1 files changed, 74 insertions, 68 deletions
diff --git a/llvm/test/CodeGen/X86/vector-trunc.ll b/llvm/test/CodeGen/X86/vector-trunc.ll
index e9472b80871..d4874c18870 100644
--- a/llvm/test/CodeGen/X86/vector-trunc.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc.ll
@@ -296,22 +296,32 @@ define void @trunc8i64_8i8(<8 x i64> %a) {
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
-; AVX2-LABEL: trunc8i64_8i8:
-; AVX2:       # %bb.0: # %entry
-; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm2
-; AVX2-NEXT:    vmovdqa {{.*#+}} xmm3 = <u,u,0,8,u,u,u,u,u,u,u,u,u,u,u,u>
-; AVX2-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
-; AVX2-NEXT:    vpshufb %xmm3, %xmm1, %xmm1
-; AVX2-NEXT:    vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-NEXT:    vmovdqa {{.*#+}} xmm3 = <0,8,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
-; AVX2-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
-; AVX2-NEXT:    vpshufb %xmm3, %xmm0, %xmm0
-; AVX2-NEXT:    vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
-; AVX2-NEXT:    vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
-; AVX2-NEXT:    vmovq %xmm0, (%rax)
-; AVX2-NEXT:    vzeroupper
-; AVX2-NEXT:    retq
+; AVX2-SLOW-LABEL: trunc8i64_8i8:
+; AVX2-SLOW:       # %bb.0: # %entry
+; AVX2-SLOW-NEXT:    vextractf128 $1, %ymm0, %xmm2
+; AVX2-SLOW-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
+; AVX2-SLOW-NEXT:    vextractf128 $1, %ymm1, %xmm2
+; AVX2-SLOW-NEXT:    vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
+; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} xmm2 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
+; AVX2-SLOW-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX2-SLOW-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
+; AVX2-SLOW-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; AVX2-SLOW-NEXT:    vmovq %xmm0, (%rax)
+; AVX2-SLOW-NEXT:    vzeroupper
+; AVX2-SLOW-NEXT:    retq
+;
+; AVX2-FAST-LABEL: trunc8i64_8i8:
+; AVX2-FAST:       # %bb.0: # %entry
+; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm2 = [0,2,4,6,4,6,6,7]
+; AVX2-FAST-NEXT:    vpermd %ymm0, %ymm2, %ymm0
+; AVX2-FAST-NEXT:    vpermd %ymm1, %ymm2, %ymm1
+; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm2 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
+; AVX2-FAST-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX2-FAST-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
+; AVX2-FAST-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; AVX2-FAST-NEXT:    vmovq %xmm0, (%rax)
+; AVX2-FAST-NEXT:    vzeroupper
+; AVX2-FAST-NEXT:    retq
 ;
 ; AVX512-LABEL: trunc8i64_8i8:
 ; AVX512:       # %bb.0: # %entry
@@ -577,11 +587,9 @@ define void @trunc8i32_8i8(<8 x i32> %a) {
 ;
 ; AVX2-LABEL: trunc8i32_8i8:
 ; AVX2:       # %bb.0: # %entry
-; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
-; AVX2-NEXT:    vmovdqa {{.*#+}} xmm2 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
-; AVX2-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
-; AVX2-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
-; AVX2-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; AVX2-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15,16,17,20,21,24,25,28,29,24,25,28,29,28,29,30,31]
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
+; AVX2-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
 ; AVX2-NEXT:    vmovq %xmm0, (%rax)
 ; AVX2-NEXT:    vzeroupper
 ; AVX2-NEXT:    retq
@@ -589,7 +597,8 @@ define void @trunc8i32_8i8(<8 x i32> %a) {
 ; AVX512F-LABEL: trunc8i32_8i8:
 ; AVX512F:       # %bb.0: # %entry
 ; AVX512F-NEXT:    # kill: def $ymm0 killed $ymm0 def $zmm0
-; AVX512F-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512F-NEXT:    vpmovdw %zmm0, %ymm0
+; AVX512F-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
 ; AVX512F-NEXT:    vmovq %xmm0, (%rax)
 ; AVX512F-NEXT:    vzeroupper
 ; AVX512F-NEXT:    retq
@@ -603,7 +612,8 @@ define void @trunc8i32_8i8(<8 x i32> %a) {
 ; AVX512BW-LABEL: trunc8i32_8i8:
 ; AVX512BW:       # %bb.0: # %entry
 ; AVX512BW-NEXT:    # kill: def $ymm0 killed $ymm0 def $zmm0
-; AVX512BW-NEXT:    vpmovdb %zmm0, %xmm0
+; AVX512BW-NEXT:    vpmovdw %zmm0, %ymm0
+; AVX512BW-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
 ; AVX512BW-NEXT:    vmovq %xmm0, (%rax)
 ; AVX512BW-NEXT:    vzeroupper
 ; AVX512BW-NEXT:    retq
@@ -1470,53 +1480,39 @@ define <8 x i16> @trunc2x4i64_8i16(<4 x i64> %a, <4 x i64> %b) {
 ;
 ; AVX1-LABEL: trunc2x4i64_8i16:
 ; AVX1:       # %bb.0: # %entry
-; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm2
-; AVX1-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; AVX1-NEXT:    vpshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
-; AVX1-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; AVX1-NEXT:    vpshuflw {{.*#+}} xmm1 = xmm1[0,1,0,2,4,5,6,7]
-; AVX1-NEXT:    vpunpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
-; AVX1-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; AVX1-NEXT:    vpshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
-; AVX1-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; AVX1-NEXT:    vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; AVX1-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; AVX1-NEXT:    vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
+; AVX1-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
+; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm2
+; AVX1-NEXT:    vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
+; AVX1-NEXT:    vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX1-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX1-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
+; AVX1-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; AVX1-NEXT:    vzeroupper
 ; AVX1-NEXT:    retq
 ;
 ; AVX2-SLOW-LABEL: trunc2x4i64_8i16:
 ; AVX2-SLOW:       # %bb.0: # %entry
-; AVX2-SLOW-NEXT:    vextracti128 $1, %ymm1, %xmm2
-; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; AVX2-SLOW-NEXT:    vpshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
-; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; AVX2-SLOW-NEXT:    vpshuflw {{.*#+}} xmm1 = xmm1[0,1,0,2,4,5,6,7]
-; AVX2-SLOW-NEXT:    vpunpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
-; AVX2-SLOW-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
-; AVX2-SLOW-NEXT:    vpshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
-; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; AVX2-SLOW-NEXT:    vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; AVX2-SLOW-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; AVX2-SLOW-NEXT:    vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; AVX2-SLOW-NEXT:    vextractf128 $1, %ymm0, %xmm2
+; AVX2-SLOW-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
+; AVX2-SLOW-NEXT:    vextractf128 $1, %ymm1, %xmm2
+; AVX2-SLOW-NEXT:    vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
+; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX2-SLOW-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX2-SLOW-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
+; AVX2-SLOW-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; AVX2-SLOW-NEXT:    vzeroupper
 ; AVX2-SLOW-NEXT:    retq
 ;
 ; AVX2-FAST-LABEL: trunc2x4i64_8i16:
 ; AVX2-FAST:       # %bb.0: # %entry
-; AVX2-FAST-NEXT:    vextracti128 $1, %ymm1, %xmm2
-; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm3 = [0,1,2,3,0,1,8,9,8,9,10,11,12,13,14,15]
-; AVX2-FAST-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
-; AVX2-FAST-NEXT:    vpshufb %xmm3, %xmm1, %xmm1
-; AVX2-FAST-NEXT:    vpunpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
-; AVX2-FAST-NEXT:    vextracti128 $1, %ymm0, %xmm2
-; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm3 = [0,1,8,9,8,9,10,11,8,9,10,11,12,13,14,15]
-; AVX2-FAST-NEXT:    vpshufb %xmm3, %xmm2, %xmm2
-; AVX2-FAST-NEXT:    vpshufb %xmm3, %xmm0, %xmm0
-; AVX2-FAST-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
-; AVX2-FAST-NEXT:    vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
+; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm2 = [0,2,4,6,4,6,6,7]
+; AVX2-FAST-NEXT:    vpermd %ymm0, %ymm2, %ymm0
+; AVX2-FAST-NEXT:    vpermd %ymm1, %ymm2, %ymm1
+; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX2-FAST-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX2-FAST-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
+; AVX2-FAST-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; AVX2-FAST-NEXT:    vzeroupper
 ; AVX2-FAST-NEXT:    retq
 ;
@@ -1524,16 +1520,22 @@ define <8 x i16> @trunc2x4i64_8i16(<4 x i64> %a, <4 x i64> %b) {
 ; AVX512F:       # %bb.0: # %entry
 ; AVX512F-NEXT:    # kill: def $ymm1 killed $ymm1 def $zmm1
 ; AVX512F-NEXT:    # kill: def $ymm0 killed $ymm0 def $zmm0
-; AVX512F-NEXT:    vpmovqw %zmm0, %xmm0
-; AVX512F-NEXT:    vpmovqw %zmm1, %xmm1
+; AVX512F-NEXT:    vpmovqd %zmm0, %ymm0
+; AVX512F-NEXT:    vpmovqd %zmm1, %ymm1
+; AVX512F-NEXT:    vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX512F-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX512F-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX512F-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; AVX512F-NEXT:    vzeroupper
 ; AVX512F-NEXT:    retq
 ;
 ; AVX512VL-LABEL: trunc2x4i64_8i16:
 ; AVX512VL:       # %bb.0: # %entry
-; AVX512VL-NEXT:    vpmovqw %ymm0, %xmm0
-; AVX512VL-NEXT:    vpmovqw %ymm1, %xmm1
+; AVX512VL-NEXT:    vpmovqd %ymm0, %xmm0
+; AVX512VL-NEXT:    vpmovqd %ymm1, %xmm1
+; AVX512VL-NEXT:    vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX512VL-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX512VL-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX512VL-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; AVX512VL-NEXT:    vzeroupper
 ; AVX512VL-NEXT:    retq
@@ -1542,17 +1544,21 @@ define <8 x i16> @trunc2x4i64_8i16(<4 x i64> %a, <4 x i64> %b) {
 ; AVX512BW:       # %bb.0: # %entry
 ; AVX512BW-NEXT:    # kill: def $ymm1 killed $ymm1 def $zmm1
 ; AVX512BW-NEXT:    # kill: def $ymm0 killed $ymm0 def $zmm0
-; AVX512BW-NEXT:    vpmovqw %zmm0, %xmm0
-; AVX512BW-NEXT:    vpmovqw %zmm1, %xmm1
+; AVX512BW-NEXT:    vpmovqd %zmm0, %ymm0
+; AVX512BW-NEXT:    vpmovqd %zmm1, %ymm1
+; AVX512BW-NEXT:    vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX512BW-NEXT:    vpshufb %xmm2, %xmm1, %xmm1
+; AVX512BW-NEXT:    vpshufb %xmm2, %xmm0, %xmm0
 ; AVX512BW-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
 ; AVX512BW-NEXT:    vzeroupper
 ; AVX512BW-NEXT:    retq
 ;
 ; AVX512BWVL-LABEL: trunc2x4i64_8i16:
 ; AVX512BWVL:       # %bb.0: # %entry
-; AVX512BWVL-NEXT:    vpmovqw %ymm0, %xmm0
-; AVX512BWVL-NEXT:    vpmovqw %ymm1, %xmm1
-; AVX512BWVL-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX512BWVL-NEXT:    vpmovqd %ymm0, %xmm2
+; AVX512BWVL-NEXT:    vpmovqd %ymm1, %xmm1
+; AVX512BWVL-NEXT:    vmovdqa {{.*#+}} xmm0 = [0,2,4,6,8,10,12,14]
+; AVX512BWVL-NEXT:    vpermi2w %xmm1, %xmm2, %xmm0
 ; AVX512BWVL-NEXT:    vzeroupper
 ; AVX512BWVL-NEXT:    retq
 entry: