[X86] Add test cases for packus/ssat/usat 32i32->v32i8 test cases. NFC

llvm-svn: 374459

diff --git a/llvm/test/CodeGen/X86/min-legal-vector-width.ll b/llvm/test/CodeGen/X86/min-legal-vector-width.ll
index 5c851c1c723845023f45b9424a4d5947b0e3f613..832692480e0e1e13ebae9b11c9c07b1d5c70edb3 100644 (file)
--- a/llvm/test/CodeGen/X86/min-legal-vector-width.ll
+++ b/llvm/test/CodeGen/X86/min-legal-vector-width.ll
@@ -1097,3 +1097,22 @@ define <16 x i8> @trunc_packus_v16i32_v16i8(<16 x i32>* %p, <16 x i8>* %q) "min-
   %f = trunc <16 x i32> %e to <16 x i8>
   ret <16 x i8> %f
 }
+
+define <32 x i8> @trunc_packus_v32i32_v32i8(<32 x i32> %a0) {
+; CHECK-LABEL: trunc_packus_v32i32_v32i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; CHECK-NEXT:    vpmaxsd %zmm2, %zmm0, %zmm0
+; CHECK-NEXT:    vpmovusdb %zmm0, %xmm0
+; CHECK-NEXT:    vpmaxsd %zmm2, %zmm1, %zmm1
+; CHECK-NEXT:    vpmovusdb %zmm1, %xmm1
+; CHECK-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; CHECK-NEXT:    retq
+  %1 = icmp slt <32 x i32> %a0, <i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255>
+  %2 = select <32 x i1> %1, <32 x i32> %a0, <32 x i32> <i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255>
+  %3 = icmp sgt <32 x i32> %2, zeroinitializer
+  %4 = select <32 x i1> %3, <32 x i32> %2, <32 x i32> zeroinitializer
+  %5 = trunc <32 x i32> %4 to <32 x i8>
+  ret <32 x i8> %5
+}
+

diff --git a/llvm/test/CodeGen/X86/vector-trunc-packus.ll b/llvm/test/CodeGen/X86/vector-trunc-packus.ll
index b0d6a20bdf38b256d26a9decaf21729677003d92..704ac81394baa47a55fd70becd8adadf87f20634 100644 (file)
--- a/llvm/test/CodeGen/X86/vector-trunc-packus.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc-packus.ll
@@ -3087,3 +3087,57 @@ define <32 x i8> @trunc_packus_v32i16_v32i8(<32 x i16> %a0) {
   %5 = trunc <32 x i16> %4 to <32 x i8>
   ret <32 x i8> %5
 }
+
+define <32 x i8> @trunc_packus_v32i32_v32i8(<32 x i32> %a0) {
+; SSE-LABEL: trunc_packus_v32i32_v32i8:
+; SSE:       # %bb.0:
+; SSE-NEXT:    packssdw %xmm3, %xmm2
+; SSE-NEXT:    packssdw %xmm1, %xmm0
+; SSE-NEXT:    packuswb %xmm2, %xmm0
+; SSE-NEXT:    packssdw %xmm7, %xmm6
+; SSE-NEXT:    packssdw %xmm5, %xmm4
+; SSE-NEXT:    packuswb %xmm6, %xmm4
+; SSE-NEXT:    movdqa %xmm4, %xmm1
+; SSE-NEXT:    retq
+;
+; AVX1-LABEL: trunc_packus_v32i32_v32i8:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vextractf128 $1, %ymm3, %xmm4
+; AVX1-NEXT:    vpackssdw %xmm4, %xmm3, %xmm3
+; AVX1-NEXT:    vextractf128 $1, %ymm2, %xmm4
+; AVX1-NEXT:    vpackssdw %xmm4, %xmm2, %xmm2
+; AVX1-NEXT:    vpackuswb %xmm3, %xmm2, %xmm2
+; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm3
+; AVX1-NEXT:    vpackssdw %xmm3, %xmm1, %xmm1
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm3
+; AVX1-NEXT:    vpackssdw %xmm3, %xmm0, %xmm0
+; AVX1-NEXT:    vpackuswb %xmm1, %xmm0, %xmm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: trunc_packus_v32i32_v32i8:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vpackssdw %ymm3, %ymm2, %ymm2
+; AVX2-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[0,2,1,3]
+; AVX2-NEXT:    vpackssdw %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
+; AVX2-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: trunc_packus_v32i32_v32i8:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; AVX512-NEXT:    vpmaxsd %zmm2, %zmm0, %zmm0
+; AVX512-NEXT:    vpmovusdb %zmm0, %xmm0
+; AVX512-NEXT:    vpmaxsd %zmm2, %zmm1, %zmm1
+; AVX512-NEXT:    vpmovusdb %zmm1, %xmm1
+; AVX512-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX512-NEXT:    retq
+  %1 = icmp slt <32 x i32> %a0, <i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255>
+  %2 = select <32 x i1> %1, <32 x i32> %a0, <32 x i32> <i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255>
+  %3 = icmp sgt <32 x i32> %2, zeroinitializer
+  %4 = select <32 x i1> %3, <32 x i32> %2, <32 x i32> zeroinitializer
+  %5 = trunc <32 x i32> %4 to <32 x i8>
+  ret <32 x i8> %5
+}

diff --git a/llvm/test/CodeGen/X86/vector-trunc-ssat.ll b/llvm/test/CodeGen/X86/vector-trunc-ssat.ll
index 774a478a5d72df2a752365cf4ae9f34a4fbd323d..54f0d35eb7486c6baf72568e363de2e4c604bbc6 100644 (file)
--- a/llvm/test/CodeGen/X86/vector-trunc-ssat.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc-ssat.ll
@@ -3049,3 +3049,54 @@ define <32 x i8> @trunc_ssat_v32i16_v32i8(<32 x i16> %a0) {
   %5 = trunc <32 x i16> %4 to <32 x i8>
   ret <32 x i8> %5
 }
+
+define <32 x i8> @trunc_ssat_v32i32_v32i8(<32 x i32> %a0) {
+; SSE-LABEL: trunc_ssat_v32i32_v32i8:
+; SSE:       # %bb.0:
+; SSE-NEXT:    packssdw %xmm3, %xmm2
+; SSE-NEXT:    packssdw %xmm1, %xmm0
+; SSE-NEXT:    packsswb %xmm2, %xmm0
+; SSE-NEXT:    packssdw %xmm7, %xmm6
+; SSE-NEXT:    packssdw %xmm5, %xmm4
+; SSE-NEXT:    packsswb %xmm6, %xmm4
+; SSE-NEXT:    movdqa %xmm4, %xmm1
+; SSE-NEXT:    retq
+;
+; AVX1-LABEL: trunc_ssat_v32i32_v32i8:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vextractf128 $1, %ymm3, %xmm4
+; AVX1-NEXT:    vpackssdw %xmm4, %xmm3, %xmm3
+; AVX1-NEXT:    vextractf128 $1, %ymm2, %xmm4
+; AVX1-NEXT:    vpackssdw %xmm4, %xmm2, %xmm2
+; AVX1-NEXT:    vpacksswb %xmm3, %xmm2, %xmm2
+; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm3
+; AVX1-NEXT:    vpackssdw %xmm3, %xmm1, %xmm1
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm3
+; AVX1-NEXT:    vpackssdw %xmm3, %xmm0, %xmm0
+; AVX1-NEXT:    vpacksswb %xmm1, %xmm0, %xmm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: trunc_ssat_v32i32_v32i8:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vpackssdw %ymm3, %ymm2, %ymm2
+; AVX2-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[0,2,1,3]
+; AVX2-NEXT:    vpackssdw %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
+; AVX2-NEXT:    vpacksswb %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: trunc_ssat_v32i32_v32i8:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vpmovsdb %zmm0, %xmm0
+; AVX512-NEXT:    vpmovsdb %zmm1, %xmm1
+; AVX512-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX512-NEXT:    retq
+  %1 = icmp slt <32 x i32> %a0, <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127>
+  %2 = select <32 x i1> %1, <32 x i32> %a0, <32 x i32> <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127>
+  %3 = icmp sgt <32 x i32> %2, <i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128>
+  %4 = select <32 x i1> %3, <32 x i32> %2, <32 x i32> <i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128>
+  %5 = trunc <32 x i32> %4 to <32 x i8>
+  ret <32 x i8> %5
+}

diff --git a/llvm/test/CodeGen/X86/vector-trunc-usat.ll b/llvm/test/CodeGen/X86/vector-trunc-usat.ll
index 7489d393585ffb56a0a9d0db0cb447bd6e51b47c..2c443db99e134402ef5da4ba1d261f53514259ff 100644 (file)
--- a/llvm/test/CodeGen/X86/vector-trunc-usat.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc-usat.ll
@@ -2453,3 +2453,211 @@ define <32 x i8> @trunc_usat_v32i16_v32i8(<32 x i16> %a0) {
   %3 = trunc <32 x i16> %2 to <32 x i8>
   ret <32 x i8> %3
 }
+
+define <32 x i8> @trunc_usat_v32i32_v32i8(<32 x i32> %a0) {
+; SSE2-LABEL: trunc_usat_v32i32_v32i8:
+; SSE2:       # %bb.0:
+; SSE2-NEXT:    movdqa %xmm1, %xmm8
+; SSE2-NEXT:    movdqa {{.*#+}} xmm10 = [255,255,255,255]
+; SSE2-NEXT:    movdqa {{.*#+}} xmm11 = [2147483648,2147483648,2147483648,2147483648]
+; SSE2-NEXT:    movdqa %xmm5, %xmm1
+; SSE2-NEXT:    pxor %xmm11, %xmm1
+; SSE2-NEXT:    movdqa {{.*#+}} xmm9 = [2147483903,2147483903,2147483903,2147483903]
+; SSE2-NEXT:    movdqa %xmm9, %xmm12
+; SSE2-NEXT:    pcmpgtd %xmm1, %xmm12
+; SSE2-NEXT:    pand %xmm12, %xmm5
+; SSE2-NEXT:    pandn %xmm10, %xmm12
+; SSE2-NEXT:    por %xmm5, %xmm12
+; SSE2-NEXT:    movdqa %xmm4, %xmm5
+; SSE2-NEXT:    pxor %xmm11, %xmm5
+; SSE2-NEXT:    movdqa %xmm9, %xmm1
+; SSE2-NEXT:    pcmpgtd %xmm5, %xmm1
+; SSE2-NEXT:    pand %xmm1, %xmm4
+; SSE2-NEXT:    pandn %xmm10, %xmm1
+; SSE2-NEXT:    por %xmm4, %xmm1
+; SSE2-NEXT:    packuswb %xmm12, %xmm1
+; SSE2-NEXT:    movdqa %xmm7, %xmm4
+; SSE2-NEXT:    pxor %xmm11, %xmm4
+; SSE2-NEXT:    movdqa %xmm9, %xmm5
+; SSE2-NEXT:    pcmpgtd %xmm4, %xmm5
+; SSE2-NEXT:    pand %xmm5, %xmm7
+; SSE2-NEXT:    pandn %xmm10, %xmm5
+; SSE2-NEXT:    por %xmm7, %xmm5
+; SSE2-NEXT:    movdqa %xmm6, %xmm4
+; SSE2-NEXT:    pxor %xmm11, %xmm4
+; SSE2-NEXT:    movdqa %xmm9, %xmm7
+; SSE2-NEXT:    pcmpgtd %xmm4, %xmm7
+; SSE2-NEXT:    pand %xmm7, %xmm6
+; SSE2-NEXT:    pandn %xmm10, %xmm7
+; SSE2-NEXT:    por %xmm6, %xmm7
+; SSE2-NEXT:    packuswb %xmm5, %xmm7
+; SSE2-NEXT:    packuswb %xmm7, %xmm1
+; SSE2-NEXT:    movdqa %xmm8, %xmm4
+; SSE2-NEXT:    pxor %xmm11, %xmm4
+; SSE2-NEXT:    movdqa %xmm9, %xmm5
+; SSE2-NEXT:    pcmpgtd %xmm4, %xmm5
+; SSE2-NEXT:    pand %xmm5, %xmm8
+; SSE2-NEXT:    pandn %xmm10, %xmm5
+; SSE2-NEXT:    por %xmm8, %xmm5
+; SSE2-NEXT:    movdqa %xmm0, %xmm4
+; SSE2-NEXT:    pxor %xmm11, %xmm4
+; SSE2-NEXT:    movdqa %xmm9, %xmm6
+; SSE2-NEXT:    pcmpgtd %xmm4, %xmm6
+; SSE2-NEXT:    pand %xmm6, %xmm0
+; SSE2-NEXT:    pandn %xmm10, %xmm6
+; SSE2-NEXT:    por %xmm6, %xmm0
+; SSE2-NEXT:    packuswb %xmm5, %xmm0
+; SSE2-NEXT:    movdqa %xmm3, %xmm4
+; SSE2-NEXT:    pxor %xmm11, %xmm4
+; SSE2-NEXT:    movdqa %xmm9, %xmm5
+; SSE2-NEXT:    pcmpgtd %xmm4, %xmm5
+; SSE2-NEXT:    pand %xmm5, %xmm3
+; SSE2-NEXT:    pandn %xmm10, %xmm5
+; SSE2-NEXT:    por %xmm3, %xmm5
+; SSE2-NEXT:    pxor %xmm2, %xmm11
+; SSE2-NEXT:    pcmpgtd %xmm11, %xmm9
+; SSE2-NEXT:    pand %xmm9, %xmm2
+; SSE2-NEXT:    pandn %xmm10, %xmm9
+; SSE2-NEXT:    por %xmm2, %xmm9
+; SSE2-NEXT:    packuswb %xmm5, %xmm9
+; SSE2-NEXT:    packuswb %xmm9, %xmm0
+; SSE2-NEXT:    retq
+;
+; SSSE3-LABEL: trunc_usat_v32i32_v32i8:
+; SSSE3:       # %bb.0:
+; SSSE3-NEXT:    movdqa %xmm1, %xmm8
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm10 = [255,255,255,255]
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm11 = [2147483648,2147483648,2147483648,2147483648]
+; SSSE3-NEXT:    movdqa %xmm5, %xmm1
+; SSSE3-NEXT:    pxor %xmm11, %xmm1
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm9 = [2147483903,2147483903,2147483903,2147483903]
+; SSSE3-NEXT:    movdqa %xmm9, %xmm12
+; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm12
+; SSSE3-NEXT:    pand %xmm12, %xmm5
+; SSSE3-NEXT:    pandn %xmm10, %xmm12
+; SSSE3-NEXT:    por %xmm5, %xmm12
+; SSSE3-NEXT:    movdqa %xmm4, %xmm5
+; SSSE3-NEXT:    pxor %xmm11, %xmm5
+; SSSE3-NEXT:    movdqa %xmm9, %xmm1
+; SSSE3-NEXT:    pcmpgtd %xmm5, %xmm1
+; SSSE3-NEXT:    pand %xmm1, %xmm4
+; SSSE3-NEXT:    pandn %xmm10, %xmm1
+; SSSE3-NEXT:    por %xmm4, %xmm1
+; SSSE3-NEXT:    packuswb %xmm12, %xmm1
+; SSSE3-NEXT:    movdqa %xmm7, %xmm4
+; SSSE3-NEXT:    pxor %xmm11, %xmm4
+; SSSE3-NEXT:    movdqa %xmm9, %xmm5
+; SSSE3-NEXT:    pcmpgtd %xmm4, %xmm5
+; SSSE3-NEXT:    pand %xmm5, %xmm7
+; SSSE3-NEXT:    pandn %xmm10, %xmm5
+; SSSE3-NEXT:    por %xmm7, %xmm5
+; SSSE3-NEXT:    movdqa %xmm6, %xmm4
+; SSSE3-NEXT:    pxor %xmm11, %xmm4
+; SSSE3-NEXT:    movdqa %xmm9, %xmm7
+; SSSE3-NEXT:    pcmpgtd %xmm4, %xmm7
+; SSSE3-NEXT:    pand %xmm7, %xmm6
+; SSSE3-NEXT:    pandn %xmm10, %xmm7
+; SSSE3-NEXT:    por %xmm6, %xmm7
+; SSSE3-NEXT:    packuswb %xmm5, %xmm7
+; SSSE3-NEXT:    packuswb %xmm7, %xmm1
+; SSSE3-NEXT:    movdqa %xmm8, %xmm4
+; SSSE3-NEXT:    pxor %xmm11, %xmm4
+; SSSE3-NEXT:    movdqa %xmm9, %xmm5
+; SSSE3-NEXT:    pcmpgtd %xmm4, %xmm5
+; SSSE3-NEXT:    pand %xmm5, %xmm8
+; SSSE3-NEXT:    pandn %xmm10, %xmm5
+; SSSE3-NEXT:    por %xmm8, %xmm5
+; SSSE3-NEXT:    movdqa %xmm0, %xmm4
+; SSSE3-NEXT:    pxor %xmm11, %xmm4
+; SSSE3-NEXT:    movdqa %xmm9, %xmm6
+; SSSE3-NEXT:    pcmpgtd %xmm4, %xmm6
+; SSSE3-NEXT:    pand %xmm6, %xmm0
+; SSSE3-NEXT:    pandn %xmm10, %xmm6
+; SSSE3-NEXT:    por %xmm6, %xmm0
+; SSSE3-NEXT:    packuswb %xmm5, %xmm0
+; SSSE3-NEXT:    movdqa %xmm3, %xmm4
+; SSSE3-NEXT:    pxor %xmm11, %xmm4
+; SSSE3-NEXT:    movdqa %xmm9, %xmm5
+; SSSE3-NEXT:    pcmpgtd %xmm4, %xmm5
+; SSSE3-NEXT:    pand %xmm5, %xmm3
+; SSSE3-NEXT:    pandn %xmm10, %xmm5
+; SSSE3-NEXT:    por %xmm3, %xmm5
+; SSSE3-NEXT:    pxor %xmm2, %xmm11
+; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm9
+; SSSE3-NEXT:    pand %xmm9, %xmm2
+; SSSE3-NEXT:    pandn %xmm10, %xmm9
+; SSSE3-NEXT:    por %xmm2, %xmm9
+; SSSE3-NEXT:    packuswb %xmm5, %xmm9
+; SSSE3-NEXT:    packuswb %xmm9, %xmm0
+; SSSE3-NEXT:    retq
+;
+; SSE41-LABEL: trunc_usat_v32i32_v32i8:
+; SSE41:       # %bb.0:
+; SSE41-NEXT:    movdqa {{.*#+}} xmm8 = [255,255,255,255]
+; SSE41-NEXT:    pminud %xmm8, %xmm5
+; SSE41-NEXT:    pminud %xmm8, %xmm4
+; SSE41-NEXT:    packusdw %xmm5, %xmm4
+; SSE41-NEXT:    pminud %xmm8, %xmm7
+; SSE41-NEXT:    pminud %xmm8, %xmm6
+; SSE41-NEXT:    packusdw %xmm7, %xmm6
+; SSE41-NEXT:    packuswb %xmm6, %xmm4
+; SSE41-NEXT:    pminud %xmm8, %xmm1
+; SSE41-NEXT:    pminud %xmm8, %xmm0
+; SSE41-NEXT:    packusdw %xmm1, %xmm0
+; SSE41-NEXT:    pminud %xmm8, %xmm3
+; SSE41-NEXT:    pminud %xmm8, %xmm2
+; SSE41-NEXT:    packusdw %xmm3, %xmm2
+; SSE41-NEXT:    packuswb %xmm2, %xmm0
+; SSE41-NEXT:    movdqa %xmm4, %xmm1
+; SSE41-NEXT:    retq
+;
+; AVX1-LABEL: trunc_usat_v32i32_v32i8:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm4
+; AVX1-NEXT:    vmovdqa {{.*#+}} xmm5 = [255,255,255,255]
+; AVX1-NEXT:    vpminud %xmm5, %xmm4, %xmm4
+; AVX1-NEXT:    vpminud %xmm5, %xmm0, %xmm0
+; AVX1-NEXT:    vpackusdw %xmm4, %xmm0, %xmm0
+; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm4
+; AVX1-NEXT:    vpminud %xmm5, %xmm4, %xmm4
+; AVX1-NEXT:    vpminud %xmm5, %xmm1, %xmm1
+; AVX1-NEXT:    vpackusdw %xmm4, %xmm1, %xmm1
+; AVX1-NEXT:    vpackuswb %xmm1, %xmm0, %xmm0
+; AVX1-NEXT:    vextractf128 $1, %ymm2, %xmm1
+; AVX1-NEXT:    vpminud %xmm5, %xmm1, %xmm1
+; AVX1-NEXT:    vpminud %xmm5, %xmm2, %xmm2
+; AVX1-NEXT:    vpackusdw %xmm1, %xmm2, %xmm1
+; AVX1-NEXT:    vextractf128 $1, %ymm3, %xmm2
+; AVX1-NEXT:    vpminud %xmm5, %xmm2, %xmm2
+; AVX1-NEXT:    vpminud %xmm5, %xmm3, %xmm3
+; AVX1-NEXT:    vpackusdw %xmm2, %xmm3, %xmm2
+; AVX1-NEXT:    vpackuswb %xmm2, %xmm1, %xmm1
+; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: trunc_usat_v32i32_v32i8:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vpbroadcastd {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255]
+; AVX2-NEXT:    vpminud %ymm4, %ymm1, %ymm1
+; AVX2-NEXT:    vpminud %ymm4, %ymm0, %ymm0
+; AVX2-NEXT:    vpackusdw %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpminud %ymm4, %ymm3, %ymm1
+; AVX2-NEXT:    vpminud %ymm4, %ymm2, %ymm2
+; AVX2-NEXT:    vpackusdw %ymm1, %ymm2, %ymm1
+; AVX2-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,2,1,3]
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
+; AVX2-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: trunc_usat_v32i32_v32i8:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vpmovusdb %zmm0, %xmm0
+; AVX512-NEXT:    vpmovusdb %zmm1, %xmm1
+; AVX512-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm0
+; AVX512-NEXT:    retq
+  %1 = icmp ult <32 x i32> %a0, <i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255>
+  %2 = select <32 x i1> %1, <32 x i32> %a0, <32 x i32> <i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255, i32 255>
+  %3 = trunc <32 x i32> %2 to <32 x i8>
+  ret <32 x i8> %3
+}