[S1]]
+; CHECK-NEXT: [[SUB:%.*]] = sub <16 x i32> [[S0]], [[S3]]
+; CHECK-NEXT: [[T0:%.*]] = shufflevector <16 x i32> [[ADD]], <16 x i32> [[SUB]], <16 x i32> <i32 20, i32 0, i32 29, i32 2, i32 15, i32 2, i32 11, i32 7, i32 14, i32 14, i32 19, i32 30, i32 20, i32 1, i32 18, i32 28>
+; CHECK-NEXT: [[T1:%.*]] = shufflevector <16 x i32> [[SUB]], <16 x i32> [[ADD]], <16 x i32> <i32 23, i32 18, i32 25, i32 14, i32 11, i32 21, i32 29, i32 21, i32 4, i32 10, i32 21, i32 19, i32 18, i32 5, i32 12, i32 30>
+; CHECK-NEXT: [[R:%.*]] = xor <16 x i32> [[T0]], [[T1]]
+; CHECK-NEXT: ret <16 x i32> [[R]]
+;
+ %s0 = shufflevector <16 x i32> %l0, <16 x i32> %l6, <16 x i32> <i32 1, i32 20, i32 15, i32 3, i32 1, i32 10, i32 17, i32 25, i32 29, i32 23, i32 20, i32 10, i32 0, i32 20, i32 30, i32 30>
+ %s1 = shufflevector <16 x i32> %l1, <16 x i32> %l1, <16 x i32> <i32 11, i32 22, i32 1, i32 7, i32 20, i32 0, i32 2, i32 24, i32 28, i32 10, i32 31, i32 12, i32 22, i32 5, i32 11, i32 4>
+ %s2 = shufflevector <16 x i32> %l7, <16 x i32> %l7, <16 x i32> <i32 23, i32 20, i32 29, i32 25, i32 14, i32 21, i32 11, i32 9, i32 2, i32 7, i32 5, i32 15, i32 24, i32 30, i32 26, i32 5>
+ %s3 = shufflevector <16 x i32> %l6, <16 x i32> %l7, <16 x i32> <i32 29, i32 19, i32 15, i32 30, i32 13, i32 0, i32 30, i32 23, i32 26, i32 3, i32 15, i32 24, i32 29, i32 8, i32 4, i32 0>
+ %add = add <16 x i32> %s1, %s1
+ %sub = sub <16 x i32> %s0, %s3
+ %t0 = shufflevector <16 x i32> %add, <16 x i32> %sub, <16 x i32> <i32 20, i32 0, i32 29, i32 2, i32 15, i32 2, i32 11, i32 7, i32 14, i32 14, i32 19, i32 30, i32 20, i32 1, i32 18, i32 28>
+ %t1 = shufflevector <16 x i32> %sub, <16 x i32> %add, <16 x i32> <i32 23, i32 18, i32 25, i32 14, i32 11, i32 21, i32 29, i32 21, i32 4, i32 10, i32 21, i32 19, i32 18, i32 5, i32 12, i32 30>
+ %r = xor <16 x i32> %t0, %t1
+ ret <16 x i32> %r
+}
+
+define void @test_1652048214(ptr %src, ptr %dst) {
+; CHECK-LABEL: @test_1652048214(
+; CHECK-NEXT: [[G0:%.*]] = getelementptr <8 x i32>, ptr [[SRC:%.*]], i32 0
+; CHECK-NEXT: [[L0:%.*]] = load <8 x i32>, ptr [[G0]], align 32
+; CHECK-NEXT: [[G1:%.*]] = getelementptr <8 x i32>, ptr [[SRC]], i32 1
+; CHECK-NEXT: [[L1:%.*]] = load <8 x i32>, ptr [[G1]], align 32
+; CHECK-NEXT: [[S0:%.*]] = shufflevector <8 x i32> [[L1]], <8 x i32> [[L0]], <4 x i32> <i32 6, i32 2, i32 14, i32 6>
+; CHECK-NEXT: [[S1:%.*]] = shufflevector <8 x i32> [[L0]], <8 x i32> [[L0]], <4 x i32> <i32 9, i32 5, i32 1, i32 0>
+; CHECK-NEXT: [[ADD:%.*]] = add <4 x i32> [[S1]], [[S1]]
+; CHECK-NEXT: [[SUB:%.*]] = sub <4 x i32> [[S0]], [[S1]]
+; CHECK-NEXT: [[T0:%.*]] = shufflevector <4 x i32> [[SUB]], <4 x i32> [[ADD]], <4 x i32> <i32 1, i32 5, i32 7, i32 3>
+; CHECK-NEXT: [[H0:%.*]] = getelementptr <4 x i32>, ptr [[DST:%.*]], i32 0
+; CHECK-NEXT: store <4 x i32> [[T0]], ptr [[H0]], align 16
+; CHECK-NEXT: ret void
+;
+ %g0 = getelementptr <8 x i32>, ptr %src, i32 0
+ %l0 = load <8 x i32>, ptr %g0, align 32
+ %g1 = getelementptr <8 x i32>, ptr %src, i32 1
+ %l1 = load <8 x i32>, ptr %g1, align 32
+ %s0 = shufflevector <8 x i32> %l1, <8 x i32> %l0, <4 x i32> <i32 6, i32 2, i32 14, i32 6>
+ %s1 = shufflevector <8 x i32> %l0, <8 x i32> %l0, <4 x i32> <i32 9, i32 5, i32 1, i32 0>
+ %add = add <4 x i32> %s1, %s1
+ %sub = sub <4 x i32> %s0, %s1
+ %t0 = shufflevector <4 x i32> %sub, <4 x i32> %add, <4 x i32> <i32 1, i32 5, i32 7, i32 3>
+ %h0 = getelementptr <4 x i32>, ptr %dst, i32 0
+ store <4 x i32> %t0, ptr %h0, align 16
+ ret void
+}
+
+define dso_local i32 @full(i8* nocapture noundef readonly %p1, i32 noundef %st1, i8* nocapture noundef readonly %p2, i32 noundef %st2) {
+; CHECK-LABEL: @full(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[IDX_EXT:%.*]] = sext i32 [[ST1:%.*]] to i64
+; CHECK-NEXT: [[IDX_EXT63:%.*]] = sext i32 [[ST2:%.*]] to i64
+; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i8, ptr [[P1:%.*]], i64 4
+; CHECK-NEXT: [[ARRAYIDX5:%.*]] = getelementptr inbounds i8, ptr [[P2:%.*]], i64 4
+; CHECK-NEXT: [[ADD_PTR:%.*]] = getelementptr inbounds i8, ptr [[P1]], i64 [[IDX_EXT]]
+; CHECK-NEXT: [[ADD_PTR64:%.*]] = getelementptr inbounds i8, ptr [[P2]], i64 [[IDX_EXT63]]
+; CHECK-NEXT: [[ARRAYIDX3_1:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR]], i64 4
+; CHECK-NEXT: [[ARRAYIDX5_1:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64]], i64 4
+; CHECK-NEXT: [[ADD_PTR_1:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR]], i64 [[IDX_EXT]]
+; CHECK-NEXT: [[ADD_PTR64_1:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64]], i64 [[IDX_EXT63]]
+; CHECK-NEXT: [[ARRAYIDX3_2:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR_1]], i64 4
+; CHECK-NEXT: [[ARRAYIDX5_2:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64_1]], i64 4
+; CHECK-NEXT: [[ADD_PTR_2:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR_1]], i64 [[IDX_EXT]]
+; CHECK-NEXT: [[ADD_PTR64_2:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64_1]], i64 [[IDX_EXT63]]
+; CHECK-NEXT: [[ARRAYIDX3_3:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR_2]], i64 4
+; CHECK-NEXT: [[ARRAYIDX5_3:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64_2]], i64 4
+; CHECK-NEXT: [[TMP0:%.*]] = bitcast ptr [[P1]] to ptr
+; CHECK-NEXT: [[TMP1:%.*]] = load <4 x i8>, ptr [[TMP0]], align 1
+; CHECK-NEXT: [[TMP2:%.*]] = bitcast ptr [[P2]] to ptr
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i8>, ptr [[TMP2]], align 1
+; CHECK-NEXT: [[TMP4:%.*]] = bitcast ptr [[ARRAYIDX3]] to ptr
+; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i8>, ptr [[TMP4]], align 1
+; CHECK-NEXT: [[TMP6:%.*]] = bitcast ptr [[ARRAYIDX5]] to ptr
+; CHECK-NEXT: [[TMP7:%.*]] = load <4 x i8>, ptr [[TMP6]], align 1
+; CHECK-NEXT: [[TMP8:%.*]] = bitcast ptr [[ADD_PTR]] to ptr
+; CHECK-NEXT: [[TMP9:%.*]] = load <4 x i8>, ptr [[TMP8]], align 1
+; CHECK-NEXT: [[TMP10:%.*]] = bitcast ptr [[ADD_PTR64]] to ptr
+; CHECK-NEXT: [[TMP11:%.*]] = load <4 x i8>, ptr [[TMP10]], align 1
+; CHECK-NEXT: [[TMP12:%.*]] = bitcast ptr [[ARRAYIDX3_1]] to ptr
+; CHECK-NEXT: [[TMP13:%.*]] = load <4 x i8>, ptr [[TMP12]], align 1
+; CHECK-NEXT: [[TMP14:%.*]] = bitcast ptr [[ARRAYIDX5_1]] to ptr
+; CHECK-NEXT: [[TMP15:%.*]] = load <4 x i8>, ptr [[TMP14]], align 1
+; CHECK-NEXT: [[TMP16:%.*]] = bitcast ptr [[ADD_PTR_1]] to ptr
+; CHECK-NEXT: [[TMP17:%.*]] = load <4 x i8>, ptr [[TMP16]], align 1
+; CHECK-NEXT: [[TMP18:%.*]] = bitcast ptr [[ADD_PTR64_1]] to ptr
+; CHECK-NEXT: [[TMP19:%.*]] = load <4 x i8>, ptr [[TMP18]], align 1
+; CHECK-NEXT: [[TMP20:%.*]] = bitcast ptr [[ARRAYIDX3_2]] to ptr
+; CHECK-NEXT: [[TMP21:%.*]] = load <4 x i8>, ptr [[TMP20]], align 1
+; CHECK-NEXT: [[TMP22:%.*]] = bitcast ptr [[ARRAYIDX5_2]] to ptr
+; CHECK-NEXT: [[TMP23:%.*]] = load <4 x i8>, ptr [[TMP22]], align 1
+; CHECK-NEXT: [[TMP24:%.*]] = bitcast ptr [[ADD_PTR_2]] to ptr
+; CHECK-NEXT: [[TMP25:%.*]] = load <4 x i8>, ptr [[TMP24]], align 1
+; CHECK-NEXT: [[TMP26:%.*]] = shufflevector <4 x i8> [[TMP25]], <4 x i8> [[TMP17]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP27:%.*]] = shufflevector <4 x i8> [[TMP9]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP28:%.*]] = shufflevector <16 x i8> [[TMP26]], <16 x i8> [[TMP27]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP29:%.*]] = shufflevector <4 x i8> [[TMP1]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP30:%.*]] = shufflevector <16 x i8> [[TMP28]], <16 x i8> [[TMP29]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: [[TMP31:%.*]] = zext <16 x i8> [[TMP30]] to <16 x i32>
+; CHECK-NEXT: [[TMP32:%.*]] = bitcast ptr [[ADD_PTR64_2]] to ptr
+; CHECK-NEXT: [[TMP33:%.*]] = load <4 x i8>, ptr [[TMP32]], align 1
+; CHECK-NEXT: [[TMP34:%.*]] = shufflevector <4 x i8> [[TMP33]], <4 x i8> [[TMP19]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP35:%.*]] = shufflevector <4 x i8> [[TMP11]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP36:%.*]] = shufflevector <16 x i8> [[TMP34]], <16 x i8> [[TMP35]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP37:%.*]] = shufflevector <4 x i8> [[TMP3]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP38:%.*]] = shufflevector <16 x i8> [[TMP36]], <16 x i8> [[TMP37]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: [[TMP39:%.*]] = zext <16 x i8> [[TMP38]] to <16 x i32>
+; CHECK-NEXT: [[TMP40:%.*]] = sub nsw <16 x i32> [[TMP31]], [[TMP39]]
+; CHECK-NEXT: [[TMP41:%.*]] = bitcast ptr [[ARRAYIDX3_3]] to ptr
+; CHECK-NEXT: [[TMP42:%.*]] = load <4 x i8>, ptr [[TMP41]], align 1
+; CHECK-NEXT: [[TMP43:%.*]] = shufflevector <4 x i8> [[TMP42]], <4 x i8> [[TMP21]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP44:%.*]] = shufflevector <4 x i8> [[TMP13]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP45:%.*]] = shufflevector <16 x i8> [[TMP43]], <16 x i8> [[TMP44]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP46:%.*]] = shufflevector <4 x i8> [[TMP5]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP47:%.*]] = shufflevector <16 x i8> [[TMP45]], <16 x i8> [[TMP46]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: [[TMP48:%.*]] = zext <16 x i8> [[TMP47]] to <16 x i32>
+; CHECK-NEXT: [[TMP49:%.*]] = bitcast ptr [[ARRAYIDX5_3]] to ptr
+; CHECK-NEXT: [[TMP50:%.*]] = load <4 x i8>, ptr [[TMP49]], align 1
+; CHECK-NEXT: [[TMP51:%.*]] = shufflevector <4 x i8> [[TMP50]], <4 x i8> [[TMP23]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP52:%.*]] = shufflevector <4 x i8> [[TMP15]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP53:%.*]] = shufflevector <16 x i8> [[TMP51]], <16 x i8> [[TMP52]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP54:%.*]] = shufflevector <4 x i8> [[TMP7]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP55:%.*]] = shufflevector <16 x i8> [[TMP53]], <16 x i8> [[TMP54]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: [[TMP56:%.*]] = zext <16 x i8> [[TMP55]] to <16 x i32>
+; CHECK-NEXT: [[TMP57:%.*]] = sub nsw <16 x i32> [[TMP48]], [[TMP56]]
+; CHECK-NEXT: [[TMP58:%.*]] = shl nsw <16 x i32> [[TMP57]], <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; CHECK-NEXT: [[TMP59:%.*]] = add nsw <16 x i32> [[TMP58]], [[TMP40]]
+; CHECK-NEXT: [[REORDER:%.*]] = shufflevector <16 x i32> [[TMP59]], <16 x i32> poison, <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 13, i32 12, i32 15, i32 14>
+; CHECK-NEXT: [[TMP60:%.*]] = add nsw <16 x i32> [[TMP59]], [[REORDER]]
+; CHECK-NEXT: [[TMP61:%.*]] = sub nsw <16 x i32> [[TMP59]], [[REORDER]]
+; CHECK-NEXT: [[TMP62:%.*]] = shufflevector <16 x i32> [[TMP60]], <16 x i32> [[TMP61]], <16 x i32> <i32 3, i32 7, i32 11, i32 15, i32 22, i32 18, i32 26, i32 30, i32 5, i32 1, i32 9, i32 13, i32 20, i32 16, i32 24, i32 28>
+; CHECK-NEXT: [[REORDER1:%.*]] = shufflevector <16 x i32> [[TMP60]], <16 x i32> [[TMP61]], <16 x i32> <i32 1, i32 5, i32 9, i32 13, i32 20, i32 16, i32 24, i32 28, i32 7, i32 3, i32 11, i32 15, i32 22, i32 18, i32 26, i32 30>
+; CHECK-NEXT: [[TMP63:%.*]] = add nsw <16 x i32> [[TMP62]], [[REORDER1]]
+; CHECK-NEXT: [[TMP64:%.*]] = sub nsw <16 x i32> [[TMP62]], [[REORDER1]]
+; CHECK-NEXT: [[TMP65:%.*]] = shufflevector <16 x i32> [[TMP63]], <16 x i32> [[TMP64]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; CHECK-NEXT: [[REORDER2:%.*]] = shufflevector <16 x i32> [[TMP63]], <16 x i32> [[TMP64]], <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 25, i32 24, i32 27, i32 26, i32 29, i32 28, i32 31, i32 30>
+; CHECK-NEXT: [[TMP66:%.*]] = add nsw <16 x i32> [[TMP65]], [[REORDER2]]
+; CHECK-NEXT: [[TMP67:%.*]] = sub nsw <16 x i32> [[TMP65]], [[REORDER2]]
+; CHECK-NEXT: [[TMP68:%.*]] = shufflevector <16 x i32> [[TMP66]], <16 x i32> [[TMP67]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 20, i32 5, i32 6, i32 23, i32 24, i32 9, i32 10, i32 27, i32 28, i32 13, i32 14, i32 31>
+; CHECK-NEXT: [[REORDER3:%.*]] = shufflevector <16 x i32> [[TMP66]], <16 x i32> [[TMP67]], <16 x i32> <i32 2, i32 19, i32 0, i32 17, i32 23, i32 6, i32 5, i32 20, i32 27, i32 10, i32 9, i32 24, i32 31, i32 14, i32 13, i32 28>
+; CHECK-NEXT: [[TMP69:%.*]] = add nsw <16 x i32> [[TMP68]], [[REORDER3]]
+; CHECK-NEXT: [[TMP70:%.*]] = sub nsw <16 x i32> [[TMP68]], [[REORDER3]]
+; CHECK-NEXT: [[TMP71:%.*]] = shufflevector <16 x i32> [[TMP69]], <16 x i32> [[TMP70]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 4, i32 5, i32 22, i32 23, i32 8, i32 9, i32 26, i32 27, i32 12, i32 13, i32 30, i32 31>
+; CHECK-NEXT: [[TMP72:%.*]] = lshr <16 x i32> [[TMP71]], <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
+; CHECK-NEXT: [[TMP73:%.*]] = and <16 x i32> [[TMP72]], <i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537>
+; CHECK-NEXT: [[TMP74:%.*]] = mul nuw <16 x i32> [[TMP73]], <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535>
+; CHECK-NEXT: [[TMP75:%.*]] = add <16 x i32> [[TMP74]], [[TMP71]]
+; CHECK-NEXT: [[TMP76:%.*]] = xor <16 x i32> [[TMP75]], [[TMP74]]
+; CHECK-NEXT: [[TMP77:%.*]] = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> [[TMP76]])
+; CHECK-NEXT: [[CONV118:%.*]] = and i32 [[TMP77]], 65535
+; CHECK-NEXT: [[SHR:%.*]] = lshr i32 [[TMP77]], 16
+; CHECK-NEXT: [[ADD119:%.*]] = add nuw nsw i32 [[CONV118]], [[SHR]]
+; CHECK-NEXT: [[SHR120:%.*]] = lshr i32 [[ADD119]], 1
+; CHECK-NEXT: ret i32 [[SHR120]]
+;
+entry:
+ %idx.ext = sext i32 %st1 to i64
+ %idx.ext63 = sext i32 %st2 to i64
+ %arrayidx3 = getelementptr inbounds i8, i8* %p1, i64 4
+ %arrayidx5 = getelementptr inbounds i8, i8* %p2, i64 4
+ %add.ptr = getelementptr inbounds i8, i8* %p1, i64 %idx.ext
+ %add.ptr64 = getelementptr inbounds i8, i8* %p2, i64 %idx.ext63
+ %arrayidx3.1 = getelementptr inbounds i8, i8* %add.ptr, i64 4
+ %arrayidx5.1 = getelementptr inbounds i8, i8* %add.ptr64, i64 4
+ %add.ptr.1 = getelementptr inbounds i8, i8* %add.ptr, i64 %idx.ext
+ %add.ptr64.1 = getelementptr inbounds i8, i8* %add.ptr64, i64 %idx.ext63
+ %arrayidx3.2 = getelementptr inbounds i8, i8* %add.ptr.1, i64 4
+ %arrayidx5.2 = getelementptr inbounds i8, i8* %add.ptr64.1, i64 4
+ %add.ptr.2 = getelementptr inbounds i8, i8* %add.ptr.1, i64 %idx.ext
+ %add.ptr64.2 = getelementptr inbounds i8, i8* %add.ptr64.1, i64 %idx.ext63
+ %arrayidx3.3 = getelementptr inbounds i8, i8* %add.ptr.2, i64 4
+ %arrayidx5.3 = getelementptr inbounds i8, i8* %add.ptr64.2, i64 4
+ %0 = bitcast i8* %p1 to <4 x i8>*
+ %1 = load <4 x i8>, <4 x i8>* %0, align 1
+ %2 = bitcast i8* %p2 to <4 x i8>*
+ %3 = load <4 x i8>, <4 x i8>* %2, align 1
+ %4 = bitcast i8* %arrayidx3 to <4 x i8>*
+ %5 = load <4 x i8>, <4 x i8>* %4, align 1
+ %6 = bitcast i8* %arrayidx5 to <4 x i8>*
+ %7 = load <4 x i8>, <4 x i8>* %6, align 1
+ %8 = bitcast i8* %add.ptr to <4 x i8>*
+ %9 = load <4 x i8>, <4 x i8>* %8, align 1
+ %10 = bitcast i8* %add.ptr64 to <4 x i8>*
+ %11 = load <4 x i8>, <4 x i8>* %10, align 1
+ %12 = bitcast i8* %arrayidx3.1 to <4 x i8>*
+ %13 = load <4 x i8>, <4 x i8>* %12, align 1
+ %14 = bitcast i8* %arrayidx5.1 to <4 x i8>*
+ %15 = load <4 x i8>, <4 x i8>* %14, align 1
+ %16 = bitcast i8* %add.ptr.1 to <4 x i8>*
+ %17 = load <4 x i8>, <4 x i8>* %16, align 1
+ %18 = bitcast i8* %add.ptr64.1 to <4 x i8>*
+ %19 = load <4 x i8>, <4 x i8>* %18, align 1
+ %20 = bitcast i8* %arrayidx3.2 to <4 x i8>*
+ %21 = load <4 x i8>, <4 x i8>* %20, align 1
+ %22 = bitcast i8* %arrayidx5.2 to <4 x i8>*
+ %23 = load <4 x i8>, <4 x i8>* %22, align 1
+ %24 = bitcast i8* %add.ptr.2 to <4 x i8>*
+ %25 = load <4 x i8>, <4 x i8>* %24, align 1
+ %26 = shufflevector <4 x i8> %25, <4 x i8> %17, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %27 = shufflevector <4 x i8> %9, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %28 = shufflevector <16 x i8> %26, <16 x i8> %27, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 undef, i32 undef, i32 undef, i32 undef>
+ %29 = shufflevector <4 x i8> %1, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %30 = shufflevector <16 x i8> %28, <16 x i8> %29, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+ %31 = zext <16 x i8> %30 to <16 x i32>
+ %32 = bitcast i8* %add.ptr64.2 to <4 x i8>*
+ %33 = load <4 x i8>, <4 x i8>* %32, align 1
+ %34 = shufflevector <4 x i8> %33, <4 x i8> %19, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %35 = shufflevector <4 x i8> %11, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %36 = shufflevector <16 x i8> %34, <16 x i8> %35, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 undef, i32 undef, i32 undef, i32 undef>
+ %37 = shufflevector <4 x i8> %3, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %38 = shufflevector <16 x i8> %36, <16 x i8> %37, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+ %39 = zext <16 x i8> %38 to <16 x i32>
+ %40 = sub nsw <16 x i32> %31, %39
+ %41 = bitcast i8* %arrayidx3.3 to <4 x i8>*
+ %42 = load <4 x i8>, <4 x i8>* %41, align 1
+ %43 = shufflevector <4 x i8> %42, <4 x i8> %21, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %44 = shufflevector <4 x i8> %13, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %45 = shufflevector <16 x i8> %43, <16 x i8> %44, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 undef, i32 undef, i32 undef, i32 undef>
+ %46 = shufflevector <4 x i8> %5, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %47 = shufflevector <16 x i8> %45, <16 x i8> %46, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+ %48 = zext <16 x i8> %47 to <16 x i32>
+ %49 = bitcast i8* %arrayidx5.3 to <4 x i8>*
+ %50 = load <4 x i8>, <4 x i8>* %49, align 1
+ %51 = shufflevector <4 x i8> %50, <4 x i8> %23, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %52 = shufflevector <4 x i8> %15, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %53 = shufflevector <16 x i8> %51, <16 x i8> %52, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 undef, i32 undef, i32 undef, i32 undef>
+ %54 = shufflevector <4 x i8> %7, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %55 = shufflevector <16 x i8> %53, <16 x i8> %54, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+ %56 = zext <16 x i8> %55 to <16 x i32>
+ %57 = sub nsw <16 x i32> %48, %56
+ %58 = shl nsw <16 x i32> %57, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+ %59 = add nsw <16 x i32> %58, %40
+ %reorder = shufflevector <16 x i32> %59, <16 x i32> poison, <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 13, i32 12, i32 15, i32 14>
+ %60 = add nsw <16 x i32> %59, %reorder
+ %61 = sub nsw <16 x i32> %59, %reorder
+ %62 = shufflevector <16 x i32> %60, <16 x i32> %61, <16 x i32> <i32 3, i32 7, i32 11, i32 15, i32 22, i32 18, i32 26, i32 30, i32 5, i32 1, i32 9, i32 13, i32 20, i32 16, i32 24, i32 28>
+ %reorder1 = shufflevector <16 x i32> %60, <16 x i32> %61, <16 x i32> <i32 1, i32 5, i32 9, i32 13, i32 20, i32 16, i32 24, i32 28, i32 7, i32 3, i32 11, i32 15, i32 22, i32 18, i32 26, i32 30>
+ %63 = add nsw <16 x i32> %62, %reorder1
+ %64 = sub nsw <16 x i32> %62, %reorder1
+ %65 = shufflevector <16 x i32> %63, <16 x i32> %64, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+ %reorder2 = shufflevector <16 x i32> %63, <16 x i32> %64, <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 25, i32 24, i32 27, i32 26, i32 29, i32 28, i32 31, i32 30>
+ %66 = add nsw <16 x i32> %65, %reorder2
+ %67 = sub nsw <16 x i32> %65, %reorder2
+ %68 = shufflevector <16 x i32> %66, <16 x i32> %67, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 20, i32 5, i32 6, i32 23, i32 24, i32 9, i32 10, i32 27, i32 28, i32 13, i32 14, i32 31>
+ %reorder3 = shufflevector <16 x i32> %66, <16 x i32> %67, <16 x i32> <i32 2, i32 19, i32 0, i32 17, i32 23, i32 6, i32 5, i32 20, i32 27, i32 10, i32 9, i32 24, i32 31, i32 14, i32 13, i32 28>
+ %69 = add nsw <16 x i32> %68, %reorder3
+ %70 = sub nsw <16 x i32> %68, %reorder3
+ %71 = shufflevector <16 x i32> %69, <16 x i32> %70, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 4, i32 5, i32 22, i32 23, i32 8, i32 9, i32 26, i32 27, i32 12, i32 13, i32 30, i32 31>
+ %72 = lshr <16 x i32> %71, <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
+ %73 = and <16 x i32> %72, <i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537>
+ %74 = mul nuw <16 x i32> %73, <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535>
+ %75 = add <16 x i32> %74, %71
+ %76 = xor <16 x i32> %75, %74
+ %77 = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %76)
+ %conv118 = and i32 %77, 65535
+ %shr = lshr i32 %77, 16
+ %add119 = add nuw nsw i32 %conv118, %shr
+ %shr120 = lshr i32 %add119, 1
+ ret i32 %shr120
+}
+
+define i32 @full_reorder(ptr nocapture noundef readonly %pix1, i32 noundef %i_pix1, ptr nocapture noundef readonly %pix2, i32 noundef %i_pix2) {
+; CHECK-LABEL: @full_reorder(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[IDX_EXT:%.*]] = sext i32 [[I_PIX1:%.*]] to i64
+; CHECK-NEXT: [[IDX_EXT63:%.*]] = sext i32 [[I_PIX2:%.*]] to i64
+; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i8, ptr [[PIX1:%.*]], i64 4
+; CHECK-NEXT: [[ARRAYIDX5:%.*]] = getelementptr inbounds i8, ptr [[PIX2:%.*]], i64 4
+; CHECK-NEXT: [[ADD_PTR:%.*]] = getelementptr inbounds i8, ptr [[PIX1]], i64 [[IDX_EXT]]
+; CHECK-NEXT: [[ADD_PTR64:%.*]] = getelementptr inbounds i8, ptr [[PIX2]], i64 [[IDX_EXT63]]
+; CHECK-NEXT: [[ARRAYIDX3_1:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR]], i64 4
+; CHECK-NEXT: [[ARRAYIDX5_1:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64]], i64 4
+; CHECK-NEXT: [[ADD_PTR_1:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR]], i64 [[IDX_EXT]]
+; CHECK-NEXT: [[ADD_PTR64_1:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64]], i64 [[IDX_EXT63]]
+; CHECK-NEXT: [[ARRAYIDX3_2:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR_1]], i64 4
+; CHECK-NEXT: [[ARRAYIDX5_2:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64_1]], i64 4
+; CHECK-NEXT: [[ADD_PTR_2:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR_1]], i64 [[IDX_EXT]]
+; CHECK-NEXT: [[ADD_PTR64_2:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64_1]], i64 [[IDX_EXT63]]
+; CHECK-NEXT: [[ARRAYIDX3_3:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR_2]], i64 4
+; CHECK-NEXT: [[ARRAYIDX5_3:%.*]] = getelementptr inbounds i8, ptr [[ADD_PTR64_2]], i64 4
+; CHECK-NEXT: [[TMP0:%.*]] = load <4 x i8>, ptr [[PIX1]], align 1
+; CHECK-NEXT: [[TMP1:%.*]] = load <4 x i8>, ptr [[PIX2]], align 1
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i8>, ptr [[ARRAYIDX3]], align 1
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i8>, ptr [[ARRAYIDX5]], align 1
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i8>, ptr [[ADD_PTR]], align 1
+; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i8>, ptr [[ADD_PTR64]], align 1
+; CHECK-NEXT: [[TMP6:%.*]] = load <4 x i8>, ptr [[ARRAYIDX3_1]], align 1
+; CHECK-NEXT: [[TMP7:%.*]] = load <4 x i8>, ptr [[ARRAYIDX5_1]], align 1
+; CHECK-NEXT: [[TMP8:%.*]] = load <4 x i8>, ptr [[ADD_PTR_1]], align 1
+; CHECK-NEXT: [[TMP9:%.*]] = load <4 x i8>, ptr [[ADD_PTR64_1]], align 1
+; CHECK-NEXT: [[TMP10:%.*]] = load <4 x i8>, ptr [[ARRAYIDX3_2]], align 1
+; CHECK-NEXT: [[TMP11:%.*]] = load <4 x i8>, ptr [[ARRAYIDX5_2]], align 1
+; CHECK-NEXT: [[TMP12:%.*]] = load <4 x i8>, ptr [[ADD_PTR_2]], align 1
+; CHECK-NEXT: [[TMP13:%.*]] = shufflevector <4 x i8> [[TMP12]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP14:%.*]] = shufflevector <4 x i8> [[TMP8]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP15:%.*]] = shufflevector <16 x i8> [[TMP13]], <16 x i8> [[TMP14]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[TMP16:%.*]] = shufflevector <4 x i8> [[TMP4]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP17:%.*]] = shufflevector <16 x i8> [[TMP15]], <16 x i8> [[TMP16]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[TMP18:%.*]] = shufflevector <4 x i8> [[TMP0]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP19:%.*]] = shufflevector <16 x i8> [[TMP17]], <16 x i8> [[TMP18]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: [[TMP20:%.*]] = zext <16 x i8> [[TMP19]] to <16 x i32>
+; CHECK-NEXT: [[TMP21:%.*]] = load <4 x i8>, ptr [[ADD_PTR64_2]], align 1
+; CHECK-NEXT: [[TMP22:%.*]] = shufflevector <4 x i8> [[TMP21]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP23:%.*]] = shufflevector <4 x i8> [[TMP9]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP24:%.*]] = shufflevector <16 x i8> [[TMP22]], <16 x i8> [[TMP23]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[TMP25:%.*]] = shufflevector <4 x i8> [[TMP5]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP26:%.*]] = shufflevector <16 x i8> [[TMP24]], <16 x i8> [[TMP25]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[TMP27:%.*]] = shufflevector <4 x i8> [[TMP1]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP28:%.*]] = shufflevector <16 x i8> [[TMP26]], <16 x i8> [[TMP27]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: [[TMP29:%.*]] = zext <16 x i8> [[TMP28]] to <16 x i32>
+; CHECK-NEXT: [[TMP30:%.*]] = sub nsw <16 x i32> [[TMP20]], [[TMP29]]
+; CHECK-NEXT: [[TMP31:%.*]] = load <4 x i8>, ptr [[ARRAYIDX3_3]], align 1
+; CHECK-NEXT: [[TMP32:%.*]] = shufflevector <4 x i8> [[TMP31]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP33:%.*]] = shufflevector <4 x i8> [[TMP10]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP34:%.*]] = shufflevector <16 x i8> [[TMP32]], <16 x i8> [[TMP33]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[TMP35:%.*]] = shufflevector <4 x i8> [[TMP6]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP36:%.*]] = shufflevector <16 x i8> [[TMP34]], <16 x i8> [[TMP35]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[TMP37:%.*]] = shufflevector <4 x i8> [[TMP2]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP38:%.*]] = shufflevector <16 x i8> [[TMP36]], <16 x i8> [[TMP37]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: [[TMP39:%.*]] = zext <16 x i8> [[TMP38]] to <16 x i32>
+; CHECK-NEXT: [[TMP40:%.*]] = load <4 x i8>, ptr [[ARRAYIDX5_3]], align 1
+; CHECK-NEXT: [[TMP41:%.*]] = shufflevector <4 x i8> [[TMP40]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP42:%.*]] = shufflevector <4 x i8> [[TMP11]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP43:%.*]] = shufflevector <16 x i8> [[TMP41]], <16 x i8> [[TMP42]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[TMP44:%.*]] = shufflevector <4 x i8> [[TMP7]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP45:%.*]] = shufflevector <16 x i8> [[TMP43]], <16 x i8> [[TMP44]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: [[TMP46:%.*]] = shufflevector <4 x i8> [[TMP3]], <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+; CHECK-NEXT: [[TMP47:%.*]] = shufflevector <16 x i8> [[TMP45]], <16 x i8> [[TMP46]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: [[TMP48:%.*]] = zext <16 x i8> [[TMP47]] to <16 x i32>
+; CHECK-NEXT: [[TMP49:%.*]] = sub nsw <16 x i32> [[TMP39]], [[TMP48]]
+; CHECK-NEXT: [[TMP50:%.*]] = shl nsw <16 x i32> [[TMP49]], <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+; CHECK-NEXT: [[TMP51:%.*]] = add nsw <16 x i32> [[TMP50]], [[TMP30]]
+; CHECK-NEXT: [[REORDER:%.*]] = shufflevector <16 x i32> [[TMP51]], <16 x i32> poison, <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 13, i32 12, i32 15, i32 14>
+; CHECK-NEXT: [[TMP52:%.*]] = add nsw <16 x i32> [[TMP51]], [[REORDER]]
+; CHECK-NEXT: [[TMP53:%.*]] = sub nsw <16 x i32> [[TMP51]], [[REORDER]]
+; CHECK-NEXT: [[TMP54:%.*]] = shufflevector <16 x i32> [[TMP52]], <16 x i32> [[TMP53]], <16 x i32> <i32 3, i32 7, i32 11, i32 15, i32 22, i32 18, i32 26, i32 30, i32 5, i32 1, i32 9, i32 13, i32 20, i32 16, i32 24, i32 28>
+; CHECK-NEXT: [[REORDER191:%.*]] = shufflevector <16 x i32> [[TMP52]], <16 x i32> [[TMP53]], <16 x i32> <i32 1, i32 5, i32 9, i32 13, i32 20, i32 16, i32 24, i32 28, i32 7, i32 3, i32 11, i32 15, i32 22, i32 18, i32 26, i32 30>
+; CHECK-NEXT: [[TMP55:%.*]] = add nsw <16 x i32> [[TMP54]], [[REORDER191]]
+; CHECK-NEXT: [[TMP56:%.*]] = sub nsw <16 x i32> [[TMP54]], [[REORDER191]]
+; CHECK-NEXT: [[TMP57:%.*]] = shufflevector <16 x i32> [[TMP55]], <16 x i32> [[TMP56]], <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; CHECK-NEXT: [[REORDER192:%.*]] = shufflevector <16 x i32> [[TMP55]], <16 x i32> [[TMP56]], <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 25, i32 24, i32 27, i32 26, i32 29, i32 28, i32 31, i32 30>
+; CHECK-NEXT: [[TMP58:%.*]] = add nsw <16 x i32> [[TMP57]], [[REORDER192]]
+; CHECK-NEXT: [[TMP59:%.*]] = sub nsw <16 x i32> [[TMP57]], [[REORDER192]]
+; CHECK-NEXT: [[TMP60:%.*]] = shufflevector <16 x i32> [[TMP58]], <16 x i32> [[TMP59]], <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 20, i32 5, i32 6, i32 23, i32 24, i32 9, i32 10, i32 27, i32 28, i32 13, i32 14, i32 31>
+; CHECK-NEXT: [[REORDER193:%.*]] = shufflevector <16 x i32> [[TMP58]], <16 x i32> [[TMP59]], <16 x i32> <i32 2, i32 19, i32 0, i32 17, i32 23, i32 6, i32 5, i32 20, i32 27, i32 10, i32 9, i32 24, i32 31, i32 14, i32 13, i32 28>
+; CHECK-NEXT: [[TMP61:%.*]] = add nsw <16 x i32> [[TMP60]], [[REORDER193]]
+; CHECK-NEXT: [[TMP62:%.*]] = sub nsw <16 x i32> [[TMP60]], [[REORDER193]]
+; CHECK-NEXT: [[TMP63:%.*]] = shufflevector <16 x i32> [[TMP61]], <16 x i32> [[TMP62]], <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 4, i32 5, i32 22, i32 23, i32 8, i32 9, i32 26, i32 27, i32 12, i32 13, i32 30, i32 31>
+; CHECK-NEXT: [[TMP64:%.*]] = lshr <16 x i32> [[TMP63]], <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
+; CHECK-NEXT: [[TMP65:%.*]] = and <16 x i32> [[TMP64]], <i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537>
+; CHECK-NEXT: [[TMP66:%.*]] = mul nuw <16 x i32> [[TMP65]], <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535>
+; CHECK-NEXT: [[TMP67:%.*]] = add <16 x i32> [[TMP66]], [[TMP63]]
+; CHECK-NEXT: [[TMP68:%.*]] = xor <16 x i32> [[TMP67]], [[TMP66]]
+; CHECK-NEXT: [[TMP69:%.*]] = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> [[TMP68]])
+; CHECK-NEXT: [[CONV118:%.*]] = and i32 [[TMP69]], 65535
+; CHECK-NEXT: [[SHR:%.*]] = lshr i32 [[TMP69]], 16
+; CHECK-NEXT: [[ADD119:%.*]] = add nuw nsw i32 [[CONV118]], [[SHR]]
+; CHECK-NEXT: [[SHR120:%.*]] = lshr i32 [[ADD119]], 1
+; CHECK-NEXT: ret i32 [[SHR120]]
+;
+entry:
+ %idx.ext = sext i32 %i_pix1 to i64
+ %idx.ext63 = sext i32 %i_pix2 to i64
+ %arrayidx3 = getelementptr inbounds i8, ptr %pix1, i64 4
+ %arrayidx5 = getelementptr inbounds i8, ptr %pix2, i64 4
+ %add.ptr = getelementptr inbounds i8, ptr %pix1, i64 %idx.ext
+ %add.ptr64 = getelementptr inbounds i8, ptr %pix2, i64 %idx.ext63
+ %arrayidx3.1 = getelementptr inbounds i8, ptr %add.ptr, i64 4
+ %arrayidx5.1 = getelementptr inbounds i8, ptr %add.ptr64, i64 4
+ %add.ptr.1 = getelementptr inbounds i8, ptr %add.ptr, i64 %idx.ext
+ %add.ptr64.1 = getelementptr inbounds i8, ptr %add.ptr64, i64 %idx.ext63
+ %arrayidx3.2 = getelementptr inbounds i8, ptr %add.ptr.1, i64 4
+ %arrayidx5.2 = getelementptr inbounds i8, ptr %add.ptr64.1, i64 4
+ %add.ptr.2 = getelementptr inbounds i8, ptr %add.ptr.1, i64 %idx.ext
+ %add.ptr64.2 = getelementptr inbounds i8, ptr %add.ptr64.1, i64 %idx.ext63
+ %arrayidx3.3 = getelementptr inbounds i8, ptr %add.ptr.2, i64 4
+ %arrayidx5.3 = getelementptr inbounds i8, ptr %add.ptr64.2, i64 4
+ %0 = load <4 x i8>, ptr %pix1, align 1
+ %1 = load <4 x i8>, ptr %pix2, align 1
+ %2 = load <4 x i8>, ptr %arrayidx3, align 1
+ %3 = load <4 x i8>, ptr %arrayidx5, align 1
+ %4 = load <4 x i8>, ptr %add.ptr, align 1
+ %5 = load <4 x i8>, ptr %add.ptr64, align 1
+ %6 = load <4 x i8>, ptr %arrayidx3.1, align 1
+ %7 = load <4 x i8>, ptr %arrayidx5.1, align 1
+ %8 = load <4 x i8>, ptr %add.ptr.1, align 1
+ %9 = load <4 x i8>, ptr %add.ptr64.1, align 1
+ %10 = load <4 x i8>, ptr %arrayidx3.2, align 1
+ %11 = load <4 x i8>, ptr %arrayidx5.2, align 1
+ %12 = load <4 x i8>, ptr %add.ptr.2, align 1
+ %13 = shufflevector <4 x i8> %12, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %14 = shufflevector <4 x i8> %8, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %15 = shufflevector <16 x i8> %13, <16 x i8> %14, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %16 = shufflevector <4 x i8> %4, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %17 = shufflevector <16 x i8> %15, <16 x i8> %16, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+ %18 = shufflevector <4 x i8> %0, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %19 = shufflevector <16 x i8> %17, <16 x i8> %18, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+ %20 = zext <16 x i8> %19 to <16 x i32>
+ %21 = load <4 x i8>, ptr %add.ptr64.2, align 1
+ %22 = shufflevector <4 x i8> %21, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %23 = shufflevector <4 x i8> %9, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %24 = shufflevector <16 x i8> %22, <16 x i8> %23, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %25 = shufflevector <4 x i8> %5, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %26 = shufflevector <16 x i8> %24, <16 x i8> %25, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+ %27 = shufflevector <4 x i8> %1, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %28 = shufflevector <16 x i8> %26, <16 x i8> %27, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+ %29 = zext <16 x i8> %28 to <16 x i32>
+ %30 = sub nsw <16 x i32> %20, %29
+ %31 = load <4 x i8>, ptr %arrayidx3.3, align 1
+ %32 = shufflevector <4 x i8> %31, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %33 = shufflevector <4 x i8> %10, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %34 = shufflevector <16 x i8> %32, <16 x i8> %33, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %35 = shufflevector <4 x i8> %6, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %36 = shufflevector <16 x i8> %34, <16 x i8> %35, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+ %37 = shufflevector <4 x i8> %2, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %38 = shufflevector <16 x i8> %36, <16 x i8> %37, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+ %39 = zext <16 x i8> %38 to <16 x i32>
+ %40 = load <4 x i8>, ptr %arrayidx5.3, align 1
+ %41 = shufflevector <4 x i8> %40, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %42 = shufflevector <4 x i8> %11, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %43 = shufflevector <16 x i8> %41, <16 x i8> %42, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %44 = shufflevector <4 x i8> %7, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %45 = shufflevector <16 x i8> %43, <16 x i8> %44, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 12, i32 13, i32 14, i32 15>
+ %46 = shufflevector <4 x i8> %3, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
+ %47 = shufflevector <16 x i8> %45, <16 x i8> %46, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 16, i32 17, i32 18, i32 19>
+ %48 = zext <16 x i8> %47 to <16 x i32>
+ %49 = sub nsw <16 x i32> %39, %48
+ %50 = shl nsw <16 x i32> %49, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
+ %51 = add nsw <16 x i32> %50, %30
+ %reorder = shufflevector <16 x i32> %51, <16 x i32> poison, <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 13, i32 12, i32 15, i32 14>
+ %52 = add nsw <16 x i32> %51, %reorder
+ %53 = sub nsw <16 x i32> %51, %reorder
+ %54 = shufflevector <16 x i32> %52, <16 x i32> %53, <16 x i32> <i32 3, i32 7, i32 11, i32 15, i32 22, i32 18, i32 26, i32 30, i32 5, i32 1, i32 9, i32 13, i32 20, i32 16, i32 24, i32 28>
+ %reorder191 = shufflevector <16 x i32> %52, <16 x i32> %53, <16 x i32> <i32 1, i32 5, i32 9, i32 13, i32 20, i32 16, i32 24, i32 28, i32 7, i32 3, i32 11, i32 15, i32 22, i32 18, i32 26, i32 30>
+ %55 = add nsw <16 x i32> %54, %reorder191
+ %56 = sub nsw <16 x i32> %54, %reorder191
+ %57 = shufflevector <16 x i32> %55, <16 x i32> %56, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+ %reorder192 = shufflevector <16 x i32> %55, <16 x i32> %56, <16 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 25, i32 24, i32 27, i32 26, i32 29, i32 28, i32 31, i32 30>
+ %58 = add nsw <16 x i32> %57, %reorder192
+ %59 = sub nsw <16 x i32> %57, %reorder192
+ %60 = shufflevector <16 x i32> %58, <16 x i32> %59, <16 x i32> <i32 0, i32 17, i32 2, i32 19, i32 20, i32 5, i32 6, i32 23, i32 24, i32 9, i32 10, i32 27, i32 28, i32 13, i32 14, i32 31>
+ %reorder193 = shufflevector <16 x i32> %58, <16 x i32> %59, <16 x i32> <i32 2, i32 19, i32 0, i32 17, i32 23, i32 6, i32 5, i32 20, i32 27, i32 10, i32 9, i32 24, i32 31, i32 14, i32 13, i32 28>
+ %61 = add nsw <16 x i32> %60, %reorder193
+ %62 = sub nsw <16 x i32> %60, %reorder193
+ %63 = shufflevector <16 x i32> %61, <16 x i32> %62, <16 x i32> <i32 0, i32 1, i32 18, i32 19, i32 4, i32 5, i32 22, i32 23, i32 8, i32 9, i32 26, i32 27, i32 12, i32 13, i32 30, i32 31>
+ %64 = lshr <16 x i32> %63, <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
+ %65 = and <16 x i32> %64, <i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537>
+ %66 = mul nuw <16 x i32> %65, <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535>
+ %67 = add <16 x i32> %66, %63
+ %68 = xor <16 x i32> %67, %66
+ %69 = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %68)
+ %conv118 = and i32 %69, 65535
+ %shr = lshr i32 %69, 16
+ %add119 = add nuw nsw i32 %conv118, %shr
+ %shr120 = lshr i32 %add119, 1
+ ret i32 %shr120
+}
+
+declare i32 @llvm.vector.reduce.add.v16i32(<16 x i32>)
More information about the llvm-commits
mailing list