[llvm] bc39a7a - [LowerMatrixIntrinsics] Fix test expectations (NFC)

Nikita Popov via llvm-commits llvm-commits at lists.llvm.org
Tue Jul 18 02:21:24 PDT 2023


Author: Nikita Popov
Date: 2023-07-18T11:21:11+02:00
New Revision: bc39a7a5e40860bfdb1c9fe9fd319d33a29a9a8e

URL: https://github.com/llvm/llvm-project/commit/bc39a7a5e40860bfdb1c9fe9fd319d33a29a9a8e
DIFF: https://github.com/llvm/llvm-project/commit/bc39a7a5e40860bfdb1c9fe9fd319d33a29a9a8e.diff

LOG: [LowerMatrixIntrinsics] Fix test expectations (NFC)

Some of the test expectation were incorrectly changed in
23c21759458014fc4d7cbea45b6fbe7349a0a4fd. Regenerate the tests.

Added: 
    

Modified: 
    llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-loops.ll
    llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-multiple-blocks.ll
    llvm/test/Transforms/LowerMatrixIntrinsics/transpose-opts.ll

Removed: 
    


################################################################################
diff  --git a/llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-loops.ll b/llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-loops.ll
index e698bd27d921ec..2798f114d61900 100644
--- a/llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-loops.ll
+++ b/llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-loops.ll
@@ -40,11 +40,11 @@ define void @multiply_noalias_4x4(ptr noalias %A, ptr noalias %B, ptr noalias %C
 ; CHECK-NEXT:    [[COL_LOAD4:%.*]] = load <2 x double>, ptr [[VEC_GEP3]], align 8
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x double> [[COL_LOAD2]], <2 x double> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP6:%.*]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD]], <2 x double> [[SPLAT_SPLAT]], <2 x double> [[RESULT_VEC_0]])
-; CHECK-NEXT:    [[SPLAT_SPLAT8:%.*]] = shufflevector <2 x double> [[COL_LOAD2]], <2 x double> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT8:%.*]] = shufflevector <2 x double> [[COL_LOAD2]], <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP7]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD1]], <2 x double> [[SPLAT_SPLAT8]], <2 x double> [[TMP6]])
 ; CHECK-NEXT:    [[SPLAT_SPLAT12:%.*]] = shufflevector <2 x double> [[COL_LOAD4]], <2 x double> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP8:%.*]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD]], <2 x double> [[SPLAT_SPLAT12]], <2 x double> [[RESULT_VEC_1]])
-; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <2 x double> [[COL_LOAD4]], <2 x double> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <2 x double> [[COL_LOAD4]], <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP9]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD1]], <2 x double> [[SPLAT_SPLAT15]], <2 x double> [[TMP8]])
 ; CHECK-NEXT:    br label [[INNER_LATCH]]
 ; CHECK:       inner.latch:
@@ -117,13 +117,13 @@ define void @multiply_noalias_2x4(ptr noalias %A, ptr noalias %B, ptr noalias %C
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x i64> [[COL_LOAD2]], <2 x i64> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP6:%.*]] = mul <2 x i64> [[COL_LOAD]], [[SPLAT_SPLAT]]
 ; CHECK-NEXT:    [[TMP7:%.*]] = add <2 x i64> [[RESULT_VEC_0]], [[TMP6]]
-; CHECK-NEXT:    [[SPLAT_SPLAT8:%.*]] = shufflevector <2 x i64> [[COL_LOAD2]], <2 x i64> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT8:%.*]] = shufflevector <2 x i64> [[COL_LOAD2]], <2 x i64> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP8:%.*]] = mul <2 x i64> [[COL_LOAD1]], [[SPLAT_SPLAT8]]
 ; CHECK-NEXT:    [[TMP9]] = add <2 x i64> [[TMP7]], [[TMP8]]
 ; CHECK-NEXT:    [[SPLAT_SPLAT12:%.*]] = shufflevector <2 x i64> [[COL_LOAD4]], <2 x i64> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP10:%.*]] = mul <2 x i64> [[COL_LOAD]], [[SPLAT_SPLAT12]]
 ; CHECK-NEXT:    [[TMP11:%.*]] = add <2 x i64> [[RESULT_VEC_1]], [[TMP10]]
-; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <2 x i64> [[COL_LOAD4]], <2 x i64> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <2 x i64> [[COL_LOAD4]], <2 x i64> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP12:%.*]] = mul <2 x i64> [[COL_LOAD1]], [[SPLAT_SPLAT15]]
 ; CHECK-NEXT:    [[TMP13]] = add <2 x i64> [[TMP11]], [[TMP12]]
 ; CHECK-NEXT:    br label [[INNER_LATCH]]
@@ -203,13 +203,13 @@ define void @multiply_noalias_4x2_2x8(ptr noalias %A, ptr noalias %B, ptr noalia
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x i64> [[COL_LOAD2]], <2 x i64> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP6:%.*]] = mul <2 x i64> [[COL_LOAD]], [[SPLAT_SPLAT]]
 ; CHECK-NEXT:    [[TMP7:%.*]] = add <2 x i64> [[RESULT_VEC_0]], [[TMP6]]
-; CHECK-NEXT:    [[SPLAT_SPLAT8:%.*]] = shufflevector <2 x i64> [[COL_LOAD2]], <2 x i64> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT8:%.*]] = shufflevector <2 x i64> [[COL_LOAD2]], <2 x i64> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP8:%.*]] = mul <2 x i64> [[COL_LOAD1]], [[SPLAT_SPLAT8]]
 ; CHECK-NEXT:    [[TMP9]] = add <2 x i64> [[TMP7]], [[TMP8]]
 ; CHECK-NEXT:    [[SPLAT_SPLAT12:%.*]] = shufflevector <2 x i64> [[COL_LOAD4]], <2 x i64> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP10:%.*]] = mul <2 x i64> [[COL_LOAD]], [[SPLAT_SPLAT12]]
 ; CHECK-NEXT:    [[TMP11:%.*]] = add <2 x i64> [[RESULT_VEC_1]], [[TMP10]]
-; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <2 x i64> [[COL_LOAD4]], <2 x i64> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <2 x i64> [[COL_LOAD4]], <2 x i64> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP12:%.*]] = mul <2 x i64> [[COL_LOAD1]], [[SPLAT_SPLAT15]]
 ; CHECK-NEXT:    [[TMP13]] = add <2 x i64> [[TMP11]], [[TMP12]]
 ; CHECK-NEXT:    br label [[INNER_LATCH]]
@@ -319,11 +319,11 @@ define void @multiply_alias_2x2(ptr %A, ptr %B, ptr %C) {
 ; CHECK-NEXT:    [[COL_LOAD11:%.*]] = load <2 x float>, ptr [[VEC_GEP10]], align 4
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x float> [[COL_LOAD9]], <2 x float> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP14:%.*]] = call contract <2 x float> @llvm.fmuladd.v2f32(<2 x float> [[COL_LOAD]], <2 x float> [[SPLAT_SPLAT]], <2 x float> [[RESULT_VEC_0]])
-; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <2 x float> [[COL_LOAD9]], <2 x float> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <2 x float> [[COL_LOAD9]], <2 x float> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP15]] = call contract <2 x float> @llvm.fmuladd.v2f32(<2 x float> [[COL_LOAD8]], <2 x float> [[SPLAT_SPLAT15]], <2 x float> [[TMP14]])
 ; CHECK-NEXT:    [[SPLAT_SPLAT19:%.*]] = shufflevector <2 x float> [[COL_LOAD11]], <2 x float> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP16:%.*]] = call contract <2 x float> @llvm.fmuladd.v2f32(<2 x float> [[COL_LOAD]], <2 x float> [[SPLAT_SPLAT19]], <2 x float> [[RESULT_VEC_1]])
-; CHECK-NEXT:    [[SPLAT_SPLAT22:%.*]] = shufflevector <2 x float> [[COL_LOAD11]], <2 x float> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT22:%.*]] = shufflevector <2 x float> [[COL_LOAD11]], <2 x float> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP17]] = call contract <2 x float> @llvm.fmuladd.v2f32(<2 x float> [[COL_LOAD8]], <2 x float> [[SPLAT_SPLAT22]], <2 x float> [[TMP16]])
 ; CHECK-NEXT:    br label [[INNER_LATCH]]
 ; CHECK:       inner.latch:

diff  --git a/llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-multiple-blocks.ll b/llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-multiple-blocks.ll
index f7ac2d321439a5..2f4846a2840e50 100644
--- a/llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-multiple-blocks.ll
+++ b/llvm/test/Transforms/LowerMatrixIntrinsics/multiply-fused-multiple-blocks.ll
@@ -55,11 +55,11 @@ define void @test(ptr %A, ptr %B, ptr %C, i1 %cond) {
 ; CHECK-NEXT:    [[COL_LOAD11:%.*]] = load <2 x double>, ptr [[VEC_GEP10]], align 8
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x double> [[COL_LOAD9]], <2 x double> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP8:%.*]] = fmul contract <2 x double> [[COL_LOAD]], [[SPLAT_SPLAT]]
-; CHECK-NEXT:    [[SPLAT_SPLAT14:%.*]] = shufflevector <2 x double> [[COL_LOAD9]], <2 x double> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT14:%.*]] = shufflevector <2 x double> [[COL_LOAD9]], <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP9:%.*]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD8]], <2 x double> [[SPLAT_SPLAT14]], <2 x double> [[TMP8]])
 ; CHECK-NEXT:    [[SPLAT_SPLAT17:%.*]] = shufflevector <2 x double> [[COL_LOAD11]], <2 x double> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP10:%.*]] = fmul contract <2 x double> [[COL_LOAD]], [[SPLAT_SPLAT17]]
-; CHECK-NEXT:    [[SPLAT_SPLAT20:%.*]] = shufflevector <2 x double> [[COL_LOAD11]], <2 x double> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT20:%.*]] = shufflevector <2 x double> [[COL_LOAD11]], <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP11:%.*]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD8]], <2 x double> [[SPLAT_SPLAT20]], <2 x double> [[TMP10]])
 ; CHECK-NEXT:    store <2 x double> [[TMP9]], ptr [[C]], align 8
 ; CHECK-NEXT:    [[VEC_GEP21:%.*]] = getelementptr double, ptr [[C]], i64 3
@@ -71,13 +71,13 @@ define void @test(ptr %A, ptr %B, ptr %C, i1 %cond) {
 ; CHECK-NEXT:    [[COL_LOAD25:%.*]] = load <2 x double>, ptr [[TMP7]], align 8
 ; CHECK-NEXT:    [[VEC_GEP26:%.*]] = getelementptr double, ptr [[TMP7]], i64 2
 ; CHECK-NEXT:    [[COL_LOAD27:%.*]] = load <2 x double>, ptr [[VEC_GEP26]], align 8
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT29:%.*]] = shufflevector <2 x double> [[COL_LOAD25]], <2 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT29:%.*]] = shufflevector <2 x double> [[COL_LOAD25]], <2 x double> undef, <1 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP13:%.*]] = fmul contract <1 x double> [[COL_LOAD22]], [[SPLAT_SPLATINSERT29]]
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT32:%.*]] = shufflevector <2 x double> [[COL_LOAD25]], <2 x double> poison, <1 x i32> <i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT32:%.*]] = shufflevector <2 x double> [[COL_LOAD25]], <2 x double> undef, <1 x i32> <i32 1>
 ; CHECK-NEXT:    [[TMP14:%.*]] = call contract <1 x double> @llvm.fmuladd.v1f64(<1 x double> [[COL_LOAD24]], <1 x double> [[SPLAT_SPLATINSERT32]], <1 x double> [[TMP13]])
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT35:%.*]] = shufflevector <2 x double> [[COL_LOAD27]], <2 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT35:%.*]] = shufflevector <2 x double> [[COL_LOAD27]], <2 x double> undef, <1 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP15:%.*]] = fmul contract <1 x double> [[COL_LOAD22]], [[SPLAT_SPLATINSERT35]]
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT38:%.*]] = shufflevector <2 x double> [[COL_LOAD27]], <2 x double> poison, <1 x i32> <i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT38:%.*]] = shufflevector <2 x double> [[COL_LOAD27]], <2 x double> undef, <1 x i32> <i32 1>
 ; CHECK-NEXT:    [[TMP16:%.*]] = call contract <1 x double> @llvm.fmuladd.v1f64(<1 x double> [[COL_LOAD24]], <1 x double> [[SPLAT_SPLATINSERT38]], <1 x double> [[TMP15]])
 ; CHECK-NEXT:    [[TMP17:%.*]] = getelementptr double, ptr [[C]], i64 2
 ; CHECK-NEXT:    store <1 x double> [[TMP14]], ptr [[TMP17]], align 8
@@ -90,7 +90,7 @@ define void @test(ptr %A, ptr %B, ptr %C, i1 %cond) {
 ; CHECK-NEXT:    [[COL_LOAD44:%.*]] = load <2 x double>, ptr [[TMP18]], align 8
 ; CHECK-NEXT:    [[SPLAT_SPLAT47:%.*]] = shufflevector <2 x double> [[COL_LOAD44]], <2 x double> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP19:%.*]] = fmul contract <2 x double> [[COL_LOAD41]], [[SPLAT_SPLAT47]]
-; CHECK-NEXT:    [[SPLAT_SPLAT50:%.*]] = shufflevector <2 x double> [[COL_LOAD44]], <2 x double> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT50:%.*]] = shufflevector <2 x double> [[COL_LOAD44]], <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP20:%.*]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD43]], <2 x double> [[SPLAT_SPLAT50]], <2 x double> [[TMP19]])
 ; CHECK-NEXT:    [[TMP21:%.*]] = getelementptr double, ptr [[C]], i64 6
 ; CHECK-NEXT:    store <2 x double> [[TMP20]], ptr [[TMP21]], align 8
@@ -100,9 +100,9 @@ define void @test(ptr %A, ptr %B, ptr %C, i1 %cond) {
 ; CHECK-NEXT:    [[COL_LOAD53:%.*]] = load <1 x double>, ptr [[VEC_GEP52]], align 8
 ; CHECK-NEXT:    [[TMP23:%.*]] = getelementptr double, ptr [[TMP7]], i64 4
 ; CHECK-NEXT:    [[COL_LOAD54:%.*]] = load <2 x double>, ptr [[TMP23]], align 8
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT56:%.*]] = shufflevector <2 x double> [[COL_LOAD54]], <2 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT56:%.*]] = shufflevector <2 x double> [[COL_LOAD54]], <2 x double> undef, <1 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP24:%.*]] = fmul contract <1 x double> [[COL_LOAD51]], [[SPLAT_SPLATINSERT56]]
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT59:%.*]] = shufflevector <2 x double> [[COL_LOAD54]], <2 x double> poison, <1 x i32> <i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT59:%.*]] = shufflevector <2 x double> [[COL_LOAD54]], <2 x double> undef, <1 x i32> <i32 1>
 ; CHECK-NEXT:    [[TMP25:%.*]] = call contract <1 x double> @llvm.fmuladd.v1f64(<1 x double> [[COL_LOAD53]], <1 x double> [[SPLAT_SPLATINSERT59]], <1 x double> [[TMP24]])
 ; CHECK-NEXT:    [[TMP26:%.*]] = getelementptr double, ptr [[C]], i64 8
 ; CHECK-NEXT:    store <1 x double> [[TMP25]], ptr [[TMP26]], align 8
@@ -163,11 +163,11 @@ define void @test(ptr %A, ptr %B, ptr %C, i1 %cond) {
 ; CHECK-NEXT:    [[COL_LOAD80:%.*]] = load <2 x double>, ptr [[VEC_GEP79]], align 8
 ; CHECK-NEXT:    [[SPLAT_SPLAT83:%.*]] = shufflevector <2 x double> [[COL_LOAD78]], <2 x double> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP40:%.*]] = fmul contract <2 x double> [[COL_LOAD75]], [[SPLAT_SPLAT83]]
-; CHECK-NEXT:    [[SPLAT_SPLAT86:%.*]] = shufflevector <2 x double> [[COL_LOAD78]], <2 x double> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT86:%.*]] = shufflevector <2 x double> [[COL_LOAD78]], <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP41:%.*]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD77]], <2 x double> [[SPLAT_SPLAT86]], <2 x double> [[TMP40]])
 ; CHECK-NEXT:    [[SPLAT_SPLAT89:%.*]] = shufflevector <2 x double> [[COL_LOAD80]], <2 x double> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP42:%.*]] = fmul contract <2 x double> [[COL_LOAD75]], [[SPLAT_SPLAT89]]
-; CHECK-NEXT:    [[SPLAT_SPLAT92:%.*]] = shufflevector <2 x double> [[COL_LOAD80]], <2 x double> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT92:%.*]] = shufflevector <2 x double> [[COL_LOAD80]], <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP43:%.*]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD77]], <2 x double> [[SPLAT_SPLAT92]], <2 x double> [[TMP42]])
 ; CHECK-NEXT:    store <2 x double> [[TMP41]], ptr [[C]], align 8
 ; CHECK-NEXT:    [[VEC_GEP93:%.*]] = getelementptr double, ptr [[C]], i64 3
@@ -179,13 +179,13 @@ define void @test(ptr %A, ptr %B, ptr %C, i1 %cond) {
 ; CHECK-NEXT:    [[COL_LOAD97:%.*]] = load <2 x double>, ptr [[TMP39]], align 8
 ; CHECK-NEXT:    [[VEC_GEP98:%.*]] = getelementptr double, ptr [[TMP39]], i64 2
 ; CHECK-NEXT:    [[COL_LOAD99:%.*]] = load <2 x double>, ptr [[VEC_GEP98]], align 8
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT101:%.*]] = shufflevector <2 x double> [[COL_LOAD97]], <2 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT101:%.*]] = shufflevector <2 x double> [[COL_LOAD97]], <2 x double> undef, <1 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP45:%.*]] = fmul contract <1 x double> [[COL_LOAD94]], [[SPLAT_SPLATINSERT101]]
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT104:%.*]] = shufflevector <2 x double> [[COL_LOAD97]], <2 x double> poison, <1 x i32> <i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT104:%.*]] = shufflevector <2 x double> [[COL_LOAD97]], <2 x double> undef, <1 x i32> <i32 1>
 ; CHECK-NEXT:    [[TMP46:%.*]] = call contract <1 x double> @llvm.fmuladd.v1f64(<1 x double> [[COL_LOAD96]], <1 x double> [[SPLAT_SPLATINSERT104]], <1 x double> [[TMP45]])
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT107:%.*]] = shufflevector <2 x double> [[COL_LOAD99]], <2 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT107:%.*]] = shufflevector <2 x double> [[COL_LOAD99]], <2 x double> undef, <1 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP47:%.*]] = fmul contract <1 x double> [[COL_LOAD94]], [[SPLAT_SPLATINSERT107]]
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT110:%.*]] = shufflevector <2 x double> [[COL_LOAD99]], <2 x double> poison, <1 x i32> <i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT110:%.*]] = shufflevector <2 x double> [[COL_LOAD99]], <2 x double> undef, <1 x i32> <i32 1>
 ; CHECK-NEXT:    [[TMP48:%.*]] = call contract <1 x double> @llvm.fmuladd.v1f64(<1 x double> [[COL_LOAD96]], <1 x double> [[SPLAT_SPLATINSERT110]], <1 x double> [[TMP47]])
 ; CHECK-NEXT:    [[TMP49:%.*]] = getelementptr double, ptr [[C]], i64 2
 ; CHECK-NEXT:    store <1 x double> [[TMP46]], ptr [[TMP49]], align 8
@@ -198,7 +198,7 @@ define void @test(ptr %A, ptr %B, ptr %C, i1 %cond) {
 ; CHECK-NEXT:    [[COL_LOAD116:%.*]] = load <2 x double>, ptr [[TMP50]], align 8
 ; CHECK-NEXT:    [[SPLAT_SPLAT119:%.*]] = shufflevector <2 x double> [[COL_LOAD116]], <2 x double> poison, <2 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP51:%.*]] = fmul contract <2 x double> [[COL_LOAD113]], [[SPLAT_SPLAT119]]
-; CHECK-NEXT:    [[SPLAT_SPLAT122:%.*]] = shufflevector <2 x double> [[COL_LOAD116]], <2 x double> poison, <2 x i32> <i32 1, i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLAT122:%.*]] = shufflevector <2 x double> [[COL_LOAD116]], <2 x double> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK-NEXT:    [[TMP52:%.*]] = call contract <2 x double> @llvm.fmuladd.v2f64(<2 x double> [[COL_LOAD115]], <2 x double> [[SPLAT_SPLAT122]], <2 x double> [[TMP51]])
 ; CHECK-NEXT:    [[TMP53:%.*]] = getelementptr double, ptr [[C]], i64 6
 ; CHECK-NEXT:    store <2 x double> [[TMP52]], ptr [[TMP53]], align 8
@@ -208,9 +208,9 @@ define void @test(ptr %A, ptr %B, ptr %C, i1 %cond) {
 ; CHECK-NEXT:    [[COL_LOAD125:%.*]] = load <1 x double>, ptr [[VEC_GEP124]], align 8
 ; CHECK-NEXT:    [[TMP55:%.*]] = getelementptr double, ptr [[TMP39]], i64 4
 ; CHECK-NEXT:    [[COL_LOAD126:%.*]] = load <2 x double>, ptr [[TMP55]], align 8
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT128:%.*]] = shufflevector <2 x double> [[COL_LOAD126]], <2 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT128:%.*]] = shufflevector <2 x double> [[COL_LOAD126]], <2 x double> undef, <1 x i32> zeroinitializer
 ; CHECK-NEXT:    [[TMP56:%.*]] = fmul contract <1 x double> [[COL_LOAD123]], [[SPLAT_SPLATINSERT128]]
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT131:%.*]] = shufflevector <2 x double> [[COL_LOAD126]], <2 x double> poison, <1 x i32> <i32 1>
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT131:%.*]] = shufflevector <2 x double> [[COL_LOAD126]], <2 x double> undef, <1 x i32> <i32 1>
 ; CHECK-NEXT:    [[TMP57:%.*]] = call contract <1 x double> @llvm.fmuladd.v1f64(<1 x double> [[COL_LOAD125]], <1 x double> [[SPLAT_SPLATINSERT131]], <1 x double> [[TMP56]])
 ; CHECK-NEXT:    [[TMP58:%.*]] = getelementptr double, ptr [[C]], i64 8
 ; CHECK-NEXT:    store <1 x double> [[TMP57]], ptr [[TMP58]], align 8

diff  --git a/llvm/test/Transforms/LowerMatrixIntrinsics/transpose-opts.ll b/llvm/test/Transforms/LowerMatrixIntrinsics/transpose-opts.ll
index 036e68b1242778..57d038e5cf9473 100644
--- a/llvm/test/Transforms/LowerMatrixIntrinsics/transpose-opts.ll
+++ b/llvm/test/Transforms/LowerMatrixIntrinsics/transpose-opts.ll
@@ -31,14 +31,14 @@ define void @double_transpose(ptr %A, ptr %B) {
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <3 x double>, ptr [[A:%.*]], align 16
 ; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr double, ptr [[A]], i64 3
-; CHECK-NEXT:    [[COL_LOAD2:%.*]] = load <3 x double>, ptr [[VEC_GEP]], align 8
-; CHECK-NEXT:    [[VEC_GEP3:%.*]] = getelementptr double, ptr [[A]], i64 6
-; CHECK-NEXT:    [[COL_LOAD5:%.*]] = load <3 x double>, ptr [[VEC_GEP3]], align 16
+; CHECK-NEXT:    [[COL_LOAD1:%.*]] = load <3 x double>, ptr [[VEC_GEP]], align 8
+; CHECK-NEXT:    [[VEC_GEP2:%.*]] = getelementptr double, ptr [[A]], i64 6
+; CHECK-NEXT:    [[COL_LOAD3:%.*]] = load <3 x double>, ptr [[VEC_GEP2]], align 16
 ; CHECK-NEXT:    store <3 x double> [[COL_LOAD]], ptr [[B:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP7:%.*]] = getelementptr double, ptr [[B]], i64 3
-; CHECK-NEXT:    store <3 x double> [[COL_LOAD2]], ptr [[VEC_GEP7]], align 8
-; CHECK-NEXT:    [[VEC_GEP9:%.*]] = getelementptr double, ptr [[B]], i64 6
-; CHECK-NEXT:    store <3 x double> [[COL_LOAD5]], ptr [[VEC_GEP9]], align 16
+; CHECK-NEXT:    [[VEC_GEP4:%.*]] = getelementptr double, ptr [[B]], i64 3
+; CHECK-NEXT:    store <3 x double> [[COL_LOAD1]], ptr [[VEC_GEP4]], align 8
+; CHECK-NEXT:    [[VEC_GEP5:%.*]] = getelementptr double, ptr [[B]], i64 6
+; CHECK-NEXT:    store <3 x double> [[COL_LOAD3]], ptr [[VEC_GEP5]], align 16
 ; CHECK-NEXT:    ret void
 ;
 entry:
@@ -75,190 +75,190 @@ define void @multiply_ntt(ptr %A, ptr %B, ptr %C, ptr %R) {
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <2 x double>, ptr [[A:%.*]], align 16
 ; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr double, ptr [[A]], i64 2
-; CHECK-NEXT:    [[COL_LOAD40:%.*]] = load <2 x double>, ptr [[VEC_GEP]], align 16
-; CHECK-NEXT:    [[VEC_GEP41:%.*]] = getelementptr double, ptr [[A]], i64 4
-; CHECK-NEXT:    [[COL_LOAD43:%.*]] = load <2 x double>, ptr [[VEC_GEP41]], align 16
-; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <2 x double> [[COL_LOAD]], <2 x double> [[COL_LOAD40]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <2 x double> [[COL_LOAD43]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x double> [[TMP1]], <4 x double> [[TMP2]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
-; CHECK-NEXT:    [[COL_LOAD45:%.*]] = load <2 x double>, ptr [[B:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP46:%.*]] = getelementptr double, ptr [[B]], i64 2
-; CHECK-NEXT:    [[COL_LOAD48:%.*]] = load <2 x double>, ptr [[VEC_GEP46]], align 16
-; CHECK-NEXT:    [[VEC_GEP49:%.*]] = getelementptr double, ptr [[B]], i64 4
-; CHECK-NEXT:    [[COL_LOAD51:%.*]] = load <2 x double>, ptr [[VEC_GEP49]], align 16
-; CHECK-NEXT:    [[COL_LOAD53:%.*]] = load <4 x double>, ptr [[C:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP54:%.*]] = getelementptr double, ptr [[C]], i64 4
-; CHECK-NEXT:    [[COL_LOAD56:%.*]] = load <4 x double>, ptr [[VEC_GEP54]], align 16
-; CHECK-NEXT:    [[BLOCK57:%.*]] = shufflevector <4 x double> [[COL_LOAD53]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <2 x double> [[COL_LOAD45]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT58:%.*]] = insertelement <2 x double> poison, double [[TMP6]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT59:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT58]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP7:%.*]] = fmul <2 x double> [[BLOCK57]], [[SPLAT_SPLAT59]]
-; CHECK-NEXT:    [[BLOCK60:%.*]] = shufflevector <4 x double> [[COL_LOAD56]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP8:%.*]] = extractelement <2 x double> [[COL_LOAD45]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT61:%.*]] = insertelement <2 x double> poison, double [[TMP8]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT62:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT61]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP9:%.*]] = fmul <2 x double> [[BLOCK60]], [[SPLAT_SPLAT62]]
-; CHECK-NEXT:    [[TMP10:%.*]] = fadd <2 x double> [[TMP7]], [[TMP9]]
-; CHECK-NEXT:    [[TMP11:%.*]] = shufflevector <2 x double> [[TMP10]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <4 x double> undef, <4 x double> [[TMP11]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
-; CHECK-NEXT:    [[BLOCK63:%.*]] = shufflevector <4 x double> [[COL_LOAD53]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP13:%.*]] = extractelement <2 x double> [[COL_LOAD45]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT64:%.*]] = insertelement <2 x double> poison, double [[TMP13]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT65:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT64]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP14:%.*]] = fmul <2 x double> [[BLOCK63]], [[SPLAT_SPLAT65]]
-; CHECK-NEXT:    [[BLOCK66:%.*]] = shufflevector <4 x double> [[COL_LOAD56]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP15:%.*]] = extractelement <2 x double> [[COL_LOAD45]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT67:%.*]] = insertelement <2 x double> poison, double [[TMP15]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT68:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT67]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP16:%.*]] = fmul <2 x double> [[BLOCK66]], [[SPLAT_SPLAT68]]
-; CHECK-NEXT:    [[TMP17:%.*]] = fadd <2 x double> [[TMP14]], [[TMP16]]
-; CHECK-NEXT:    [[TMP18:%.*]] = shufflevector <2 x double> [[TMP17]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP19:%.*]] = shufflevector <4 x double> [[TMP12]], <4 x double> [[TMP18]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    [[BLOCK69:%.*]] = shufflevector <4 x double> [[COL_LOAD53]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP20:%.*]] = extractelement <2 x double> [[COL_LOAD48]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT70:%.*]] = insertelement <2 x double> poison, double [[TMP20]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT71:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT70]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP21:%.*]] = fmul <2 x double> [[BLOCK69]], [[SPLAT_SPLAT71]]
-; CHECK-NEXT:    [[BLOCK72:%.*]] = shufflevector <4 x double> [[COL_LOAD56]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP22:%.*]] = extractelement <2 x double> [[COL_LOAD48]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT73:%.*]] = insertelement <2 x double> poison, double [[TMP22]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT74:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT73]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP23:%.*]] = fmul <2 x double> [[BLOCK72]], [[SPLAT_SPLAT74]]
-; CHECK-NEXT:    [[TMP24:%.*]] = fadd <2 x double> [[TMP21]], [[TMP23]]
-; CHECK-NEXT:    [[TMP25:%.*]] = shufflevector <2 x double> [[TMP24]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP26:%.*]] = shufflevector <4 x double> undef, <4 x double> [[TMP25]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
-; CHECK-NEXT:    [[BLOCK75:%.*]] = shufflevector <4 x double> [[COL_LOAD53]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP27:%.*]] = extractelement <2 x double> [[COL_LOAD48]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT76:%.*]] = insertelement <2 x double> poison, double [[TMP27]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT77:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT76]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP28:%.*]] = fmul <2 x double> [[BLOCK75]], [[SPLAT_SPLAT77]]
-; CHECK-NEXT:    [[BLOCK78:%.*]] = shufflevector <4 x double> [[COL_LOAD56]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP29:%.*]] = extractelement <2 x double> [[COL_LOAD48]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT79:%.*]] = insertelement <2 x double> poison, double [[TMP29]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT80:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT79]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP30:%.*]] = fmul <2 x double> [[BLOCK78]], [[SPLAT_SPLAT80]]
-; CHECK-NEXT:    [[TMP31:%.*]] = fadd <2 x double> [[TMP28]], [[TMP30]]
-; CHECK-NEXT:    [[TMP32:%.*]] = shufflevector <2 x double> [[TMP31]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP33:%.*]] = shufflevector <4 x double> [[TMP26]], <4 x double> [[TMP32]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    [[BLOCK81:%.*]] = shufflevector <4 x double> [[COL_LOAD53]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP34:%.*]] = extractelement <2 x double> [[COL_LOAD51]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT82:%.*]] = insertelement <2 x double> poison, double [[TMP34]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT83:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT82]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP35:%.*]] = fmul <2 x double> [[BLOCK81]], [[SPLAT_SPLAT83]]
-; CHECK-NEXT:    [[BLOCK84:%.*]] = shufflevector <4 x double> [[COL_LOAD56]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP36:%.*]] = extractelement <2 x double> [[COL_LOAD51]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT85:%.*]] = insertelement <2 x double> poison, double [[TMP36]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT86:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT85]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP37:%.*]] = fmul <2 x double> [[BLOCK84]], [[SPLAT_SPLAT86]]
-; CHECK-NEXT:    [[TMP38:%.*]] = fadd <2 x double> [[TMP35]], [[TMP37]]
-; CHECK-NEXT:    [[TMP39:%.*]] = shufflevector <2 x double> [[TMP38]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP40:%.*]] = shufflevector <4 x double> undef, <4 x double> [[TMP39]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
-; CHECK-NEXT:    [[BLOCK87:%.*]] = shufflevector <4 x double> [[COL_LOAD53]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP41:%.*]] = extractelement <2 x double> [[COL_LOAD51]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT88:%.*]] = insertelement <2 x double> poison, double [[TMP41]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT89:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT88]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP42:%.*]] = fmul <2 x double> [[BLOCK87]], [[SPLAT_SPLAT89]]
-; CHECK-NEXT:    [[BLOCK90:%.*]] = shufflevector <4 x double> [[COL_LOAD56]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP43:%.*]] = extractelement <2 x double> [[COL_LOAD51]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT91:%.*]] = insertelement <2 x double> poison, double [[TMP43]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT92:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT91]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP44:%.*]] = fmul <2 x double> [[BLOCK90]], [[SPLAT_SPLAT92]]
-; CHECK-NEXT:    [[TMP45:%.*]] = fadd <2 x double> [[TMP42]], [[TMP44]]
-; CHECK-NEXT:    [[TMP46:%.*]] = shufflevector <2 x double> [[TMP45]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP47:%.*]] = shufflevector <4 x double> [[TMP40]], <4 x double> [[TMP46]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    [[TMP48:%.*]] = shufflevector <4 x double> [[TMP19]], <4 x double> [[TMP33]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[TMP49:%.*]] = shufflevector <4 x double> [[TMP47]], <4 x double> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP50:%.*]] = shufflevector <8 x double> [[TMP48]], <8 x double> [[TMP49]], <12 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>
-; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <6 x double> [[TMP3]], <6 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[SPLIT1:%.*]] = shufflevector <6 x double> [[TMP3]], <6 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[SPLIT2:%.*]] = shufflevector <6 x double> [[TMP3]], <6 x double> poison, <2 x i32> <i32 4, i32 5>
-; CHECK-NEXT:    [[SPLIT3:%.*]] = shufflevector <12 x double> [[TMP50]], <12 x double> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[SPLIT4:%.*]] = shufflevector <12 x double> [[TMP50]], <12 x double> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[SPLIT5:%.*]] = shufflevector <12 x double> [[TMP50]], <12 x double> poison, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; CHECK-NEXT:    [[COL_LOAD39:%.*]] = load <2 x double>, ptr [[VEC_GEP]], align 16
+; CHECK-NEXT:    [[VEC_GEP40:%.*]] = getelementptr double, ptr [[A]], i64 4
+; CHECK-NEXT:    [[COL_LOAD41:%.*]] = load <2 x double>, ptr [[VEC_GEP40]], align 16
+; CHECK-NEXT:    [[TMP0:%.*]] = shufflevector <2 x double> [[COL_LOAD]], <2 x double> [[COL_LOAD39]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <2 x double> [[COL_LOAD41]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[TMP0]], <4 x double> [[TMP1]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
+; CHECK-NEXT:    [[COL_LOAD42:%.*]] = load <2 x double>, ptr [[B:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP43:%.*]] = getelementptr double, ptr [[B]], i64 2
+; CHECK-NEXT:    [[COL_LOAD44:%.*]] = load <2 x double>, ptr [[VEC_GEP43]], align 16
+; CHECK-NEXT:    [[VEC_GEP45:%.*]] = getelementptr double, ptr [[B]], i64 4
+; CHECK-NEXT:    [[COL_LOAD46:%.*]] = load <2 x double>, ptr [[VEC_GEP45]], align 16
+; CHECK-NEXT:    [[COL_LOAD47:%.*]] = load <4 x double>, ptr [[C:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP48:%.*]] = getelementptr double, ptr [[C]], i64 4
+; CHECK-NEXT:    [[COL_LOAD49:%.*]] = load <4 x double>, ptr [[VEC_GEP48]], align 16
+; CHECK-NEXT:    [[BLOCK50:%.*]] = shufflevector <4 x double> [[COL_LOAD47]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[COL_LOAD42]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT51:%.*]] = insertelement <2 x double> poison, double [[TMP3]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT52:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT51]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP4:%.*]] = fmul <2 x double> [[BLOCK50]], [[SPLAT_SPLAT52]]
+; CHECK-NEXT:    [[BLOCK53:%.*]] = shufflevector <4 x double> [[COL_LOAD49]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <2 x double> [[COL_LOAD42]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT54:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT55:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT54]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP6:%.*]] = fmul <2 x double> [[BLOCK53]], [[SPLAT_SPLAT55]]
+; CHECK-NEXT:    [[TMP7:%.*]] = fadd <2 x double> [[TMP4]], [[TMP6]]
+; CHECK-NEXT:    [[TMP8:%.*]] = shufflevector <2 x double> [[TMP7]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP9:%.*]] = shufflevector <4 x double> poison, <4 x double> [[TMP8]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[BLOCK56:%.*]] = shufflevector <4 x double> [[COL_LOAD47]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP10:%.*]] = extractelement <2 x double> [[COL_LOAD42]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT57:%.*]] = insertelement <2 x double> poison, double [[TMP10]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT58:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT57]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP11:%.*]] = fmul <2 x double> [[BLOCK56]], [[SPLAT_SPLAT58]]
+; CHECK-NEXT:    [[BLOCK59:%.*]] = shufflevector <4 x double> [[COL_LOAD49]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP12:%.*]] = extractelement <2 x double> [[COL_LOAD42]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT60:%.*]] = insertelement <2 x double> poison, double [[TMP12]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT61:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT60]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP13:%.*]] = fmul <2 x double> [[BLOCK59]], [[SPLAT_SPLAT61]]
+; CHECK-NEXT:    [[TMP14:%.*]] = fadd <2 x double> [[TMP11]], [[TMP13]]
+; CHECK-NEXT:    [[TMP15:%.*]] = shufflevector <2 x double> [[TMP14]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP16:%.*]] = shufflevector <4 x double> [[TMP9]], <4 x double> [[TMP15]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
+; CHECK-NEXT:    [[BLOCK62:%.*]] = shufflevector <4 x double> [[COL_LOAD47]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP17:%.*]] = extractelement <2 x double> [[COL_LOAD44]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT63:%.*]] = insertelement <2 x double> poison, double [[TMP17]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT64:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT63]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP18:%.*]] = fmul <2 x double> [[BLOCK62]], [[SPLAT_SPLAT64]]
+; CHECK-NEXT:    [[BLOCK65:%.*]] = shufflevector <4 x double> [[COL_LOAD49]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP19:%.*]] = extractelement <2 x double> [[COL_LOAD44]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT66:%.*]] = insertelement <2 x double> poison, double [[TMP19]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT67:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT66]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP20:%.*]] = fmul <2 x double> [[BLOCK65]], [[SPLAT_SPLAT67]]
+; CHECK-NEXT:    [[TMP21:%.*]] = fadd <2 x double> [[TMP18]], [[TMP20]]
+; CHECK-NEXT:    [[TMP22:%.*]] = shufflevector <2 x double> [[TMP21]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP23:%.*]] = shufflevector <4 x double> poison, <4 x double> [[TMP22]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[BLOCK68:%.*]] = shufflevector <4 x double> [[COL_LOAD47]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP24:%.*]] = extractelement <2 x double> [[COL_LOAD44]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT69:%.*]] = insertelement <2 x double> poison, double [[TMP24]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT70:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT69]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP25:%.*]] = fmul <2 x double> [[BLOCK68]], [[SPLAT_SPLAT70]]
+; CHECK-NEXT:    [[BLOCK71:%.*]] = shufflevector <4 x double> [[COL_LOAD49]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP26:%.*]] = extractelement <2 x double> [[COL_LOAD44]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT72:%.*]] = insertelement <2 x double> poison, double [[TMP26]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT73:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT72]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP27:%.*]] = fmul <2 x double> [[BLOCK71]], [[SPLAT_SPLAT73]]
+; CHECK-NEXT:    [[TMP28:%.*]] = fadd <2 x double> [[TMP25]], [[TMP27]]
+; CHECK-NEXT:    [[TMP29:%.*]] = shufflevector <2 x double> [[TMP28]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP30:%.*]] = shufflevector <4 x double> [[TMP23]], <4 x double> [[TMP29]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
+; CHECK-NEXT:    [[BLOCK74:%.*]] = shufflevector <4 x double> [[COL_LOAD47]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP31:%.*]] = extractelement <2 x double> [[COL_LOAD46]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT75:%.*]] = insertelement <2 x double> poison, double [[TMP31]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT76:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT75]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP32:%.*]] = fmul <2 x double> [[BLOCK74]], [[SPLAT_SPLAT76]]
+; CHECK-NEXT:    [[BLOCK77:%.*]] = shufflevector <4 x double> [[COL_LOAD49]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP33:%.*]] = extractelement <2 x double> [[COL_LOAD46]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT78:%.*]] = insertelement <2 x double> poison, double [[TMP33]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT79:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT78]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP34:%.*]] = fmul <2 x double> [[BLOCK77]], [[SPLAT_SPLAT79]]
+; CHECK-NEXT:    [[TMP35:%.*]] = fadd <2 x double> [[TMP32]], [[TMP34]]
+; CHECK-NEXT:    [[TMP36:%.*]] = shufflevector <2 x double> [[TMP35]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP37:%.*]] = shufflevector <4 x double> poison, <4 x double> [[TMP36]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[BLOCK80:%.*]] = shufflevector <4 x double> [[COL_LOAD47]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP38:%.*]] = extractelement <2 x double> [[COL_LOAD46]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT81:%.*]] = insertelement <2 x double> poison, double [[TMP38]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT82:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT81]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP39:%.*]] = fmul <2 x double> [[BLOCK80]], [[SPLAT_SPLAT82]]
+; CHECK-NEXT:    [[BLOCK83:%.*]] = shufflevector <4 x double> [[COL_LOAD49]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP40:%.*]] = extractelement <2 x double> [[COL_LOAD46]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT84:%.*]] = insertelement <2 x double> poison, double [[TMP40]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT85:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT84]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP41:%.*]] = fmul <2 x double> [[BLOCK83]], [[SPLAT_SPLAT85]]
+; CHECK-NEXT:    [[TMP42:%.*]] = fadd <2 x double> [[TMP39]], [[TMP41]]
+; CHECK-NEXT:    [[TMP43:%.*]] = shufflevector <2 x double> [[TMP42]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP44:%.*]] = shufflevector <4 x double> [[TMP37]], <4 x double> [[TMP43]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
+; CHECK-NEXT:    [[TMP45:%.*]] = shufflevector <4 x double> [[TMP16]], <4 x double> [[TMP30]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    [[TMP46:%.*]] = shufflevector <4 x double> [[TMP44]], <4 x double> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP47:%.*]] = shufflevector <8 x double> [[TMP45]], <8 x double> [[TMP46]], <12 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>
+; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <6 x double> [[TMP2]], <6 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[SPLIT1:%.*]] = shufflevector <6 x double> [[TMP2]], <6 x double> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[SPLIT2:%.*]] = shufflevector <6 x double> [[TMP2]], <6 x double> poison, <2 x i32> <i32 4, i32 5>
+; CHECK-NEXT:    [[SPLIT3:%.*]] = shufflevector <12 x double> [[TMP47]], <12 x double> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT:    [[SPLIT4:%.*]] = shufflevector <12 x double> [[TMP47]], <12 x double> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    [[SPLIT5:%.*]] = shufflevector <12 x double> [[TMP47]], <12 x double> poison, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
 ; CHECK-NEXT:    [[BLOCK:%.*]] = shufflevector <2 x double> [[SPLIT]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP51:%.*]] = extractelement <4 x double> [[SPLIT3]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP51]], i64 0
+; CHECK-NEXT:    [[TMP48:%.*]] = extractelement <4 x double> [[SPLIT3]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP48]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP52:%.*]] = fmul <2 x double> [[BLOCK]], [[SPLAT_SPLAT]]
+; CHECK-NEXT:    [[TMP49:%.*]] = fmul <2 x double> [[BLOCK]], [[SPLAT_SPLAT]]
 ; CHECK-NEXT:    [[BLOCK6:%.*]] = shufflevector <2 x double> [[SPLIT1]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP53:%.*]] = extractelement <4 x double> [[SPLIT4]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT7:%.*]] = insertelement <2 x double> poison, double [[TMP53]], i64 0
+; CHECK-NEXT:    [[TMP50:%.*]] = extractelement <4 x double> [[SPLIT4]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT7:%.*]] = insertelement <2 x double> poison, double [[TMP50]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT8:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT7]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP54:%.*]] = fmul <2 x double> [[BLOCK6]], [[SPLAT_SPLAT8]]
-; CHECK-NEXT:    [[TMP55:%.*]] = fadd <2 x double> [[TMP52]], [[TMP54]]
+; CHECK-NEXT:    [[TMP51:%.*]] = fmul <2 x double> [[BLOCK6]], [[SPLAT_SPLAT8]]
+; CHECK-NEXT:    [[TMP52:%.*]] = fadd <2 x double> [[TMP49]], [[TMP51]]
 ; CHECK-NEXT:    [[BLOCK9:%.*]] = shufflevector <2 x double> [[SPLIT2]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP56:%.*]] = extractelement <4 x double> [[SPLIT5]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT10:%.*]] = insertelement <2 x double> poison, double [[TMP56]], i64 0
+; CHECK-NEXT:    [[TMP53:%.*]] = extractelement <4 x double> [[SPLIT5]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT10:%.*]] = insertelement <2 x double> poison, double [[TMP53]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT11:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT10]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP57:%.*]] = fmul <2 x double> [[BLOCK9]], [[SPLAT_SPLAT11]]
-; CHECK-NEXT:    [[TMP58:%.*]] = fadd <2 x double> [[TMP55]], [[TMP57]]
-; CHECK-NEXT:    [[TMP59:%.*]] = shufflevector <2 x double> [[TMP58]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP60:%.*]] = shufflevector <2 x double> undef, <2 x double> [[TMP59]], <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP54:%.*]] = fmul <2 x double> [[BLOCK9]], [[SPLAT_SPLAT11]]
+; CHECK-NEXT:    [[TMP55:%.*]] = fadd <2 x double> [[TMP52]], [[TMP54]]
+; CHECK-NEXT:    [[TMP56:%.*]] = shufflevector <2 x double> [[TMP55]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP57:%.*]] = shufflevector <2 x double> poison, <2 x double> [[TMP56]], <2 x i32> <i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK12:%.*]] = shufflevector <2 x double> [[SPLIT]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP61:%.*]] = extractelement <4 x double> [[SPLIT3]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT13:%.*]] = insertelement <2 x double> poison, double [[TMP61]], i64 0
+; CHECK-NEXT:    [[TMP58:%.*]] = extractelement <4 x double> [[SPLIT3]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT13:%.*]] = insertelement <2 x double> poison, double [[TMP58]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT14:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT13]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP62:%.*]] = fmul <2 x double> [[BLOCK12]], [[SPLAT_SPLAT14]]
+; CHECK-NEXT:    [[TMP59:%.*]] = fmul <2 x double> [[BLOCK12]], [[SPLAT_SPLAT14]]
 ; CHECK-NEXT:    [[BLOCK15:%.*]] = shufflevector <2 x double> [[SPLIT1]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP63:%.*]] = extractelement <4 x double> [[SPLIT4]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT16:%.*]] = insertelement <2 x double> poison, double [[TMP63]], i64 0
+; CHECK-NEXT:    [[TMP60:%.*]] = extractelement <4 x double> [[SPLIT4]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT16:%.*]] = insertelement <2 x double> poison, double [[TMP60]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT17:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT16]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP64:%.*]] = fmul <2 x double> [[BLOCK15]], [[SPLAT_SPLAT17]]
-; CHECK-NEXT:    [[TMP65:%.*]] = fadd <2 x double> [[TMP62]], [[TMP64]]
+; CHECK-NEXT:    [[TMP61:%.*]] = fmul <2 x double> [[BLOCK15]], [[SPLAT_SPLAT17]]
+; CHECK-NEXT:    [[TMP62:%.*]] = fadd <2 x double> [[TMP59]], [[TMP61]]
 ; CHECK-NEXT:    [[BLOCK18:%.*]] = shufflevector <2 x double> [[SPLIT2]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP66:%.*]] = extractelement <4 x double> [[SPLIT5]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT19:%.*]] = insertelement <2 x double> poison, double [[TMP66]], i64 0
+; CHECK-NEXT:    [[TMP63:%.*]] = extractelement <4 x double> [[SPLIT5]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT19:%.*]] = insertelement <2 x double> poison, double [[TMP63]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT20:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT19]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP67:%.*]] = fmul <2 x double> [[BLOCK18]], [[SPLAT_SPLAT20]]
-; CHECK-NEXT:    [[TMP68:%.*]] = fadd <2 x double> [[TMP65]], [[TMP67]]
-; CHECK-NEXT:    [[TMP69:%.*]] = shufflevector <2 x double> [[TMP68]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP70:%.*]] = shufflevector <2 x double> undef, <2 x double> [[TMP69]], <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP64:%.*]] = fmul <2 x double> [[BLOCK18]], [[SPLAT_SPLAT20]]
+; CHECK-NEXT:    [[TMP65:%.*]] = fadd <2 x double> [[TMP62]], [[TMP64]]
+; CHECK-NEXT:    [[TMP66:%.*]] = shufflevector <2 x double> [[TMP65]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP67:%.*]] = shufflevector <2 x double> poison, <2 x double> [[TMP66]], <2 x i32> <i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK21:%.*]] = shufflevector <2 x double> [[SPLIT]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP71:%.*]] = extractelement <4 x double> [[SPLIT3]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT22:%.*]] = insertelement <2 x double> poison, double [[TMP71]], i64 0
+; CHECK-NEXT:    [[TMP68:%.*]] = extractelement <4 x double> [[SPLIT3]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT22:%.*]] = insertelement <2 x double> poison, double [[TMP68]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT23:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT22]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP72:%.*]] = fmul <2 x double> [[BLOCK21]], [[SPLAT_SPLAT23]]
+; CHECK-NEXT:    [[TMP69:%.*]] = fmul <2 x double> [[BLOCK21]], [[SPLAT_SPLAT23]]
 ; CHECK-NEXT:    [[BLOCK24:%.*]] = shufflevector <2 x double> [[SPLIT1]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP73:%.*]] = extractelement <4 x double> [[SPLIT4]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT25:%.*]] = insertelement <2 x double> poison, double [[TMP73]], i64 0
+; CHECK-NEXT:    [[TMP70:%.*]] = extractelement <4 x double> [[SPLIT4]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT25:%.*]] = insertelement <2 x double> poison, double [[TMP70]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT26:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT25]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP74:%.*]] = fmul <2 x double> [[BLOCK24]], [[SPLAT_SPLAT26]]
-; CHECK-NEXT:    [[TMP75:%.*]] = fadd <2 x double> [[TMP72]], [[TMP74]]
+; CHECK-NEXT:    [[TMP71:%.*]] = fmul <2 x double> [[BLOCK24]], [[SPLAT_SPLAT26]]
+; CHECK-NEXT:    [[TMP72:%.*]] = fadd <2 x double> [[TMP69]], [[TMP71]]
 ; CHECK-NEXT:    [[BLOCK27:%.*]] = shufflevector <2 x double> [[SPLIT2]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP76:%.*]] = extractelement <4 x double> [[SPLIT5]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT28:%.*]] = insertelement <2 x double> poison, double [[TMP76]], i64 0
+; CHECK-NEXT:    [[TMP73:%.*]] = extractelement <4 x double> [[SPLIT5]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT28:%.*]] = insertelement <2 x double> poison, double [[TMP73]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT29:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT28]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP77:%.*]] = fmul <2 x double> [[BLOCK27]], [[SPLAT_SPLAT29]]
-; CHECK-NEXT:    [[TMP78:%.*]] = fadd <2 x double> [[TMP75]], [[TMP77]]
-; CHECK-NEXT:    [[TMP79:%.*]] = shufflevector <2 x double> [[TMP78]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP80:%.*]] = shufflevector <2 x double> undef, <2 x double> [[TMP79]], <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP74:%.*]] = fmul <2 x double> [[BLOCK27]], [[SPLAT_SPLAT29]]
+; CHECK-NEXT:    [[TMP75:%.*]] = fadd <2 x double> [[TMP72]], [[TMP74]]
+; CHECK-NEXT:    [[TMP76:%.*]] = shufflevector <2 x double> [[TMP75]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP77:%.*]] = shufflevector <2 x double> poison, <2 x double> [[TMP76]], <2 x i32> <i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK30:%.*]] = shufflevector <2 x double> [[SPLIT]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP81:%.*]] = extractelement <4 x double> [[SPLIT3]], i64 3
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT31:%.*]] = insertelement <2 x double> poison, double [[TMP81]], i64 0
+; CHECK-NEXT:    [[TMP78:%.*]] = extractelement <4 x double> [[SPLIT3]], i64 3
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT31:%.*]] = insertelement <2 x double> poison, double [[TMP78]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT32:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT31]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP82:%.*]] = fmul <2 x double> [[BLOCK30]], [[SPLAT_SPLAT32]]
+; CHECK-NEXT:    [[TMP79:%.*]] = fmul <2 x double> [[BLOCK30]], [[SPLAT_SPLAT32]]
 ; CHECK-NEXT:    [[BLOCK33:%.*]] = shufflevector <2 x double> [[SPLIT1]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP83:%.*]] = extractelement <4 x double> [[SPLIT4]], i64 3
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT34:%.*]] = insertelement <2 x double> poison, double [[TMP83]], i64 0
+; CHECK-NEXT:    [[TMP80:%.*]] = extractelement <4 x double> [[SPLIT4]], i64 3
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT34:%.*]] = insertelement <2 x double> poison, double [[TMP80]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT35:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT34]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP84:%.*]] = fmul <2 x double> [[BLOCK33]], [[SPLAT_SPLAT35]]
-; CHECK-NEXT:    [[TMP85:%.*]] = fadd <2 x double> [[TMP82]], [[TMP84]]
+; CHECK-NEXT:    [[TMP81:%.*]] = fmul <2 x double> [[BLOCK33]], [[SPLAT_SPLAT35]]
+; CHECK-NEXT:    [[TMP82:%.*]] = fadd <2 x double> [[TMP79]], [[TMP81]]
 ; CHECK-NEXT:    [[BLOCK36:%.*]] = shufflevector <2 x double> [[SPLIT2]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP86:%.*]] = extractelement <4 x double> [[SPLIT5]], i64 3
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT37:%.*]] = insertelement <2 x double> poison, double [[TMP86]], i64 0
+; CHECK-NEXT:    [[TMP83:%.*]] = extractelement <4 x double> [[SPLIT5]], i64 3
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT37:%.*]] = insertelement <2 x double> poison, double [[TMP83]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT38:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT37]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP87:%.*]] = fmul <2 x double> [[BLOCK36]], [[SPLAT_SPLAT38]]
-; CHECK-NEXT:    [[TMP88:%.*]] = fadd <2 x double> [[TMP85]], [[TMP87]]
-; CHECK-NEXT:    [[TMP89:%.*]] = shufflevector <2 x double> [[TMP88]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP90:%.*]] = shufflevector <2 x double> undef, <2 x double> [[TMP89]], <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    store <2 x double> [[TMP60]], ptr [[R:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP94:%.*]] = getelementptr double, ptr [[R]], i64 2
-; CHECK-NEXT:    store <2 x double> [[TMP70]], ptr [[VEC_GEP94]], align 16
-; CHECK-NEXT:    [[VEC_GEP96:%.*]] = getelementptr double, ptr [[R]], i64 4
-; CHECK-NEXT:    store <2 x double> [[TMP80]], ptr [[VEC_GEP96]], align 16
-; CHECK-NEXT:    [[VEC_GEP98:%.*]] = getelementptr double, ptr [[R]], i64 6
-; CHECK-NEXT:    store <2 x double> [[TMP90]], ptr [[VEC_GEP98]], align 16
+; CHECK-NEXT:    [[TMP84:%.*]] = fmul <2 x double> [[BLOCK36]], [[SPLAT_SPLAT38]]
+; CHECK-NEXT:    [[TMP85:%.*]] = fadd <2 x double> [[TMP82]], [[TMP84]]
+; CHECK-NEXT:    [[TMP86:%.*]] = shufflevector <2 x double> [[TMP85]], <2 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP87:%.*]] = shufflevector <2 x double> poison, <2 x double> [[TMP86]], <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    store <2 x double> [[TMP57]], ptr [[R:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP86:%.*]] = getelementptr double, ptr [[R]], i64 2
+; CHECK-NEXT:    store <2 x double> [[TMP67]], ptr [[VEC_GEP86]], align 16
+; CHECK-NEXT:    [[VEC_GEP87:%.*]] = getelementptr double, ptr [[R]], i64 4
+; CHECK-NEXT:    store <2 x double> [[TMP77]], ptr [[VEC_GEP87]], align 16
+; CHECK-NEXT:    [[VEC_GEP88:%.*]] = getelementptr double, ptr [[R]], i64 6
+; CHECK-NEXT:    store <2 x double> [[TMP87]], ptr [[VEC_GEP88]], align 16
 ; CHECK-NEXT:    ret void
 ;
 entry:
@@ -297,133 +297,133 @@ define void @multiply_tt_t(ptr %A, ptr %B, ptr %C) {
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <3 x double>, ptr [[A:%.*]], align 16
 ; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr double, ptr [[A]], i64 3
-; CHECK-NEXT:    [[COL_LOAD2:%.*]] = load <3 x double>, ptr [[VEC_GEP]], align 8
-; CHECK-NEXT:    [[VEC_GEP3:%.*]] = getelementptr double, ptr [[A]], i64 6
-; CHECK-NEXT:    [[COL_LOAD5:%.*]] = load <3 x double>, ptr [[VEC_GEP3]], align 16
-; CHECK-NEXT:    [[COL_LOAD7:%.*]] = load <3 x double>, ptr [[B:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP8:%.*]] = getelementptr double, ptr [[B]], i64 3
-; CHECK-NEXT:    [[COL_LOAD10:%.*]] = load <3 x double>, ptr [[VEC_GEP8]], align 8
-; CHECK-NEXT:    [[VEC_GEP11:%.*]] = getelementptr double, ptr [[B]], i64 6
-; CHECK-NEXT:    [[COL_LOAD13:%.*]] = load <3 x double>, ptr [[VEC_GEP11]], align 16
-; CHECK-NEXT:    [[BLOCK:%.*]] = shufflevector <3 x double> [[COL_LOAD7]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP2]], i64 0
+; CHECK-NEXT:    [[COL_LOAD1:%.*]] = load <3 x double>, ptr [[VEC_GEP]], align 8
+; CHECK-NEXT:    [[VEC_GEP2:%.*]] = getelementptr double, ptr [[A]], i64 6
+; CHECK-NEXT:    [[COL_LOAD3:%.*]] = load <3 x double>, ptr [[VEC_GEP2]], align 16
+; CHECK-NEXT:    [[COL_LOAD4:%.*]] = load <3 x double>, ptr [[B:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP5:%.*]] = getelementptr double, ptr [[B]], i64 3
+; CHECK-NEXT:    [[COL_LOAD6:%.*]] = load <3 x double>, ptr [[VEC_GEP5]], align 8
+; CHECK-NEXT:    [[VEC_GEP7:%.*]] = getelementptr double, ptr [[B]], i64 6
+; CHECK-NEXT:    [[COL_LOAD8:%.*]] = load <3 x double>, ptr [[VEC_GEP7]], align 16
+; CHECK-NEXT:    [[BLOCK:%.*]] = shufflevector <3 x double> [[COL_LOAD4]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP0]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP3:%.*]] = fmul <2 x double> [[BLOCK]], [[SPLAT_SPLAT]]
-; CHECK-NEXT:    [[BLOCK14:%.*]] = shufflevector <3 x double> [[COL_LOAD10]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT15:%.*]] = insertelement <2 x double> poison, double [[TMP4]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT16:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT15]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP5:%.*]] = fmul <2 x double> [[BLOCK14]], [[SPLAT_SPLAT16]]
-; CHECK-NEXT:    [[TMP6:%.*]] = fadd <2 x double> [[TMP3]], [[TMP5]]
-; CHECK-NEXT:    [[BLOCK17:%.*]] = shufflevector <3 x double> [[COL_LOAD13]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT18:%.*]] = insertelement <2 x double> poison, double [[TMP7]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT19:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT18]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP8:%.*]] = fmul <2 x double> [[BLOCK17]], [[SPLAT_SPLAT19]]
-; CHECK-NEXT:    [[TMP9:%.*]] = fadd <2 x double> [[TMP6]], [[TMP8]]
-; CHECK-NEXT:    [[TMP10:%.*]] = shufflevector <2 x double> [[TMP9]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP11:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP10]], <3 x i32> <i32 3, i32 4, i32 2>
-; CHECK-NEXT:    [[BLOCK20:%.*]] = shufflevector <3 x double> [[COL_LOAD7]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP12:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT21:%.*]] = insertelement <1 x double> poison, double [[TMP12]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT22:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT21]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP13:%.*]] = fmul <1 x double> [[BLOCK20]], [[SPLAT_SPLAT22]]
-; CHECK-NEXT:    [[BLOCK23:%.*]] = shufflevector <3 x double> [[COL_LOAD10]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP14:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT24:%.*]] = insertelement <1 x double> poison, double [[TMP14]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT25:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT24]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP15:%.*]] = fmul <1 x double> [[BLOCK23]], [[SPLAT_SPLAT25]]
-; CHECK-NEXT:    [[TMP16:%.*]] = fadd <1 x double> [[TMP13]], [[TMP15]]
-; CHECK-NEXT:    [[BLOCK26:%.*]] = shufflevector <3 x double> [[COL_LOAD13]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP17:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT27:%.*]] = insertelement <1 x double> poison, double [[TMP17]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT28:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT27]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP18:%.*]] = fmul <1 x double> [[BLOCK26]], [[SPLAT_SPLAT28]]
-; CHECK-NEXT:    [[TMP19:%.*]] = fadd <1 x double> [[TMP16]], [[TMP18]]
-; CHECK-NEXT:    [[TMP20:%.*]] = shufflevector <1 x double> [[TMP19]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP21:%.*]] = shufflevector <3 x double> [[TMP11]], <3 x double> [[TMP20]], <3 x i32> <i32 0, i32 1, i32 3>
-; CHECK-NEXT:    [[BLOCK29:%.*]] = shufflevector <3 x double> [[COL_LOAD7]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP22:%.*]] = extractelement <3 x double> [[COL_LOAD2]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT30:%.*]] = insertelement <2 x double> poison, double [[TMP22]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT31:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT30]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP23:%.*]] = fmul <2 x double> [[BLOCK29]], [[SPLAT_SPLAT31]]
-; CHECK-NEXT:    [[BLOCK32:%.*]] = shufflevector <3 x double> [[COL_LOAD10]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP24:%.*]] = extractelement <3 x double> [[COL_LOAD2]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT33:%.*]] = insertelement <2 x double> poison, double [[TMP24]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT34:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT33]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP25:%.*]] = fmul <2 x double> [[BLOCK32]], [[SPLAT_SPLAT34]]
-; CHECK-NEXT:    [[TMP26:%.*]] = fadd <2 x double> [[TMP23]], [[TMP25]]
-; CHECK-NEXT:    [[BLOCK35:%.*]] = shufflevector <3 x double> [[COL_LOAD13]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP27:%.*]] = extractelement <3 x double> [[COL_LOAD2]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT36:%.*]] = insertelement <2 x double> poison, double [[TMP27]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT37:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT36]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP28:%.*]] = fmul <2 x double> [[BLOCK35]], [[SPLAT_SPLAT37]]
-; CHECK-NEXT:    [[TMP29:%.*]] = fadd <2 x double> [[TMP26]], [[TMP28]]
-; CHECK-NEXT:    [[TMP30:%.*]] = shufflevector <2 x double> [[TMP29]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP31:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP30]], <3 x i32> <i32 3, i32 4, i32 2>
-; CHECK-NEXT:    [[BLOCK38:%.*]] = shufflevector <3 x double> [[COL_LOAD7]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP32:%.*]] = extractelement <3 x double> [[COL_LOAD2]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT39:%.*]] = insertelement <1 x double> poison, double [[TMP32]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT40:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT39]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP33:%.*]] = fmul <1 x double> [[BLOCK38]], [[SPLAT_SPLAT40]]
-; CHECK-NEXT:    [[BLOCK41:%.*]] = shufflevector <3 x double> [[COL_LOAD10]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP34:%.*]] = extractelement <3 x double> [[COL_LOAD2]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT42:%.*]] = insertelement <1 x double> poison, double [[TMP34]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT43:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT42]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP35:%.*]] = fmul <1 x double> [[BLOCK41]], [[SPLAT_SPLAT43]]
-; CHECK-NEXT:    [[TMP36:%.*]] = fadd <1 x double> [[TMP33]], [[TMP35]]
-; CHECK-NEXT:    [[BLOCK44:%.*]] = shufflevector <3 x double> [[COL_LOAD13]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP37:%.*]] = extractelement <3 x double> [[COL_LOAD2]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT45:%.*]] = insertelement <1 x double> poison, double [[TMP37]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT46:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT45]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP38:%.*]] = fmul <1 x double> [[BLOCK44]], [[SPLAT_SPLAT46]]
-; CHECK-NEXT:    [[TMP39:%.*]] = fadd <1 x double> [[TMP36]], [[TMP38]]
-; CHECK-NEXT:    [[TMP40:%.*]] = shufflevector <1 x double> [[TMP39]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP41:%.*]] = shufflevector <3 x double> [[TMP31]], <3 x double> [[TMP40]], <3 x i32> <i32 0, i32 1, i32 3>
-; CHECK-NEXT:    [[BLOCK47:%.*]] = shufflevector <3 x double> [[COL_LOAD7]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP42:%.*]] = extractelement <3 x double> [[COL_LOAD5]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT48:%.*]] = insertelement <2 x double> poison, double [[TMP42]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT49:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT48]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP43:%.*]] = fmul <2 x double> [[BLOCK47]], [[SPLAT_SPLAT49]]
-; CHECK-NEXT:    [[BLOCK50:%.*]] = shufflevector <3 x double> [[COL_LOAD10]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP44:%.*]] = extractelement <3 x double> [[COL_LOAD5]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT51:%.*]] = insertelement <2 x double> poison, double [[TMP44]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT52:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT51]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP45:%.*]] = fmul <2 x double> [[BLOCK50]], [[SPLAT_SPLAT52]]
-; CHECK-NEXT:    [[TMP46:%.*]] = fadd <2 x double> [[TMP43]], [[TMP45]]
-; CHECK-NEXT:    [[BLOCK53:%.*]] = shufflevector <3 x double> [[COL_LOAD13]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP47:%.*]] = extractelement <3 x double> [[COL_LOAD5]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT54:%.*]] = insertelement <2 x double> poison, double [[TMP47]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT55:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT54]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP48:%.*]] = fmul <2 x double> [[BLOCK53]], [[SPLAT_SPLAT55]]
-; CHECK-NEXT:    [[TMP49:%.*]] = fadd <2 x double> [[TMP46]], [[TMP48]]
-; CHECK-NEXT:    [[TMP50:%.*]] = shufflevector <2 x double> [[TMP49]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP51:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP50]], <3 x i32> <i32 3, i32 4, i32 2>
-; CHECK-NEXT:    [[BLOCK56:%.*]] = shufflevector <3 x double> [[COL_LOAD7]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP52:%.*]] = extractelement <3 x double> [[COL_LOAD5]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT57:%.*]] = insertelement <1 x double> poison, double [[TMP52]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT58:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT57]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP53:%.*]] = fmul <1 x double> [[BLOCK56]], [[SPLAT_SPLAT58]]
-; CHECK-NEXT:    [[BLOCK59:%.*]] = shufflevector <3 x double> [[COL_LOAD10]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP54:%.*]] = extractelement <3 x double> [[COL_LOAD5]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT60:%.*]] = insertelement <1 x double> poison, double [[TMP54]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT61:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT60]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP55:%.*]] = fmul <1 x double> [[BLOCK59]], [[SPLAT_SPLAT61]]
-; CHECK-NEXT:    [[TMP56:%.*]] = fadd <1 x double> [[TMP53]], [[TMP55]]
-; CHECK-NEXT:    [[BLOCK62:%.*]] = shufflevector <3 x double> [[COL_LOAD13]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP57:%.*]] = extractelement <3 x double> [[COL_LOAD5]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT63:%.*]] = insertelement <1 x double> poison, double [[TMP57]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT64:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT63]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP58:%.*]] = fmul <1 x double> [[BLOCK62]], [[SPLAT_SPLAT64]]
-; CHECK-NEXT:    [[TMP59:%.*]] = fadd <1 x double> [[TMP56]], [[TMP58]]
-; CHECK-NEXT:    [[TMP60:%.*]] = shufflevector <1 x double> [[TMP59]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP61:%.*]] = shufflevector <3 x double> [[TMP51]], <3 x double> [[TMP60]], <3 x i32> <i32 0, i32 1, i32 3>
-; CHECK-NEXT:    store <3 x double> [[TMP21]], ptr [[C:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP66:%.*]] = getelementptr double, ptr [[C]], i64 3
-; CHECK-NEXT:    store <3 x double> [[TMP41]], ptr [[VEC_GEP66]], align 8
-; CHECK-NEXT:    [[VEC_GEP68:%.*]] = getelementptr double, ptr [[C]], i64 6
-; CHECK-NEXT:    store <3 x double> [[TMP61]], ptr [[VEC_GEP68]], align 16
+; CHECK-NEXT:    [[TMP1:%.*]] = fmul <2 x double> [[BLOCK]], [[SPLAT_SPLAT]]
+; CHECK-NEXT:    [[BLOCK9:%.*]] = shufflevector <3 x double> [[COL_LOAD6]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT10:%.*]] = insertelement <2 x double> poison, double [[TMP2]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT11:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT10]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP3:%.*]] = fmul <2 x double> [[BLOCK9]], [[SPLAT_SPLAT11]]
+; CHECK-NEXT:    [[TMP4:%.*]] = fadd <2 x double> [[TMP1]], [[TMP3]]
+; CHECK-NEXT:    [[BLOCK12:%.*]] = shufflevector <3 x double> [[COL_LOAD8]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT13:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT14:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT13]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP6:%.*]] = fmul <2 x double> [[BLOCK12]], [[SPLAT_SPLAT14]]
+; CHECK-NEXT:    [[TMP7:%.*]] = fadd <2 x double> [[TMP4]], [[TMP6]]
+; CHECK-NEXT:    [[TMP8:%.*]] = shufflevector <2 x double> [[TMP7]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP9:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP8]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[BLOCK15:%.*]] = shufflevector <3 x double> [[COL_LOAD4]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP10:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT16:%.*]] = insertelement <1 x double> poison, double [[TMP10]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT17:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT16]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP11:%.*]] = fmul <1 x double> [[BLOCK15]], [[SPLAT_SPLAT17]]
+; CHECK-NEXT:    [[BLOCK18:%.*]] = shufflevector <3 x double> [[COL_LOAD6]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP12:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT19:%.*]] = insertelement <1 x double> poison, double [[TMP12]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT20:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT19]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP13:%.*]] = fmul <1 x double> [[BLOCK18]], [[SPLAT_SPLAT20]]
+; CHECK-NEXT:    [[TMP14:%.*]] = fadd <1 x double> [[TMP11]], [[TMP13]]
+; CHECK-NEXT:    [[BLOCK21:%.*]] = shufflevector <3 x double> [[COL_LOAD8]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP15:%.*]] = extractelement <3 x double> [[COL_LOAD]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT22:%.*]] = insertelement <1 x double> poison, double [[TMP15]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT23:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT22]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP16:%.*]] = fmul <1 x double> [[BLOCK21]], [[SPLAT_SPLAT23]]
+; CHECK-NEXT:    [[TMP17:%.*]] = fadd <1 x double> [[TMP14]], [[TMP16]]
+; CHECK-NEXT:    [[TMP18:%.*]] = shufflevector <1 x double> [[TMP17]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP19:%.*]] = shufflevector <3 x double> [[TMP9]], <3 x double> [[TMP18]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    [[BLOCK24:%.*]] = shufflevector <3 x double> [[COL_LOAD4]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP20:%.*]] = extractelement <3 x double> [[COL_LOAD1]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT25:%.*]] = insertelement <2 x double> poison, double [[TMP20]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT26:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT25]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP21:%.*]] = fmul <2 x double> [[BLOCK24]], [[SPLAT_SPLAT26]]
+; CHECK-NEXT:    [[BLOCK27:%.*]] = shufflevector <3 x double> [[COL_LOAD6]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP22:%.*]] = extractelement <3 x double> [[COL_LOAD1]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT28:%.*]] = insertelement <2 x double> poison, double [[TMP22]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT29:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT28]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP23:%.*]] = fmul <2 x double> [[BLOCK27]], [[SPLAT_SPLAT29]]
+; CHECK-NEXT:    [[TMP24:%.*]] = fadd <2 x double> [[TMP21]], [[TMP23]]
+; CHECK-NEXT:    [[BLOCK30:%.*]] = shufflevector <3 x double> [[COL_LOAD8]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP25:%.*]] = extractelement <3 x double> [[COL_LOAD1]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT31:%.*]] = insertelement <2 x double> poison, double [[TMP25]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT32:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT31]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP26:%.*]] = fmul <2 x double> [[BLOCK30]], [[SPLAT_SPLAT32]]
+; CHECK-NEXT:    [[TMP27:%.*]] = fadd <2 x double> [[TMP24]], [[TMP26]]
+; CHECK-NEXT:    [[TMP28:%.*]] = shufflevector <2 x double> [[TMP27]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP29:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP28]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[BLOCK33:%.*]] = shufflevector <3 x double> [[COL_LOAD4]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP30:%.*]] = extractelement <3 x double> [[COL_LOAD1]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT34:%.*]] = insertelement <1 x double> poison, double [[TMP30]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT35:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT34]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP31:%.*]] = fmul <1 x double> [[BLOCK33]], [[SPLAT_SPLAT35]]
+; CHECK-NEXT:    [[BLOCK36:%.*]] = shufflevector <3 x double> [[COL_LOAD6]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP32:%.*]] = extractelement <3 x double> [[COL_LOAD1]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT37:%.*]] = insertelement <1 x double> poison, double [[TMP32]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT38:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT37]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP33:%.*]] = fmul <1 x double> [[BLOCK36]], [[SPLAT_SPLAT38]]
+; CHECK-NEXT:    [[TMP34:%.*]] = fadd <1 x double> [[TMP31]], [[TMP33]]
+; CHECK-NEXT:    [[BLOCK39:%.*]] = shufflevector <3 x double> [[COL_LOAD8]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP35:%.*]] = extractelement <3 x double> [[COL_LOAD1]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT40:%.*]] = insertelement <1 x double> poison, double [[TMP35]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT41:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT40]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP36:%.*]] = fmul <1 x double> [[BLOCK39]], [[SPLAT_SPLAT41]]
+; CHECK-NEXT:    [[TMP37:%.*]] = fadd <1 x double> [[TMP34]], [[TMP36]]
+; CHECK-NEXT:    [[TMP38:%.*]] = shufflevector <1 x double> [[TMP37]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP39:%.*]] = shufflevector <3 x double> [[TMP29]], <3 x double> [[TMP38]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    [[BLOCK42:%.*]] = shufflevector <3 x double> [[COL_LOAD4]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP40:%.*]] = extractelement <3 x double> [[COL_LOAD3]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT43:%.*]] = insertelement <2 x double> poison, double [[TMP40]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT44:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT43]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP41:%.*]] = fmul <2 x double> [[BLOCK42]], [[SPLAT_SPLAT44]]
+; CHECK-NEXT:    [[BLOCK45:%.*]] = shufflevector <3 x double> [[COL_LOAD6]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP42:%.*]] = extractelement <3 x double> [[COL_LOAD3]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT46:%.*]] = insertelement <2 x double> poison, double [[TMP42]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT47:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT46]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP43:%.*]] = fmul <2 x double> [[BLOCK45]], [[SPLAT_SPLAT47]]
+; CHECK-NEXT:    [[TMP44:%.*]] = fadd <2 x double> [[TMP41]], [[TMP43]]
+; CHECK-NEXT:    [[BLOCK48:%.*]] = shufflevector <3 x double> [[COL_LOAD8]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP45:%.*]] = extractelement <3 x double> [[COL_LOAD3]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT49:%.*]] = insertelement <2 x double> poison, double [[TMP45]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT50:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT49]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP46:%.*]] = fmul <2 x double> [[BLOCK48]], [[SPLAT_SPLAT50]]
+; CHECK-NEXT:    [[TMP47:%.*]] = fadd <2 x double> [[TMP44]], [[TMP46]]
+; CHECK-NEXT:    [[TMP48:%.*]] = shufflevector <2 x double> [[TMP47]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP49:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP48]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[BLOCK51:%.*]] = shufflevector <3 x double> [[COL_LOAD4]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP50:%.*]] = extractelement <3 x double> [[COL_LOAD3]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT52:%.*]] = insertelement <1 x double> poison, double [[TMP50]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT53:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT52]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP51:%.*]] = fmul <1 x double> [[BLOCK51]], [[SPLAT_SPLAT53]]
+; CHECK-NEXT:    [[BLOCK54:%.*]] = shufflevector <3 x double> [[COL_LOAD6]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP52:%.*]] = extractelement <3 x double> [[COL_LOAD3]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT55:%.*]] = insertelement <1 x double> poison, double [[TMP52]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT56:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT55]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP53:%.*]] = fmul <1 x double> [[BLOCK54]], [[SPLAT_SPLAT56]]
+; CHECK-NEXT:    [[TMP54:%.*]] = fadd <1 x double> [[TMP51]], [[TMP53]]
+; CHECK-NEXT:    [[BLOCK57:%.*]] = shufflevector <3 x double> [[COL_LOAD8]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP55:%.*]] = extractelement <3 x double> [[COL_LOAD3]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT58:%.*]] = insertelement <1 x double> poison, double [[TMP55]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT59:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT58]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP56:%.*]] = fmul <1 x double> [[BLOCK57]], [[SPLAT_SPLAT59]]
+; CHECK-NEXT:    [[TMP57:%.*]] = fadd <1 x double> [[TMP54]], [[TMP56]]
+; CHECK-NEXT:    [[TMP58:%.*]] = shufflevector <1 x double> [[TMP57]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP59:%.*]] = shufflevector <3 x double> [[TMP49]], <3 x double> [[TMP58]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    store <3 x double> [[TMP19]], ptr [[C:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP60:%.*]] = getelementptr double, ptr [[C]], i64 3
+; CHECK-NEXT:    store <3 x double> [[TMP39]], ptr [[VEC_GEP60]], align 8
+; CHECK-NEXT:    [[VEC_GEP61:%.*]] = getelementptr double, ptr [[C]], i64 6
+; CHECK-NEXT:    store <3 x double> [[TMP59]], ptr [[VEC_GEP61]], align 16
 ; CHECK-NEXT:    ret void
 ;
 entry:
@@ -461,105 +461,105 @@ define void @multiply_nt_t(ptr %A, ptr %B, ptr %C) {
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <2 x double>, ptr [[A:%.*]], align 16
 ; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr double, ptr [[A]], i64 2
-; CHECK-NEXT:    [[COL_LOAD40:%.*]] = load <2 x double>, ptr [[VEC_GEP]], align 16
-; CHECK-NEXT:    [[VEC_GEP41:%.*]] = getelementptr double, ptr [[A]], i64 4
-; CHECK-NEXT:    [[COL_LOAD43:%.*]] = load <2 x double>, ptr [[VEC_GEP41]], align 16
-; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <2 x double> [[COL_LOAD]], <2 x double> [[COL_LOAD40]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <2 x double> [[COL_LOAD43]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x double> [[TMP1]], <4 x double> [[TMP2]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
-; CHECK-NEXT:    [[COL_LOAD45:%.*]] = load <4 x double>, ptr [[B:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP46:%.*]] = getelementptr double, ptr [[B]], i64 4
-; CHECK-NEXT:    [[COL_LOAD48:%.*]] = load <4 x double>, ptr [[VEC_GEP46]], align 16
-; CHECK-NEXT:    [[VEC_GEP49:%.*]] = getelementptr double, ptr [[B]], i64 8
-; CHECK-NEXT:    [[COL_LOAD51:%.*]] = load <4 x double>, ptr [[VEC_GEP49]], align 16
-; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <4 x double> [[COL_LOAD45]], <4 x double> [[COL_LOAD48]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[TMP6:%.*]] = shufflevector <4 x double> [[COL_LOAD51]], <4 x double> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP7:%.*]] = shufflevector <8 x double> [[TMP5]], <8 x double> [[TMP6]], <12 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>
-; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <12 x double> [[TMP7]], <12 x double> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[SPLIT1:%.*]] = shufflevector <12 x double> [[TMP7]], <12 x double> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
-; CHECK-NEXT:    [[SPLIT2:%.*]] = shufflevector <12 x double> [[TMP7]], <12 x double> poison, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
-; CHECK-NEXT:    [[SPLIT3:%.*]] = shufflevector <6 x double> [[TMP3]], <6 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[SPLIT4:%.*]] = shufflevector <6 x double> [[TMP3]], <6 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[SPLIT5:%.*]] = shufflevector <6 x double> [[TMP3]], <6 x double> poison, <2 x i32> <i32 4, i32 5>
+; CHECK-NEXT:    [[COL_LOAD39:%.*]] = load <2 x double>, ptr [[VEC_GEP]], align 16
+; CHECK-NEXT:    [[VEC_GEP40:%.*]] = getelementptr double, ptr [[A]], i64 4
+; CHECK-NEXT:    [[COL_LOAD41:%.*]] = load <2 x double>, ptr [[VEC_GEP40]], align 16
+; CHECK-NEXT:    [[TMP0:%.*]] = shufflevector <2 x double> [[COL_LOAD]], <2 x double> [[COL_LOAD39]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <2 x double> [[COL_LOAD41]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[TMP0]], <4 x double> [[TMP1]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
+; CHECK-NEXT:    [[COL_LOAD42:%.*]] = load <4 x double>, ptr [[B:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP43:%.*]] = getelementptr double, ptr [[B]], i64 4
+; CHECK-NEXT:    [[COL_LOAD44:%.*]] = load <4 x double>, ptr [[VEC_GEP43]], align 16
+; CHECK-NEXT:    [[VEC_GEP45:%.*]] = getelementptr double, ptr [[B]], i64 8
+; CHECK-NEXT:    [[COL_LOAD46:%.*]] = load <4 x double>, ptr [[VEC_GEP45]], align 16
+; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x double> [[COL_LOAD42]], <4 x double> [[COL_LOAD44]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    [[TMP4:%.*]] = shufflevector <4 x double> [[COL_LOAD46]], <4 x double> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 poison, i32 poison, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <8 x double> [[TMP3]], <8 x double> [[TMP4]], <12 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>
+; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <12 x double> [[TMP5]], <12 x double> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT:    [[SPLIT1:%.*]] = shufflevector <12 x double> [[TMP5]], <12 x double> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT:    [[SPLIT2:%.*]] = shufflevector <12 x double> [[TMP5]], <12 x double> poison, <4 x i32> <i32 8, i32 9, i32 10, i32 11>
+; CHECK-NEXT:    [[SPLIT3:%.*]] = shufflevector <6 x double> [[TMP2]], <6 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[SPLIT4:%.*]] = shufflevector <6 x double> [[TMP2]], <6 x double> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[SPLIT5:%.*]] = shufflevector <6 x double> [[TMP2]], <6 x double> poison, <2 x i32> <i32 4, i32 5>
 ; CHECK-NEXT:    [[BLOCK:%.*]] = shufflevector <4 x double> [[SPLIT]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP8:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP8]], i64 0
+; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP6]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP9:%.*]] = fmul <2 x double> [[BLOCK]], [[SPLAT_SPLAT]]
+; CHECK-NEXT:    [[TMP7:%.*]] = fmul <2 x double> [[BLOCK]], [[SPLAT_SPLAT]]
 ; CHECK-NEXT:    [[BLOCK6:%.*]] = shufflevector <4 x double> [[SPLIT1]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP10:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT7:%.*]] = insertelement <2 x double> poison, double [[TMP10]], i64 0
+; CHECK-NEXT:    [[TMP8:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT7:%.*]] = insertelement <2 x double> poison, double [[TMP8]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT8:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT7]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP11:%.*]] = fmul <2 x double> [[BLOCK6]], [[SPLAT_SPLAT8]]
-; CHECK-NEXT:    [[TMP12:%.*]] = fadd <2 x double> [[TMP9]], [[TMP11]]
+; CHECK-NEXT:    [[TMP9:%.*]] = fmul <2 x double> [[BLOCK6]], [[SPLAT_SPLAT8]]
+; CHECK-NEXT:    [[TMP10:%.*]] = fadd <2 x double> [[TMP7]], [[TMP9]]
 ; CHECK-NEXT:    [[BLOCK9:%.*]] = shufflevector <4 x double> [[SPLIT2]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP13:%.*]] = extractelement <2 x double> [[SPLIT5]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT10:%.*]] = insertelement <2 x double> poison, double [[TMP13]], i64 0
+; CHECK-NEXT:    [[TMP11:%.*]] = extractelement <2 x double> [[SPLIT5]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT10:%.*]] = insertelement <2 x double> poison, double [[TMP11]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT11:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT10]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP14:%.*]] = fmul <2 x double> [[BLOCK9]], [[SPLAT_SPLAT11]]
-; CHECK-NEXT:    [[TMP15:%.*]] = fadd <2 x double> [[TMP12]], [[TMP14]]
-; CHECK-NEXT:    [[TMP16:%.*]] = shufflevector <2 x double> [[TMP15]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP17:%.*]] = shufflevector <4 x double> undef, <4 x double> [[TMP16]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP12:%.*]] = fmul <2 x double> [[BLOCK9]], [[SPLAT_SPLAT11]]
+; CHECK-NEXT:    [[TMP13:%.*]] = fadd <2 x double> [[TMP10]], [[TMP12]]
+; CHECK-NEXT:    [[TMP14:%.*]] = shufflevector <2 x double> [[TMP13]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP15:%.*]] = shufflevector <4 x double> poison, <4 x double> [[TMP14]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK12:%.*]] = shufflevector <4 x double> [[SPLIT]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP18:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT13:%.*]] = insertelement <2 x double> poison, double [[TMP18]], i64 0
+; CHECK-NEXT:    [[TMP16:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT13:%.*]] = insertelement <2 x double> poison, double [[TMP16]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT14:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT13]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP19:%.*]] = fmul <2 x double> [[BLOCK12]], [[SPLAT_SPLAT14]]
+; CHECK-NEXT:    [[TMP17:%.*]] = fmul <2 x double> [[BLOCK12]], [[SPLAT_SPLAT14]]
 ; CHECK-NEXT:    [[BLOCK15:%.*]] = shufflevector <4 x double> [[SPLIT1]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP20:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT16:%.*]] = insertelement <2 x double> poison, double [[TMP20]], i64 0
+; CHECK-NEXT:    [[TMP18:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT16:%.*]] = insertelement <2 x double> poison, double [[TMP18]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT17:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT16]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP21:%.*]] = fmul <2 x double> [[BLOCK15]], [[SPLAT_SPLAT17]]
-; CHECK-NEXT:    [[TMP22:%.*]] = fadd <2 x double> [[TMP19]], [[TMP21]]
+; CHECK-NEXT:    [[TMP19:%.*]] = fmul <2 x double> [[BLOCK15]], [[SPLAT_SPLAT17]]
+; CHECK-NEXT:    [[TMP20:%.*]] = fadd <2 x double> [[TMP17]], [[TMP19]]
 ; CHECK-NEXT:    [[BLOCK18:%.*]] = shufflevector <4 x double> [[SPLIT2]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP23:%.*]] = extractelement <2 x double> [[SPLIT5]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT19:%.*]] = insertelement <2 x double> poison, double [[TMP23]], i64 0
+; CHECK-NEXT:    [[TMP21:%.*]] = extractelement <2 x double> [[SPLIT5]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT19:%.*]] = insertelement <2 x double> poison, double [[TMP21]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT20:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT19]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP24:%.*]] = fmul <2 x double> [[BLOCK18]], [[SPLAT_SPLAT20]]
-; CHECK-NEXT:    [[TMP25:%.*]] = fadd <2 x double> [[TMP22]], [[TMP24]]
-; CHECK-NEXT:    [[TMP26:%.*]] = shufflevector <2 x double> [[TMP25]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP27:%.*]] = shufflevector <4 x double> [[TMP17]], <4 x double> [[TMP26]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
+; CHECK-NEXT:    [[TMP22:%.*]] = fmul <2 x double> [[BLOCK18]], [[SPLAT_SPLAT20]]
+; CHECK-NEXT:    [[TMP23:%.*]] = fadd <2 x double> [[TMP20]], [[TMP22]]
+; CHECK-NEXT:    [[TMP24:%.*]] = shufflevector <2 x double> [[TMP23]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP25:%.*]] = shufflevector <4 x double> [[TMP15]], <4 x double> [[TMP24]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
 ; CHECK-NEXT:    [[BLOCK21:%.*]] = shufflevector <4 x double> [[SPLIT]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP28:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT22:%.*]] = insertelement <2 x double> poison, double [[TMP28]], i64 0
+; CHECK-NEXT:    [[TMP26:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT22:%.*]] = insertelement <2 x double> poison, double [[TMP26]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT23:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT22]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP29:%.*]] = fmul <2 x double> [[BLOCK21]], [[SPLAT_SPLAT23]]
+; CHECK-NEXT:    [[TMP27:%.*]] = fmul <2 x double> [[BLOCK21]], [[SPLAT_SPLAT23]]
 ; CHECK-NEXT:    [[BLOCK24:%.*]] = shufflevector <4 x double> [[SPLIT1]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP30:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT25:%.*]] = insertelement <2 x double> poison, double [[TMP30]], i64 0
+; CHECK-NEXT:    [[TMP28:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT25:%.*]] = insertelement <2 x double> poison, double [[TMP28]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT26:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT25]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP31:%.*]] = fmul <2 x double> [[BLOCK24]], [[SPLAT_SPLAT26]]
-; CHECK-NEXT:    [[TMP32:%.*]] = fadd <2 x double> [[TMP29]], [[TMP31]]
+; CHECK-NEXT:    [[TMP29:%.*]] = fmul <2 x double> [[BLOCK24]], [[SPLAT_SPLAT26]]
+; CHECK-NEXT:    [[TMP30:%.*]] = fadd <2 x double> [[TMP27]], [[TMP29]]
 ; CHECK-NEXT:    [[BLOCK27:%.*]] = shufflevector <4 x double> [[SPLIT2]], <4 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP33:%.*]] = extractelement <2 x double> [[SPLIT5]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT28:%.*]] = insertelement <2 x double> poison, double [[TMP33]], i64 0
+; CHECK-NEXT:    [[TMP31:%.*]] = extractelement <2 x double> [[SPLIT5]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT28:%.*]] = insertelement <2 x double> poison, double [[TMP31]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT29:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT28]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP34:%.*]] = fmul <2 x double> [[BLOCK27]], [[SPLAT_SPLAT29]]
-; CHECK-NEXT:    [[TMP35:%.*]] = fadd <2 x double> [[TMP32]], [[TMP34]]
-; CHECK-NEXT:    [[TMP36:%.*]] = shufflevector <2 x double> [[TMP35]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP37:%.*]] = shufflevector <4 x double> undef, <4 x double> [[TMP36]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP32:%.*]] = fmul <2 x double> [[BLOCK27]], [[SPLAT_SPLAT29]]
+; CHECK-NEXT:    [[TMP33:%.*]] = fadd <2 x double> [[TMP30]], [[TMP32]]
+; CHECK-NEXT:    [[TMP34:%.*]] = shufflevector <2 x double> [[TMP33]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP35:%.*]] = shufflevector <4 x double> poison, <4 x double> [[TMP34]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK30:%.*]] = shufflevector <4 x double> [[SPLIT]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP38:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT31:%.*]] = insertelement <2 x double> poison, double [[TMP38]], i64 0
+; CHECK-NEXT:    [[TMP36:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT31:%.*]] = insertelement <2 x double> poison, double [[TMP36]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT32:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT31]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP39:%.*]] = fmul <2 x double> [[BLOCK30]], [[SPLAT_SPLAT32]]
+; CHECK-NEXT:    [[TMP37:%.*]] = fmul <2 x double> [[BLOCK30]], [[SPLAT_SPLAT32]]
 ; CHECK-NEXT:    [[BLOCK33:%.*]] = shufflevector <4 x double> [[SPLIT1]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP40:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT34:%.*]] = insertelement <2 x double> poison, double [[TMP40]], i64 0
+; CHECK-NEXT:    [[TMP38:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT34:%.*]] = insertelement <2 x double> poison, double [[TMP38]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT35:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT34]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP41:%.*]] = fmul <2 x double> [[BLOCK33]], [[SPLAT_SPLAT35]]
-; CHECK-NEXT:    [[TMP42:%.*]] = fadd <2 x double> [[TMP39]], [[TMP41]]
+; CHECK-NEXT:    [[TMP39:%.*]] = fmul <2 x double> [[BLOCK33]], [[SPLAT_SPLAT35]]
+; CHECK-NEXT:    [[TMP40:%.*]] = fadd <2 x double> [[TMP37]], [[TMP39]]
 ; CHECK-NEXT:    [[BLOCK36:%.*]] = shufflevector <4 x double> [[SPLIT2]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
-; CHECK-NEXT:    [[TMP43:%.*]] = extractelement <2 x double> [[SPLIT5]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT37:%.*]] = insertelement <2 x double> poison, double [[TMP43]], i64 0
+; CHECK-NEXT:    [[TMP41:%.*]] = extractelement <2 x double> [[SPLIT5]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT37:%.*]] = insertelement <2 x double> poison, double [[TMP41]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT38:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT37]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP44:%.*]] = fmul <2 x double> [[BLOCK36]], [[SPLAT_SPLAT38]]
-; CHECK-NEXT:    [[TMP45:%.*]] = fadd <2 x double> [[TMP42]], [[TMP44]]
-; CHECK-NEXT:    [[TMP46:%.*]] = shufflevector <2 x double> [[TMP45]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP47:%.*]] = shufflevector <4 x double> [[TMP37]], <4 x double> [[TMP46]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
-; CHECK-NEXT:    store <4 x double> [[TMP27]], ptr [[C:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP53:%.*]] = getelementptr double, ptr [[C]], i64 4
-; CHECK-NEXT:    store <4 x double> [[TMP47]], ptr [[VEC_GEP53]], align 16
+; CHECK-NEXT:    [[TMP42:%.*]] = fmul <2 x double> [[BLOCK36]], [[SPLAT_SPLAT38]]
+; CHECK-NEXT:    [[TMP43:%.*]] = fadd <2 x double> [[TMP40]], [[TMP42]]
+; CHECK-NEXT:    [[TMP44:%.*]] = shufflevector <2 x double> [[TMP43]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP45:%.*]] = shufflevector <4 x double> [[TMP35]], <4 x double> [[TMP44]], <4 x i32> <i32 0, i32 1, i32 4, i32 5>
+; CHECK-NEXT:    store <4 x double> [[TMP25]], ptr [[C:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP47:%.*]] = getelementptr double, ptr [[C]], i64 4
+; CHECK-NEXT:    store <4 x double> [[TMP45]], ptr [[VEC_GEP47]], align 16
 ; CHECK-NEXT:    ret void
 ;
 entry:
@@ -598,264 +598,264 @@ define void @multiply_ntt_t(ptr %A, ptr %B, ptr %C, ptr %R) {
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <3 x double>, ptr [[A:%.*]], align 16
 ; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr double, ptr [[A]], i64 3
-; CHECK-NEXT:    [[COL_LOAD59:%.*]] = load <3 x double>, ptr [[VEC_GEP]], align 8
-; CHECK-NEXT:    [[VEC_GEP60:%.*]] = getelementptr double, ptr [[A]], i64 6
-; CHECK-NEXT:    [[COL_LOAD62:%.*]] = load <3 x double>, ptr [[VEC_GEP60]], align 16
-; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <3 x double> [[COL_LOAD]], <3 x double> [[COL_LOAD59]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
-; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <3 x double> [[COL_LOAD62]], <3 x double> poison, <6 x i32> <i32 0, i32 1, i32 2, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <6 x double> [[TMP1]], <6 x double> [[TMP2]], <9 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
-; CHECK-NEXT:    [[COL_LOAD64:%.*]] = load <3 x double>, ptr [[B:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP65:%.*]] = getelementptr double, ptr [[B]], i64 3
-; CHECK-NEXT:    [[COL_LOAD67:%.*]] = load <3 x double>, ptr [[VEC_GEP65]], align 8
-; CHECK-NEXT:    [[VEC_GEP68:%.*]] = getelementptr double, ptr [[B]], i64 6
-; CHECK-NEXT:    [[COL_LOAD70:%.*]] = load <3 x double>, ptr [[VEC_GEP68]], align 16
-; CHECK-NEXT:    [[COL_LOAD72:%.*]] = load <3 x double>, ptr [[C:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP73:%.*]] = getelementptr double, ptr [[C]], i64 3
-; CHECK-NEXT:    [[COL_LOAD75:%.*]] = load <3 x double>, ptr [[VEC_GEP73]], align 8
-; CHECK-NEXT:    [[VEC_GEP76:%.*]] = getelementptr double, ptr [[C]], i64 6
-; CHECK-NEXT:    [[COL_LOAD78:%.*]] = load <3 x double>, ptr [[VEC_GEP76]], align 16
-; CHECK-NEXT:    [[BLOCK79:%.*]] = shufflevector <3 x double> [[COL_LOAD72]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <3 x double> [[COL_LOAD64]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT80:%.*]] = insertelement <2 x double> poison, double [[TMP6]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT81:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT80]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP7:%.*]] = fmul <2 x double> [[BLOCK79]], [[SPLAT_SPLAT81]]
-; CHECK-NEXT:    [[BLOCK82:%.*]] = shufflevector <3 x double> [[COL_LOAD75]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP8:%.*]] = extractelement <3 x double> [[COL_LOAD64]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT83:%.*]] = insertelement <2 x double> poison, double [[TMP8]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT84:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT83]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP9:%.*]] = fmul <2 x double> [[BLOCK82]], [[SPLAT_SPLAT84]]
+; CHECK-NEXT:    [[COL_LOAD58:%.*]] = load <3 x double>, ptr [[VEC_GEP]], align 8
+; CHECK-NEXT:    [[VEC_GEP59:%.*]] = getelementptr double, ptr [[A]], i64 6
+; CHECK-NEXT:    [[COL_LOAD60:%.*]] = load <3 x double>, ptr [[VEC_GEP59]], align 16
+; CHECK-NEXT:    [[TMP0:%.*]] = shufflevector <3 x double> [[COL_LOAD]], <3 x double> [[COL_LOAD58]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <3 x double> [[COL_LOAD60]], <3 x double> poison, <6 x i32> <i32 0, i32 1, i32 2, i32 poison, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <6 x double> [[TMP0]], <6 x double> [[TMP1]], <9 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
+; CHECK-NEXT:    [[COL_LOAD61:%.*]] = load <3 x double>, ptr [[B:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP62:%.*]] = getelementptr double, ptr [[B]], i64 3
+; CHECK-NEXT:    [[COL_LOAD63:%.*]] = load <3 x double>, ptr [[VEC_GEP62]], align 8
+; CHECK-NEXT:    [[VEC_GEP64:%.*]] = getelementptr double, ptr [[B]], i64 6
+; CHECK-NEXT:    [[COL_LOAD65:%.*]] = load <3 x double>, ptr [[VEC_GEP64]], align 16
+; CHECK-NEXT:    [[COL_LOAD66:%.*]] = load <3 x double>, ptr [[C:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP67:%.*]] = getelementptr double, ptr [[C]], i64 3
+; CHECK-NEXT:    [[COL_LOAD68:%.*]] = load <3 x double>, ptr [[VEC_GEP67]], align 8
+; CHECK-NEXT:    [[VEC_GEP69:%.*]] = getelementptr double, ptr [[C]], i64 6
+; CHECK-NEXT:    [[COL_LOAD70:%.*]] = load <3 x double>, ptr [[VEC_GEP69]], align 16
+; CHECK-NEXT:    [[BLOCK71:%.*]] = shufflevector <3 x double> [[COL_LOAD66]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <3 x double> [[COL_LOAD61]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT72:%.*]] = insertelement <2 x double> poison, double [[TMP3]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT73:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT72]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP4:%.*]] = fmul <2 x double> [[BLOCK71]], [[SPLAT_SPLAT73]]
+; CHECK-NEXT:    [[BLOCK74:%.*]] = shufflevector <3 x double> [[COL_LOAD68]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <3 x double> [[COL_LOAD61]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT75:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT76:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT75]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP6:%.*]] = fmul <2 x double> [[BLOCK74]], [[SPLAT_SPLAT76]]
+; CHECK-NEXT:    [[TMP7:%.*]] = fadd <2 x double> [[TMP4]], [[TMP6]]
+; CHECK-NEXT:    [[BLOCK77:%.*]] = shufflevector <3 x double> [[COL_LOAD70]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP8:%.*]] = extractelement <3 x double> [[COL_LOAD61]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT78:%.*]] = insertelement <2 x double> poison, double [[TMP8]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT79:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT78]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP9:%.*]] = fmul <2 x double> [[BLOCK77]], [[SPLAT_SPLAT79]]
 ; CHECK-NEXT:    [[TMP10:%.*]] = fadd <2 x double> [[TMP7]], [[TMP9]]
-; CHECK-NEXT:    [[BLOCK85:%.*]] = shufflevector <3 x double> [[COL_LOAD78]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP11:%.*]] = extractelement <3 x double> [[COL_LOAD64]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT86:%.*]] = insertelement <2 x double> poison, double [[TMP11]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT87:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT86]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP12:%.*]] = fmul <2 x double> [[BLOCK85]], [[SPLAT_SPLAT87]]
-; CHECK-NEXT:    [[TMP13:%.*]] = fadd <2 x double> [[TMP10]], [[TMP12]]
-; CHECK-NEXT:    [[TMP14:%.*]] = shufflevector <2 x double> [[TMP13]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP15:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP14]], <3 x i32> <i32 3, i32 4, i32 2>
-; CHECK-NEXT:    [[BLOCK88:%.*]] = shufflevector <3 x double> [[COL_LOAD72]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP16:%.*]] = extractelement <3 x double> [[COL_LOAD64]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT89:%.*]] = insertelement <1 x double> poison, double [[TMP16]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT90:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT89]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP17:%.*]] = fmul <1 x double> [[BLOCK88]], [[SPLAT_SPLAT90]]
-; CHECK-NEXT:    [[BLOCK91:%.*]] = shufflevector <3 x double> [[COL_LOAD75]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP18:%.*]] = extractelement <3 x double> [[COL_LOAD64]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT92:%.*]] = insertelement <1 x double> poison, double [[TMP18]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT93:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT92]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP19:%.*]] = fmul <1 x double> [[BLOCK91]], [[SPLAT_SPLAT93]]
+; CHECK-NEXT:    [[TMP11:%.*]] = shufflevector <2 x double> [[TMP10]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP11]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[BLOCK80:%.*]] = shufflevector <3 x double> [[COL_LOAD66]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP13:%.*]] = extractelement <3 x double> [[COL_LOAD61]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT81:%.*]] = insertelement <1 x double> poison, double [[TMP13]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT82:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT81]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP14:%.*]] = fmul <1 x double> [[BLOCK80]], [[SPLAT_SPLAT82]]
+; CHECK-NEXT:    [[BLOCK83:%.*]] = shufflevector <3 x double> [[COL_LOAD68]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP15:%.*]] = extractelement <3 x double> [[COL_LOAD61]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT84:%.*]] = insertelement <1 x double> poison, double [[TMP15]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT85:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT84]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP16:%.*]] = fmul <1 x double> [[BLOCK83]], [[SPLAT_SPLAT85]]
+; CHECK-NEXT:    [[TMP17:%.*]] = fadd <1 x double> [[TMP14]], [[TMP16]]
+; CHECK-NEXT:    [[BLOCK86:%.*]] = shufflevector <3 x double> [[COL_LOAD70]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP18:%.*]] = extractelement <3 x double> [[COL_LOAD61]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT87:%.*]] = insertelement <1 x double> poison, double [[TMP18]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT88:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT87]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP19:%.*]] = fmul <1 x double> [[BLOCK86]], [[SPLAT_SPLAT88]]
 ; CHECK-NEXT:    [[TMP20:%.*]] = fadd <1 x double> [[TMP17]], [[TMP19]]
-; CHECK-NEXT:    [[BLOCK94:%.*]] = shufflevector <3 x double> [[COL_LOAD78]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP21:%.*]] = extractelement <3 x double> [[COL_LOAD64]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT95:%.*]] = insertelement <1 x double> poison, double [[TMP21]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT96:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT95]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP22:%.*]] = fmul <1 x double> [[BLOCK94]], [[SPLAT_SPLAT96]]
-; CHECK-NEXT:    [[TMP23:%.*]] = fadd <1 x double> [[TMP20]], [[TMP22]]
-; CHECK-NEXT:    [[TMP24:%.*]] = shufflevector <1 x double> [[TMP23]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP25:%.*]] = shufflevector <3 x double> [[TMP15]], <3 x double> [[TMP24]], <3 x i32> <i32 0, i32 1, i32 3>
-; CHECK-NEXT:    [[BLOCK97:%.*]] = shufflevector <3 x double> [[COL_LOAD72]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP26:%.*]] = extractelement <3 x double> [[COL_LOAD67]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT98:%.*]] = insertelement <2 x double> poison, double [[TMP26]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT99:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT98]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP27:%.*]] = fmul <2 x double> [[BLOCK97]], [[SPLAT_SPLAT99]]
-; CHECK-NEXT:    [[BLOCK100:%.*]] = shufflevector <3 x double> [[COL_LOAD75]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP28:%.*]] = extractelement <3 x double> [[COL_LOAD67]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT101:%.*]] = insertelement <2 x double> poison, double [[TMP28]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT102:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT101]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP29:%.*]] = fmul <2 x double> [[BLOCK100]], [[SPLAT_SPLAT102]]
+; CHECK-NEXT:    [[TMP21:%.*]] = shufflevector <1 x double> [[TMP20]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP22:%.*]] = shufflevector <3 x double> [[TMP12]], <3 x double> [[TMP21]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    [[BLOCK89:%.*]] = shufflevector <3 x double> [[COL_LOAD66]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP23:%.*]] = extractelement <3 x double> [[COL_LOAD63]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT90:%.*]] = insertelement <2 x double> poison, double [[TMP23]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT91:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT90]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP24:%.*]] = fmul <2 x double> [[BLOCK89]], [[SPLAT_SPLAT91]]
+; CHECK-NEXT:    [[BLOCK92:%.*]] = shufflevector <3 x double> [[COL_LOAD68]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP25:%.*]] = extractelement <3 x double> [[COL_LOAD63]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT93:%.*]] = insertelement <2 x double> poison, double [[TMP25]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT94:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT93]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP26:%.*]] = fmul <2 x double> [[BLOCK92]], [[SPLAT_SPLAT94]]
+; CHECK-NEXT:    [[TMP27:%.*]] = fadd <2 x double> [[TMP24]], [[TMP26]]
+; CHECK-NEXT:    [[BLOCK95:%.*]] = shufflevector <3 x double> [[COL_LOAD70]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP28:%.*]] = extractelement <3 x double> [[COL_LOAD63]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT96:%.*]] = insertelement <2 x double> poison, double [[TMP28]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT97:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT96]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP29:%.*]] = fmul <2 x double> [[BLOCK95]], [[SPLAT_SPLAT97]]
 ; CHECK-NEXT:    [[TMP30:%.*]] = fadd <2 x double> [[TMP27]], [[TMP29]]
-; CHECK-NEXT:    [[BLOCK103:%.*]] = shufflevector <3 x double> [[COL_LOAD78]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP31:%.*]] = extractelement <3 x double> [[COL_LOAD67]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT104:%.*]] = insertelement <2 x double> poison, double [[TMP31]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT105:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT104]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP32:%.*]] = fmul <2 x double> [[BLOCK103]], [[SPLAT_SPLAT105]]
-; CHECK-NEXT:    [[TMP33:%.*]] = fadd <2 x double> [[TMP30]], [[TMP32]]
-; CHECK-NEXT:    [[TMP34:%.*]] = shufflevector <2 x double> [[TMP33]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP35:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP34]], <3 x i32> <i32 3, i32 4, i32 2>
-; CHECK-NEXT:    [[BLOCK106:%.*]] = shufflevector <3 x double> [[COL_LOAD72]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP36:%.*]] = extractelement <3 x double> [[COL_LOAD67]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT107:%.*]] = insertelement <1 x double> poison, double [[TMP36]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT108:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT107]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP37:%.*]] = fmul <1 x double> [[BLOCK106]], [[SPLAT_SPLAT108]]
-; CHECK-NEXT:    [[BLOCK109:%.*]] = shufflevector <3 x double> [[COL_LOAD75]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP38:%.*]] = extractelement <3 x double> [[COL_LOAD67]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT110:%.*]] = insertelement <1 x double> poison, double [[TMP38]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT111:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT110]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP39:%.*]] = fmul <1 x double> [[BLOCK109]], [[SPLAT_SPLAT111]]
+; CHECK-NEXT:    [[TMP31:%.*]] = shufflevector <2 x double> [[TMP30]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP32:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP31]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[BLOCK98:%.*]] = shufflevector <3 x double> [[COL_LOAD66]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP33:%.*]] = extractelement <3 x double> [[COL_LOAD63]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT99:%.*]] = insertelement <1 x double> poison, double [[TMP33]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT100:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT99]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP34:%.*]] = fmul <1 x double> [[BLOCK98]], [[SPLAT_SPLAT100]]
+; CHECK-NEXT:    [[BLOCK101:%.*]] = shufflevector <3 x double> [[COL_LOAD68]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP35:%.*]] = extractelement <3 x double> [[COL_LOAD63]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT102:%.*]] = insertelement <1 x double> poison, double [[TMP35]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT103:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT102]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP36:%.*]] = fmul <1 x double> [[BLOCK101]], [[SPLAT_SPLAT103]]
+; CHECK-NEXT:    [[TMP37:%.*]] = fadd <1 x double> [[TMP34]], [[TMP36]]
+; CHECK-NEXT:    [[BLOCK104:%.*]] = shufflevector <3 x double> [[COL_LOAD70]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP38:%.*]] = extractelement <3 x double> [[COL_LOAD63]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT105:%.*]] = insertelement <1 x double> poison, double [[TMP38]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT106:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT105]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP39:%.*]] = fmul <1 x double> [[BLOCK104]], [[SPLAT_SPLAT106]]
 ; CHECK-NEXT:    [[TMP40:%.*]] = fadd <1 x double> [[TMP37]], [[TMP39]]
-; CHECK-NEXT:    [[BLOCK112:%.*]] = shufflevector <3 x double> [[COL_LOAD78]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP41:%.*]] = extractelement <3 x double> [[COL_LOAD67]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT113:%.*]] = insertelement <1 x double> poison, double [[TMP41]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT114:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT113]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP42:%.*]] = fmul <1 x double> [[BLOCK112]], [[SPLAT_SPLAT114]]
-; CHECK-NEXT:    [[TMP43:%.*]] = fadd <1 x double> [[TMP40]], [[TMP42]]
-; CHECK-NEXT:    [[TMP44:%.*]] = shufflevector <1 x double> [[TMP43]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP45:%.*]] = shufflevector <3 x double> [[TMP35]], <3 x double> [[TMP44]], <3 x i32> <i32 0, i32 1, i32 3>
-; CHECK-NEXT:    [[BLOCK115:%.*]] = shufflevector <3 x double> [[COL_LOAD72]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP46:%.*]] = extractelement <3 x double> [[COL_LOAD70]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT116:%.*]] = insertelement <2 x double> poison, double [[TMP46]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT117:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT116]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP47:%.*]] = fmul <2 x double> [[BLOCK115]], [[SPLAT_SPLAT117]]
-; CHECK-NEXT:    [[BLOCK118:%.*]] = shufflevector <3 x double> [[COL_LOAD75]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP48:%.*]] = extractelement <3 x double> [[COL_LOAD70]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT119:%.*]] = insertelement <2 x double> poison, double [[TMP48]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT120:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT119]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP49:%.*]] = fmul <2 x double> [[BLOCK118]], [[SPLAT_SPLAT120]]
+; CHECK-NEXT:    [[TMP41:%.*]] = shufflevector <1 x double> [[TMP40]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP42:%.*]] = shufflevector <3 x double> [[TMP32]], <3 x double> [[TMP41]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    [[BLOCK107:%.*]] = shufflevector <3 x double> [[COL_LOAD66]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP43:%.*]] = extractelement <3 x double> [[COL_LOAD65]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT108:%.*]] = insertelement <2 x double> poison, double [[TMP43]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT109:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT108]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP44:%.*]] = fmul <2 x double> [[BLOCK107]], [[SPLAT_SPLAT109]]
+; CHECK-NEXT:    [[BLOCK110:%.*]] = shufflevector <3 x double> [[COL_LOAD68]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP45:%.*]] = extractelement <3 x double> [[COL_LOAD65]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT111:%.*]] = insertelement <2 x double> poison, double [[TMP45]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT112:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT111]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP46:%.*]] = fmul <2 x double> [[BLOCK110]], [[SPLAT_SPLAT112]]
+; CHECK-NEXT:    [[TMP47:%.*]] = fadd <2 x double> [[TMP44]], [[TMP46]]
+; CHECK-NEXT:    [[BLOCK113:%.*]] = shufflevector <3 x double> [[COL_LOAD70]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT:    [[TMP48:%.*]] = extractelement <3 x double> [[COL_LOAD65]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT114:%.*]] = insertelement <2 x double> poison, double [[TMP48]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT115:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT114]], <2 x double> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP49:%.*]] = fmul <2 x double> [[BLOCK113]], [[SPLAT_SPLAT115]]
 ; CHECK-NEXT:    [[TMP50:%.*]] = fadd <2 x double> [[TMP47]], [[TMP49]]
-; CHECK-NEXT:    [[BLOCK121:%.*]] = shufflevector <3 x double> [[COL_LOAD78]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP51:%.*]] = extractelement <3 x double> [[COL_LOAD70]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT122:%.*]] = insertelement <2 x double> poison, double [[TMP51]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT123:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT122]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP52:%.*]] = fmul <2 x double> [[BLOCK121]], [[SPLAT_SPLAT123]]
-; CHECK-NEXT:    [[TMP53:%.*]] = fadd <2 x double> [[TMP50]], [[TMP52]]
-; CHECK-NEXT:    [[TMP54:%.*]] = shufflevector <2 x double> [[TMP53]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP55:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP54]], <3 x i32> <i32 3, i32 4, i32 2>
-; CHECK-NEXT:    [[BLOCK124:%.*]] = shufflevector <3 x double> [[COL_LOAD72]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP56:%.*]] = extractelement <3 x double> [[COL_LOAD70]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT125:%.*]] = insertelement <1 x double> poison, double [[TMP56]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT126:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT125]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP57:%.*]] = fmul <1 x double> [[BLOCK124]], [[SPLAT_SPLAT126]]
-; CHECK-NEXT:    [[BLOCK127:%.*]] = shufflevector <3 x double> [[COL_LOAD75]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP58:%.*]] = extractelement <3 x double> [[COL_LOAD70]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT128:%.*]] = insertelement <1 x double> poison, double [[TMP58]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT129:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT128]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP59:%.*]] = fmul <1 x double> [[BLOCK127]], [[SPLAT_SPLAT129]]
+; CHECK-NEXT:    [[TMP51:%.*]] = shufflevector <2 x double> [[TMP50]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP52:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP51]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[BLOCK116:%.*]] = shufflevector <3 x double> [[COL_LOAD66]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP53:%.*]] = extractelement <3 x double> [[COL_LOAD65]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT117:%.*]] = insertelement <1 x double> poison, double [[TMP53]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT118:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT117]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP54:%.*]] = fmul <1 x double> [[BLOCK116]], [[SPLAT_SPLAT118]]
+; CHECK-NEXT:    [[BLOCK119:%.*]] = shufflevector <3 x double> [[COL_LOAD68]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP55:%.*]] = extractelement <3 x double> [[COL_LOAD65]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT120:%.*]] = insertelement <1 x double> poison, double [[TMP55]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT121:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT120]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP56:%.*]] = fmul <1 x double> [[BLOCK119]], [[SPLAT_SPLAT121]]
+; CHECK-NEXT:    [[TMP57:%.*]] = fadd <1 x double> [[TMP54]], [[TMP56]]
+; CHECK-NEXT:    [[BLOCK122:%.*]] = shufflevector <3 x double> [[COL_LOAD70]], <3 x double> poison, <1 x i32> <i32 2>
+; CHECK-NEXT:    [[TMP58:%.*]] = extractelement <3 x double> [[COL_LOAD65]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT123:%.*]] = insertelement <1 x double> poison, double [[TMP58]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLAT124:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT123]], <1 x double> poison, <1 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP59:%.*]] = fmul <1 x double> [[BLOCK122]], [[SPLAT_SPLAT124]]
 ; CHECK-NEXT:    [[TMP60:%.*]] = fadd <1 x double> [[TMP57]], [[TMP59]]
-; CHECK-NEXT:    [[BLOCK130:%.*]] = shufflevector <3 x double> [[COL_LOAD78]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP61:%.*]] = extractelement <3 x double> [[COL_LOAD70]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT131:%.*]] = insertelement <1 x double> poison, double [[TMP61]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLAT132:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT131]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP62:%.*]] = fmul <1 x double> [[BLOCK130]], [[SPLAT_SPLAT132]]
-; CHECK-NEXT:    [[TMP63:%.*]] = fadd <1 x double> [[TMP60]], [[TMP62]]
-; CHECK-NEXT:    [[TMP64:%.*]] = shufflevector <1 x double> [[TMP63]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP65:%.*]] = shufflevector <3 x double> [[TMP55]], <3 x double> [[TMP64]], <3 x i32> <i32 0, i32 1, i32 3>
-; CHECK-NEXT:    [[TMP66:%.*]] = shufflevector <3 x double> [[TMP25]], <3 x double> [[TMP45]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
-; CHECK-NEXT:    [[TMP67:%.*]] = shufflevector <3 x double> [[TMP65]], <3 x double> poison, <6 x i32> <i32 0, i32 1, i32 2, i32 poison, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP68:%.*]] = shufflevector <6 x double> [[TMP66]], <6 x double> [[TMP67]], <9 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
-; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <9 x double> [[TMP68]], <9 x double> poison, <3 x i32> <i32 0, i32 1, i32 2>
-; CHECK-NEXT:    [[SPLIT2:%.*]] = shufflevector <9 x double> [[TMP68]], <9 x double> poison, <3 x i32> <i32 3, i32 4, i32 5>
-; CHECK-NEXT:    [[SPLIT3:%.*]] = shufflevector <9 x double> [[TMP68]], <9 x double> poison, <3 x i32> <i32 6, i32 7, i32 8>
-; CHECK-NEXT:    [[SPLIT4:%.*]] = shufflevector <9 x double> [[TMP3]], <9 x double> poison, <3 x i32> <i32 0, i32 1, i32 2>
-; CHECK-NEXT:    [[SPLIT5:%.*]] = shufflevector <9 x double> [[TMP3]], <9 x double> poison, <3 x i32> <i32 3, i32 4, i32 5>
-; CHECK-NEXT:    [[SPLIT6:%.*]] = shufflevector <9 x double> [[TMP3]], <9 x double> poison, <3 x i32> <i32 6, i32 7, i32 8>
+; CHECK-NEXT:    [[TMP61:%.*]] = shufflevector <1 x double> [[TMP60]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP62:%.*]] = shufflevector <3 x double> [[TMP52]], <3 x double> [[TMP61]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    [[TMP63:%.*]] = shufflevector <3 x double> [[TMP22]], <3 x double> [[TMP42]], <6 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5>
+; CHECK-NEXT:    [[TMP64:%.*]] = shufflevector <3 x double> [[TMP62]], <3 x double> poison, <6 x i32> <i32 0, i32 1, i32 2, i32 poison, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP65:%.*]] = shufflevector <6 x double> [[TMP63]], <6 x double> [[TMP64]], <9 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
+; CHECK-NEXT:    [[SPLIT:%.*]] = shufflevector <9 x double> [[TMP65]], <9 x double> poison, <3 x i32> <i32 0, i32 1, i32 2>
+; CHECK-NEXT:    [[SPLIT2:%.*]] = shufflevector <9 x double> [[TMP65]], <9 x double> poison, <3 x i32> <i32 3, i32 4, i32 5>
+; CHECK-NEXT:    [[SPLIT3:%.*]] = shufflevector <9 x double> [[TMP65]], <9 x double> poison, <3 x i32> <i32 6, i32 7, i32 8>
+; CHECK-NEXT:    [[SPLIT4:%.*]] = shufflevector <9 x double> [[TMP2]], <9 x double> poison, <3 x i32> <i32 0, i32 1, i32 2>
+; CHECK-NEXT:    [[SPLIT5:%.*]] = shufflevector <9 x double> [[TMP2]], <9 x double> poison, <3 x i32> <i32 3, i32 4, i32 5>
+; CHECK-NEXT:    [[SPLIT6:%.*]] = shufflevector <9 x double> [[TMP2]], <9 x double> poison, <3 x i32> <i32 6, i32 7, i32 8>
 ; CHECK-NEXT:    [[BLOCK:%.*]] = shufflevector <3 x double> [[SPLIT]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP69:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP69]], i64 0
+; CHECK-NEXT:    [[TMP66:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP66]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP70:%.*]] = fmul <2 x double> [[BLOCK]], [[SPLAT_SPLAT]]
+; CHECK-NEXT:    [[TMP67:%.*]] = fmul <2 x double> [[BLOCK]], [[SPLAT_SPLAT]]
 ; CHECK-NEXT:    [[BLOCK7:%.*]] = shufflevector <3 x double> [[SPLIT2]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP71:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT8:%.*]] = insertelement <2 x double> poison, double [[TMP71]], i64 0
+; CHECK-NEXT:    [[TMP68:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT8:%.*]] = insertelement <2 x double> poison, double [[TMP68]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT9:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT8]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP72:%.*]] = fmul <2 x double> [[BLOCK7]], [[SPLAT_SPLAT9]]
-; CHECK-NEXT:    [[TMP73:%.*]] = fadd <2 x double> [[TMP70]], [[TMP72]]
+; CHECK-NEXT:    [[TMP69:%.*]] = fmul <2 x double> [[BLOCK7]], [[SPLAT_SPLAT9]]
+; CHECK-NEXT:    [[TMP70:%.*]] = fadd <2 x double> [[TMP67]], [[TMP69]]
 ; CHECK-NEXT:    [[BLOCK10:%.*]] = shufflevector <3 x double> [[SPLIT3]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP74:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT11:%.*]] = insertelement <2 x double> poison, double [[TMP74]], i64 0
+; CHECK-NEXT:    [[TMP71:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT11:%.*]] = insertelement <2 x double> poison, double [[TMP71]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT12:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT11]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP75:%.*]] = fmul <2 x double> [[BLOCK10]], [[SPLAT_SPLAT12]]
-; CHECK-NEXT:    [[TMP76:%.*]] = fadd <2 x double> [[TMP73]], [[TMP75]]
-; CHECK-NEXT:    [[TMP77:%.*]] = shufflevector <2 x double> [[TMP76]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP78:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP77]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[TMP72:%.*]] = fmul <2 x double> [[BLOCK10]], [[SPLAT_SPLAT12]]
+; CHECK-NEXT:    [[TMP73:%.*]] = fadd <2 x double> [[TMP70]], [[TMP72]]
+; CHECK-NEXT:    [[TMP74:%.*]] = shufflevector <2 x double> [[TMP73]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP75:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP74]], <3 x i32> <i32 3, i32 4, i32 2>
 ; CHECK-NEXT:    [[BLOCK13:%.*]] = shufflevector <3 x double> [[SPLIT]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP79:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT14:%.*]] = insertelement <1 x double> poison, double [[TMP79]], i64 0
+; CHECK-NEXT:    [[TMP76:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT14:%.*]] = insertelement <1 x double> poison, double [[TMP76]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT15:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT14]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP80:%.*]] = fmul <1 x double> [[BLOCK13]], [[SPLAT_SPLAT15]]
+; CHECK-NEXT:    [[TMP77:%.*]] = fmul <1 x double> [[BLOCK13]], [[SPLAT_SPLAT15]]
 ; CHECK-NEXT:    [[BLOCK16:%.*]] = shufflevector <3 x double> [[SPLIT2]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP81:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT17:%.*]] = insertelement <1 x double> poison, double [[TMP81]], i64 0
+; CHECK-NEXT:    [[TMP78:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT17:%.*]] = insertelement <1 x double> poison, double [[TMP78]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT18:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT17]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP82:%.*]] = fmul <1 x double> [[BLOCK16]], [[SPLAT_SPLAT18]]
-; CHECK-NEXT:    [[TMP83:%.*]] = fadd <1 x double> [[TMP80]], [[TMP82]]
+; CHECK-NEXT:    [[TMP79:%.*]] = fmul <1 x double> [[BLOCK16]], [[SPLAT_SPLAT18]]
+; CHECK-NEXT:    [[TMP80:%.*]] = fadd <1 x double> [[TMP77]], [[TMP79]]
 ; CHECK-NEXT:    [[BLOCK19:%.*]] = shufflevector <3 x double> [[SPLIT3]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP84:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 0
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT20:%.*]] = insertelement <1 x double> poison, double [[TMP84]], i64 0
+; CHECK-NEXT:    [[TMP81:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 0
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT20:%.*]] = insertelement <1 x double> poison, double [[TMP81]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT21:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT20]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP85:%.*]] = fmul <1 x double> [[BLOCK19]], [[SPLAT_SPLAT21]]
-; CHECK-NEXT:    [[TMP86:%.*]] = fadd <1 x double> [[TMP83]], [[TMP85]]
-; CHECK-NEXT:    [[TMP87:%.*]] = shufflevector <1 x double> [[TMP86]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP88:%.*]] = shufflevector <3 x double> [[TMP78]], <3 x double> [[TMP87]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    [[TMP82:%.*]] = fmul <1 x double> [[BLOCK19]], [[SPLAT_SPLAT21]]
+; CHECK-NEXT:    [[TMP83:%.*]] = fadd <1 x double> [[TMP80]], [[TMP82]]
+; CHECK-NEXT:    [[TMP84:%.*]] = shufflevector <1 x double> [[TMP83]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP85:%.*]] = shufflevector <3 x double> [[TMP75]], <3 x double> [[TMP84]], <3 x i32> <i32 0, i32 1, i32 3>
 ; CHECK-NEXT:    [[BLOCK22:%.*]] = shufflevector <3 x double> [[SPLIT]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP89:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT23:%.*]] = insertelement <2 x double> poison, double [[TMP89]], i64 0
+; CHECK-NEXT:    [[TMP86:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT23:%.*]] = insertelement <2 x double> poison, double [[TMP86]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT24:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT23]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP90:%.*]] = fmul <2 x double> [[BLOCK22]], [[SPLAT_SPLAT24]]
+; CHECK-NEXT:    [[TMP87:%.*]] = fmul <2 x double> [[BLOCK22]], [[SPLAT_SPLAT24]]
 ; CHECK-NEXT:    [[BLOCK25:%.*]] = shufflevector <3 x double> [[SPLIT2]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP91:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT26:%.*]] = insertelement <2 x double> poison, double [[TMP91]], i64 0
+; CHECK-NEXT:    [[TMP88:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT26:%.*]] = insertelement <2 x double> poison, double [[TMP88]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT27:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT26]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP92:%.*]] = fmul <2 x double> [[BLOCK25]], [[SPLAT_SPLAT27]]
-; CHECK-NEXT:    [[TMP93:%.*]] = fadd <2 x double> [[TMP90]], [[TMP92]]
+; CHECK-NEXT:    [[TMP89:%.*]] = fmul <2 x double> [[BLOCK25]], [[SPLAT_SPLAT27]]
+; CHECK-NEXT:    [[TMP90:%.*]] = fadd <2 x double> [[TMP87]], [[TMP89]]
 ; CHECK-NEXT:    [[BLOCK28:%.*]] = shufflevector <3 x double> [[SPLIT3]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP94:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT29:%.*]] = insertelement <2 x double> poison, double [[TMP94]], i64 0
+; CHECK-NEXT:    [[TMP91:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT29:%.*]] = insertelement <2 x double> poison, double [[TMP91]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT30:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT29]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP95:%.*]] = fmul <2 x double> [[BLOCK28]], [[SPLAT_SPLAT30]]
-; CHECK-NEXT:    [[TMP96:%.*]] = fadd <2 x double> [[TMP93]], [[TMP95]]
-; CHECK-NEXT:    [[TMP97:%.*]] = shufflevector <2 x double> [[TMP96]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP98:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP97]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[TMP92:%.*]] = fmul <2 x double> [[BLOCK28]], [[SPLAT_SPLAT30]]
+; CHECK-NEXT:    [[TMP93:%.*]] = fadd <2 x double> [[TMP90]], [[TMP92]]
+; CHECK-NEXT:    [[TMP94:%.*]] = shufflevector <2 x double> [[TMP93]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP95:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP94]], <3 x i32> <i32 3, i32 4, i32 2>
 ; CHECK-NEXT:    [[BLOCK31:%.*]] = shufflevector <3 x double> [[SPLIT]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP99:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT32:%.*]] = insertelement <1 x double> poison, double [[TMP99]], i64 0
+; CHECK-NEXT:    [[TMP96:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT32:%.*]] = insertelement <1 x double> poison, double [[TMP96]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT33:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT32]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP100:%.*]] = fmul <1 x double> [[BLOCK31]], [[SPLAT_SPLAT33]]
+; CHECK-NEXT:    [[TMP97:%.*]] = fmul <1 x double> [[BLOCK31]], [[SPLAT_SPLAT33]]
 ; CHECK-NEXT:    [[BLOCK34:%.*]] = shufflevector <3 x double> [[SPLIT2]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP101:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT35:%.*]] = insertelement <1 x double> poison, double [[TMP101]], i64 0
+; CHECK-NEXT:    [[TMP98:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT35:%.*]] = insertelement <1 x double> poison, double [[TMP98]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT36:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT35]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP102:%.*]] = fmul <1 x double> [[BLOCK34]], [[SPLAT_SPLAT36]]
-; CHECK-NEXT:    [[TMP103:%.*]] = fadd <1 x double> [[TMP100]], [[TMP102]]
+; CHECK-NEXT:    [[TMP99:%.*]] = fmul <1 x double> [[BLOCK34]], [[SPLAT_SPLAT36]]
+; CHECK-NEXT:    [[TMP100:%.*]] = fadd <1 x double> [[TMP97]], [[TMP99]]
 ; CHECK-NEXT:    [[BLOCK37:%.*]] = shufflevector <3 x double> [[SPLIT3]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP104:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 1
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT38:%.*]] = insertelement <1 x double> poison, double [[TMP104]], i64 0
+; CHECK-NEXT:    [[TMP101:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 1
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT38:%.*]] = insertelement <1 x double> poison, double [[TMP101]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT39:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT38]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP105:%.*]] = fmul <1 x double> [[BLOCK37]], [[SPLAT_SPLAT39]]
-; CHECK-NEXT:    [[TMP106:%.*]] = fadd <1 x double> [[TMP103]], [[TMP105]]
-; CHECK-NEXT:    [[TMP107:%.*]] = shufflevector <1 x double> [[TMP106]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP108:%.*]] = shufflevector <3 x double> [[TMP98]], <3 x double> [[TMP107]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    [[TMP102:%.*]] = fmul <1 x double> [[BLOCK37]], [[SPLAT_SPLAT39]]
+; CHECK-NEXT:    [[TMP103:%.*]] = fadd <1 x double> [[TMP100]], [[TMP102]]
+; CHECK-NEXT:    [[TMP104:%.*]] = shufflevector <1 x double> [[TMP103]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP105:%.*]] = shufflevector <3 x double> [[TMP95]], <3 x double> [[TMP104]], <3 x i32> <i32 0, i32 1, i32 3>
 ; CHECK-NEXT:    [[BLOCK40:%.*]] = shufflevector <3 x double> [[SPLIT]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP109:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT41:%.*]] = insertelement <2 x double> poison, double [[TMP109]], i64 0
+; CHECK-NEXT:    [[TMP106:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT41:%.*]] = insertelement <2 x double> poison, double [[TMP106]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT42:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT41]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP110:%.*]] = fmul <2 x double> [[BLOCK40]], [[SPLAT_SPLAT42]]
+; CHECK-NEXT:    [[TMP107:%.*]] = fmul <2 x double> [[BLOCK40]], [[SPLAT_SPLAT42]]
 ; CHECK-NEXT:    [[BLOCK43:%.*]] = shufflevector <3 x double> [[SPLIT2]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP111:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT44:%.*]] = insertelement <2 x double> poison, double [[TMP111]], i64 0
+; CHECK-NEXT:    [[TMP108:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT44:%.*]] = insertelement <2 x double> poison, double [[TMP108]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT45:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT44]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP112:%.*]] = fmul <2 x double> [[BLOCK43]], [[SPLAT_SPLAT45]]
-; CHECK-NEXT:    [[TMP113:%.*]] = fadd <2 x double> [[TMP110]], [[TMP112]]
+; CHECK-NEXT:    [[TMP109:%.*]] = fmul <2 x double> [[BLOCK43]], [[SPLAT_SPLAT45]]
+; CHECK-NEXT:    [[TMP110:%.*]] = fadd <2 x double> [[TMP107]], [[TMP109]]
 ; CHECK-NEXT:    [[BLOCK46:%.*]] = shufflevector <3 x double> [[SPLIT3]], <3 x double> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP114:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT47:%.*]] = insertelement <2 x double> poison, double [[TMP114]], i64 0
+; CHECK-NEXT:    [[TMP111:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT47:%.*]] = insertelement <2 x double> poison, double [[TMP111]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT48:%.*]] = shufflevector <2 x double> [[SPLAT_SPLATINSERT47]], <2 x double> poison, <2 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP115:%.*]] = fmul <2 x double> [[BLOCK46]], [[SPLAT_SPLAT48]]
-; CHECK-NEXT:    [[TMP116:%.*]] = fadd <2 x double> [[TMP113]], [[TMP115]]
-; CHECK-NEXT:    [[TMP117:%.*]] = shufflevector <2 x double> [[TMP116]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
-; CHECK-NEXT:    [[TMP118:%.*]] = shufflevector <3 x double> undef, <3 x double> [[TMP117]], <3 x i32> <i32 3, i32 4, i32 2>
+; CHECK-NEXT:    [[TMP112:%.*]] = fmul <2 x double> [[BLOCK46]], [[SPLAT_SPLAT48]]
+; CHECK-NEXT:    [[TMP113:%.*]] = fadd <2 x double> [[TMP110]], [[TMP112]]
+; CHECK-NEXT:    [[TMP114:%.*]] = shufflevector <2 x double> [[TMP113]], <2 x double> poison, <3 x i32> <i32 0, i32 1, i32 poison>
+; CHECK-NEXT:    [[TMP115:%.*]] = shufflevector <3 x double> poison, <3 x double> [[TMP114]], <3 x i32> <i32 3, i32 4, i32 2>
 ; CHECK-NEXT:    [[BLOCK49:%.*]] = shufflevector <3 x double> [[SPLIT]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP119:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT50:%.*]] = insertelement <1 x double> poison, double [[TMP119]], i64 0
+; CHECK-NEXT:    [[TMP116:%.*]] = extractelement <3 x double> [[SPLIT4]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT50:%.*]] = insertelement <1 x double> poison, double [[TMP116]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT51:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT50]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP120:%.*]] = fmul <1 x double> [[BLOCK49]], [[SPLAT_SPLAT51]]
+; CHECK-NEXT:    [[TMP117:%.*]] = fmul <1 x double> [[BLOCK49]], [[SPLAT_SPLAT51]]
 ; CHECK-NEXT:    [[BLOCK52:%.*]] = shufflevector <3 x double> [[SPLIT2]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP121:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT53:%.*]] = insertelement <1 x double> poison, double [[TMP121]], i64 0
+; CHECK-NEXT:    [[TMP118:%.*]] = extractelement <3 x double> [[SPLIT5]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT53:%.*]] = insertelement <1 x double> poison, double [[TMP118]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT54:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT53]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP122:%.*]] = fmul <1 x double> [[BLOCK52]], [[SPLAT_SPLAT54]]
-; CHECK-NEXT:    [[TMP123:%.*]] = fadd <1 x double> [[TMP120]], [[TMP122]]
+; CHECK-NEXT:    [[TMP119:%.*]] = fmul <1 x double> [[BLOCK52]], [[SPLAT_SPLAT54]]
+; CHECK-NEXT:    [[TMP120:%.*]] = fadd <1 x double> [[TMP117]], [[TMP119]]
 ; CHECK-NEXT:    [[BLOCK55:%.*]] = shufflevector <3 x double> [[SPLIT3]], <3 x double> poison, <1 x i32> <i32 2>
-; CHECK-NEXT:    [[TMP124:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 2
-; CHECK-NEXT:    [[SPLAT_SPLATINSERT56:%.*]] = insertelement <1 x double> poison, double [[TMP124]], i64 0
+; CHECK-NEXT:    [[TMP121:%.*]] = extractelement <3 x double> [[SPLIT6]], i64 2
+; CHECK-NEXT:    [[SPLAT_SPLATINSERT56:%.*]] = insertelement <1 x double> poison, double [[TMP121]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLAT57:%.*]] = shufflevector <1 x double> [[SPLAT_SPLATINSERT56]], <1 x double> poison, <1 x i32> zeroinitializer
-; CHECK-NEXT:    [[TMP125:%.*]] = fmul <1 x double> [[BLOCK55]], [[SPLAT_SPLAT57]]
-; CHECK-NEXT:    [[TMP126:%.*]] = fadd <1 x double> [[TMP123]], [[TMP125]]
-; CHECK-NEXT:    [[TMP127:%.*]] = shufflevector <1 x double> [[TMP126]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP128:%.*]] = shufflevector <3 x double> [[TMP118]], <3 x double> [[TMP127]], <3 x i32> <i32 0, i32 1, i32 3>
-; CHECK-NEXT:    store <3 x double> [[TMP88]], ptr [[R:%.*]], align 16
-; CHECK-NEXT:    [[VEC_GEP134:%.*]] = getelementptr double, ptr [[R]], i64 3
-; CHECK-NEXT:    store <3 x double> [[TMP108]], ptr [[VEC_GEP134]], align 8
-; CHECK-NEXT:    [[VEC_GEP136:%.*]] = getelementptr double, ptr [[R]], i64 6
-; CHECK-NEXT:    store <3 x double> [[TMP128]], ptr [[VEC_GEP136]], align 16
+; CHECK-NEXT:    [[TMP122:%.*]] = fmul <1 x double> [[BLOCK55]], [[SPLAT_SPLAT57]]
+; CHECK-NEXT:    [[TMP123:%.*]] = fadd <1 x double> [[TMP120]], [[TMP122]]
+; CHECK-NEXT:    [[TMP124:%.*]] = shufflevector <1 x double> [[TMP123]], <1 x double> poison, <3 x i32> <i32 0, i32 poison, i32 poison>
+; CHECK-NEXT:    [[TMP125:%.*]] = shufflevector <3 x double> [[TMP115]], <3 x double> [[TMP124]], <3 x i32> <i32 0, i32 1, i32 3>
+; CHECK-NEXT:    store <3 x double> [[TMP85]], ptr [[R:%.*]], align 16
+; CHECK-NEXT:    [[VEC_GEP125:%.*]] = getelementptr double, ptr [[R]], i64 3
+; CHECK-NEXT:    store <3 x double> [[TMP105]], ptr [[VEC_GEP125]], align 8
+; CHECK-NEXT:    [[VEC_GEP126:%.*]] = getelementptr double, ptr [[R]], i64 6
+; CHECK-NEXT:    store <3 x double> [[TMP125]], ptr [[VEC_GEP126]], align 16
 ; CHECK-NEXT:    ret void
 ;
 entry:
@@ -890,7 +890,7 @@ define <4 x float> @mult_tt_same_op(<4 x float> %a) {
 ; CHECK-NEXT:    [[TMP3:%.*]] = fmul <2 x float> [[BLOCK4]], [[SPLAT_SPLAT6]]
 ; CHECK-NEXT:    [[TMP4:%.*]] = fadd <2 x float> [[TMP1]], [[TMP3]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float> [[TMP4]], <2 x float> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP6:%.*]] = shufflevector <2 x float> undef, <2 x float> [[TMP5]], <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP6:%.*]] = shufflevector <2 x float> poison, <2 x float> [[TMP5]], <2 x i32> <i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK7:%.*]] = shufflevector <2 x float> [[SPLIT]], <2 x float> poison, <2 x i32> <i32 0, i32 1>
 ; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <2 x float> [[SPLIT3]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLATINSERT8:%.*]] = insertelement <2 x float> poison, float [[TMP7]], i64 0
@@ -903,7 +903,7 @@ define <4 x float> @mult_tt_same_op(<4 x float> %a) {
 ; CHECK-NEXT:    [[TMP10:%.*]] = fmul <2 x float> [[BLOCK10]], [[SPLAT_SPLAT12]]
 ; CHECK-NEXT:    [[TMP11:%.*]] = fadd <2 x float> [[TMP8]], [[TMP10]]
 ; CHECK-NEXT:    [[TMP12:%.*]] = shufflevector <2 x float> [[TMP11]], <2 x float> poison, <2 x i32> <i32 0, i32 1>
-; CHECK-NEXT:    [[TMP13:%.*]] = shufflevector <2 x float> undef, <2 x float> [[TMP12]], <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT:    [[TMP13:%.*]] = shufflevector <2 x float> poison, <2 x float> [[TMP12]], <2 x i32> <i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP14:%.*]] = extractelement <2 x float> [[TMP6]], i64 0
 ; CHECK-NEXT:    [[TMP15:%.*]] = insertelement <2 x float> poison, float [[TMP14]], i64 0
 ; CHECK-NEXT:    [[TMP16:%.*]] = extractelement <2 x float> [[TMP13]], i64 0
@@ -925,13 +925,13 @@ define <6 x double> @transpose_of_transpose_of_non_matrix_op(ptr %a) {
 ; CHECK-LABEL: @transpose_of_transpose_of_non_matrix_op(
 ; CHECK-NEXT:    [[COL_LOAD:%.*]] = load <2 x double>, ptr [[A:%.*]], align 8
 ; CHECK-NEXT:    [[VEC_GEP:%.*]] = getelementptr double, ptr [[A]], i64 4
-; CHECK-NEXT:    [[COL_LOAD2:%.*]] = load <2 x double>, ptr [[VEC_GEP]], align 8
-; CHECK-NEXT:    [[VEC_GEP3:%.*]] = getelementptr double, ptr [[A]], i64 8
-; CHECK-NEXT:    [[COL_LOAD5:%.*]] = load <2 x double>, ptr [[VEC_GEP3]], align 8
-; CHECK-NEXT:    [[VEC_GEP6:%.*]] = getelementptr double, ptr [[A]], i64 12
-; CHECK-NEXT:    [[COL_LOAD8:%.*]] = load <2 x double>, ptr [[VEC_GEP6]], align 8
-; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <2 x double> [[COL_LOAD]], <2 x double> [[COL_LOAD2]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
-; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <2 x double> [[COL_LOAD5]], <2 x double> [[COL_LOAD8]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT:    [[COL_LOAD1:%.*]] = load <2 x double>, ptr [[VEC_GEP]], align 8
+; CHECK-NEXT:    [[VEC_GEP2:%.*]] = getelementptr double, ptr [[A]], i64 8
+; CHECK-NEXT:    [[COL_LOAD3:%.*]] = load <2 x double>, ptr [[VEC_GEP2]], align 8
+; CHECK-NEXT:    [[VEC_GEP4:%.*]] = getelementptr double, ptr [[A]], i64 12
+; CHECK-NEXT:    [[COL_LOAD5:%.*]] = load <2 x double>, ptr [[VEC_GEP4]], align 8
+; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <2 x double> [[COL_LOAD]], <2 x double> [[COL_LOAD1]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <2 x double> [[COL_LOAD3]], <2 x double> [[COL_LOAD5]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP3:%.*]] = shufflevector <4 x double> [[TMP1]], <4 x double> [[TMP2]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
 ; CHECK-NEXT:    [[SHUF:%.*]] = shufflevector <8 x double> [[TMP3]], <8 x double> poison, <6 x i32> <i32 0, i32 1, i32 2, i32 4, i32 5, i32 6>
 ; CHECK-NEXT:    ret <6 x double> [[SHUF]]
@@ -962,7 +962,7 @@ define <12 x double> @factor_transpose(<6 x double> %a, <8 x double> %b) {
 ; CHECK-NEXT:    [[TMP4:%.*]] = fmul <2 x double> [[BLOCK5]], [[SPLAT_SPLAT7]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = fadd <2 x double> [[TMP2]], [[TMP4]]
 ; CHECK-NEXT:    [[TMP6:%.*]] = shufflevector <2 x double> [[TMP5]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP7:%.*]] = shufflevector <4 x double> undef, <4 x double> [[TMP6]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP7:%.*]] = shufflevector <4 x double> poison, <4 x double> [[TMP6]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK8:%.*]] = shufflevector <4 x double> [[SPLIT]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP8:%.*]] = extractelement <2 x double> [[SPLIT2]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLATINSERT9:%.*]] = insertelement <2 x double> poison, double [[TMP8]], i64 0
@@ -988,7 +988,7 @@ define <12 x double> @factor_transpose(<6 x double> %a, <8 x double> %b) {
 ; CHECK-NEXT:    [[TMP18:%.*]] = fmul <2 x double> [[BLOCK17]], [[SPLAT_SPLAT19]]
 ; CHECK-NEXT:    [[TMP19:%.*]] = fadd <2 x double> [[TMP16]], [[TMP18]]
 ; CHECK-NEXT:    [[TMP20:%.*]] = shufflevector <2 x double> [[TMP19]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP21:%.*]] = shufflevector <4 x double> undef, <4 x double> [[TMP20]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP21:%.*]] = shufflevector <4 x double> poison, <4 x double> [[TMP20]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK20:%.*]] = shufflevector <4 x double> [[SPLIT]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP22:%.*]] = extractelement <2 x double> [[SPLIT3]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLATINSERT21:%.*]] = insertelement <2 x double> poison, double [[TMP22]], i64 0
@@ -1014,7 +1014,7 @@ define <12 x double> @factor_transpose(<6 x double> %a, <8 x double> %b) {
 ; CHECK-NEXT:    [[TMP32:%.*]] = fmul <2 x double> [[BLOCK29]], [[SPLAT_SPLAT31]]
 ; CHECK-NEXT:    [[TMP33:%.*]] = fadd <2 x double> [[TMP30]], [[TMP32]]
 ; CHECK-NEXT:    [[TMP34:%.*]] = shufflevector <2 x double> [[TMP33]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
-; CHECK-NEXT:    [[TMP35:%.*]] = shufflevector <4 x double> undef, <4 x double> [[TMP34]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
+; CHECK-NEXT:    [[TMP35:%.*]] = shufflevector <4 x double> poison, <4 x double> [[TMP34]], <4 x i32> <i32 4, i32 5, i32 2, i32 3>
 ; CHECK-NEXT:    [[BLOCK32:%.*]] = shufflevector <4 x double> [[SPLIT]], <4 x double> poison, <2 x i32> <i32 2, i32 3>
 ; CHECK-NEXT:    [[TMP36:%.*]] = extractelement <2 x double> [[SPLIT4]], i64 0
 ; CHECK-NEXT:    [[SPLAT_SPLATINSERT33:%.*]] = insertelement <2 x double> poison, double [[TMP36]], i64 0


        


More information about the llvm-commits mailing list