diff options
Diffstat (limited to 'llvm/test/CodeGen/X86/vector-trunc-ssat-widen.ll')
| -rw-r--r-- | llvm/test/CodeGen/X86/vector-trunc-ssat-widen.ll | 3050 |
1 files changed, 3050 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/X86/vector-trunc-ssat-widen.ll b/llvm/test/CodeGen/X86/vector-trunc-ssat-widen.ll new file mode 100644 index 00000000000..4d32267b61e --- /dev/null +++ b/llvm/test/CodeGen/X86/vector-trunc-ssat-widen.ll @@ -0,0 +1,3050 @@ +; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=SSE --check-prefix=SSE2 +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+ssse3 | FileCheck %s --check-prefix=SSE --check-prefix=SSSE3 +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=SSE --check-prefix=SSE41 +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=AVX --check-prefix=AVX1 +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2,AVX2-SLOW +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx2,+fast-variable-shuffle | FileCheck %s --check-prefixes=AVX,AVX2,AVX2-FAST +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=AVX512 --check-prefix=AVX512F +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+fast-variable-shuffle | FileCheck %s --check-prefixes=AVX512,AVX512VL +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512bw,+fast-variable-shuffle | FileCheck %s --check-prefixes=AVX512,AVX512BW +; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512bw,+avx512vl,+fast-variable-shuffle | FileCheck %s --check-prefixes=AVX512,AVX512BWVL + +; +; Signed saturation truncation to vXi32 +; + +define <4 x i32> @trunc_ssat_v4i64_v4i32(<4 x i64> %a0) { +; SSE2-LABEL: trunc_ssat_v4i64_v4i32: +; SSE2: # %bb.0: +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [2147483647,2147483647] +; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648] +; SSE2-NEXT: movdqa %xmm0, %xmm3 +; SSE2-NEXT: pxor %xmm2, %xmm3 +; SSE2-NEXT: movdqa {{.*#+}} xmm5 = [4294967295,4294967295] +; SSE2-NEXT: movdqa %xmm5, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm3, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm5, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm3[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm4, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm0 +; SSE2-NEXT: pandn %xmm8, %xmm3 +; SSE2-NEXT: por %xmm0, %xmm3 +; SSE2-NEXT: movdqa %xmm1, %xmm0 +; SSE2-NEXT: pxor %xmm2, %xmm0 +; SSE2-NEXT: movdqa %xmm5, %xmm4 +; SSE2-NEXT: pcmpgtd %xmm0, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm5, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSE2-NEXT: por %xmm0, %xmm4 +; SSE2-NEXT: pand %xmm4, %xmm1 +; SSE2-NEXT: pandn %xmm8, %xmm4 +; SSE2-NEXT: por %xmm1, %xmm4 +; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [18446744071562067968,18446744071562067968] +; SSE2-NEXT: movdqa %xmm4, %xmm0 +; SSE2-NEXT: pxor %xmm2, %xmm0 +; SSE2-NEXT: movdqa {{.*#+}} xmm5 = [18446744069414584320,18446744069414584320] +; SSE2-NEXT: movdqa %xmm0, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm5, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm5, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm0, %xmm6 +; SSE2-NEXT: pand %xmm6, %xmm4 +; SSE2-NEXT: pandn %xmm1, %xmm6 +; SSE2-NEXT: por %xmm4, %xmm6 +; SSE2-NEXT: pxor %xmm3, %xmm2 +; SSE2-NEXT: movdqa %xmm2, %xmm0 +; SSE2-NEXT: pcmpgtd %xmm5, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm0[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm5, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: pand %xmm4, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSE2-NEXT: por %xmm2, %xmm0 +; SSE2-NEXT: pand %xmm0, %xmm3 +; SSE2-NEXT: pandn %xmm1, %xmm0 +; SSE2-NEXT: por %xmm3, %xmm0 +; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm6[0,2] +; SSE2-NEXT: retq +; +; SSSE3-LABEL: trunc_ssat_v4i64_v4i32: +; SSSE3: # %bb.0: +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [2147483647,2147483647] +; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648] +; SSSE3-NEXT: movdqa %xmm0, %xmm3 +; SSSE3-NEXT: pxor %xmm2, %xmm3 +; SSSE3-NEXT: movdqa {{.*#+}} xmm5 = [4294967295,4294967295] +; SSSE3-NEXT: movdqa %xmm5, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm3, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm5, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm3[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm4, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm0 +; SSSE3-NEXT: pandn %xmm8, %xmm3 +; SSSE3-NEXT: por %xmm0, %xmm3 +; SSSE3-NEXT: movdqa %xmm1, %xmm0 +; SSSE3-NEXT: pxor %xmm2, %xmm0 +; SSSE3-NEXT: movdqa %xmm5, %xmm4 +; SSSE3-NEXT: pcmpgtd %xmm0, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm5, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSSE3-NEXT: por %xmm0, %xmm4 +; SSSE3-NEXT: pand %xmm4, %xmm1 +; SSSE3-NEXT: pandn %xmm8, %xmm4 +; SSSE3-NEXT: por %xmm1, %xmm4 +; SSSE3-NEXT: movdqa {{.*#+}} xmm1 = [18446744071562067968,18446744071562067968] +; SSSE3-NEXT: movdqa %xmm4, %xmm0 +; SSSE3-NEXT: pxor %xmm2, %xmm0 +; SSSE3-NEXT: movdqa {{.*#+}} xmm5 = [18446744069414584320,18446744069414584320] +; SSSE3-NEXT: movdqa %xmm0, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm5, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm5, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm0, %xmm6 +; SSSE3-NEXT: pand %xmm6, %xmm4 +; SSSE3-NEXT: pandn %xmm1, %xmm6 +; SSSE3-NEXT: por %xmm4, %xmm6 +; SSSE3-NEXT: pxor %xmm3, %xmm2 +; SSSE3-NEXT: movdqa %xmm2, %xmm0 +; SSSE3-NEXT: pcmpgtd %xmm5, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm0[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm5, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: pand %xmm4, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSSE3-NEXT: por %xmm2, %xmm0 +; SSSE3-NEXT: pand %xmm0, %xmm3 +; SSSE3-NEXT: pandn %xmm1, %xmm0 +; SSSE3-NEXT: por %xmm3, %xmm0 +; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm6[0,2] +; SSSE3-NEXT: retq +; +; SSE41-LABEL: trunc_ssat_v4i64_v4i32: +; SSE41: # %bb.0: +; SSE41-NEXT: movdqa %xmm0, %xmm2 +; SSE41-NEXT: movapd {{.*#+}} xmm4 = [2147483647,2147483647] +; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [2147483648,2147483648] +; SSE41-NEXT: pxor %xmm3, %xmm0 +; SSE41-NEXT: movdqa {{.*#+}} xmm6 = [4294967295,4294967295] +; SSE41-NEXT: movdqa %xmm6, %xmm5 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm5 +; SSE41-NEXT: movdqa %xmm6, %xmm7 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm7 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm7[0,0,2,2] +; SSE41-NEXT: pand %xmm5, %xmm0 +; SSE41-NEXT: por %xmm7, %xmm0 +; SSE41-NEXT: movapd %xmm4, %xmm5 +; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm5 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: pxor %xmm3, %xmm0 +; SSE41-NEXT: movdqa %xmm6, %xmm2 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm2 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm2, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm4 +; SSE41-NEXT: movapd {{.*#+}} xmm1 = [18446744071562067968,18446744071562067968] +; SSE41-NEXT: movapd %xmm4, %xmm2 +; SSE41-NEXT: xorpd %xmm3, %xmm2 +; SSE41-NEXT: movdqa {{.*#+}} xmm6 = [18446744069414584320,18446744069414584320] +; SSE41-NEXT: movapd %xmm2, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm6, %xmm7 +; SSE41-NEXT: pcmpgtd %xmm6, %xmm2 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm2, %xmm0 +; SSE41-NEXT: movapd %xmm1, %xmm2 +; SSE41-NEXT: blendvpd %xmm0, %xmm4, %xmm2 +; SSE41-NEXT: xorpd %xmm5, %xmm3 +; SSE41-NEXT: movapd %xmm3, %xmm4 +; SSE41-NEXT: pcmpeqd %xmm6, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm6, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm4, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm5, %xmm1 +; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2] +; SSE41-NEXT: movaps %xmm1, %xmm0 +; SSE41-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v4i64_v4i32: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [2147483647,2147483647] +; AVX1-NEXT: vpcmpgtq %xmm1, %xmm2, %xmm3 +; AVX1-NEXT: vpcmpgtq %xmm0, %xmm2, %xmm4 +; AVX1-NEXT: vblendvpd %xmm4, %xmm0, %xmm2, %xmm0 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [18446744071562067968,18446744071562067968] +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm0, %xmm5 +; AVX1-NEXT: vblendvpd %xmm3, %xmm1, %xmm2, %xmm1 +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm1, %xmm2 +; AVX1-NEXT: vblendvpd %xmm2, %xmm1, %xmm4, %xmm1 +; AVX1-NEXT: vblendvpd %xmm5, %xmm0, %xmm4, %xmm0 +; AVX1-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2] +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-SLOW-LABEL: trunc_ssat_v4i64_v4i32: +; AVX2-SLOW: # %bb.0: +; AVX2-SLOW-NEXT: vpbroadcastq {{.*#+}} ymm1 = [2147483647,2147483647,2147483647,2147483647] +; AVX2-SLOW-NEXT: vpcmpgtq %ymm0, %ymm1, %ymm2 +; AVX2-SLOW-NEXT: vblendvpd %ymm2, %ymm0, %ymm1, %ymm0 +; AVX2-SLOW-NEXT: vpbroadcastq {{.*#+}} ymm1 = [18446744071562067968,18446744071562067968,18446744071562067968,18446744071562067968] +; AVX2-SLOW-NEXT: vpcmpgtq %ymm1, %ymm0, %ymm2 +; AVX2-SLOW-NEXT: vblendvpd %ymm2, %ymm0, %ymm1, %ymm0 +; AVX2-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2] +; AVX2-SLOW-NEXT: vzeroupper +; AVX2-SLOW-NEXT: retq +; +; AVX2-FAST-LABEL: trunc_ssat_v4i64_v4i32: +; AVX2-FAST: # %bb.0: +; AVX2-FAST-NEXT: vpbroadcastq {{.*#+}} ymm1 = [2147483647,2147483647,2147483647,2147483647] +; AVX2-FAST-NEXT: vpcmpgtq %ymm0, %ymm1, %ymm2 +; AVX2-FAST-NEXT: vblendvpd %ymm2, %ymm0, %ymm1, %ymm0 +; AVX2-FAST-NEXT: vpbroadcastq {{.*#+}} ymm1 = [18446744071562067968,18446744071562067968,18446744071562067968,18446744071562067968] +; AVX2-FAST-NEXT: vpcmpgtq %ymm1, %ymm0, %ymm2 +; AVX2-FAST-NEXT: vblendvpd %ymm2, %ymm0, %ymm1, %ymm0 +; AVX2-FAST-NEXT: vmovapd {{.*#+}} ymm1 = [0,2,4,6,4,6,6,7] +; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0 +; AVX2-FAST-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0 +; AVX2-FAST-NEXT: vzeroupper +; AVX2-FAST-NEXT: retq +; +; AVX512F-LABEL: trunc_ssat_v4i64_v4i32: +; AVX512F: # %bb.0: +; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0 +; AVX512F-NEXT: vpbroadcastq {{.*#+}} ymm1 = [2147483647,2147483647,2147483647,2147483647] +; AVX512F-NEXT: vpminsq %zmm1, %zmm0, %zmm0 +; AVX512F-NEXT: vpbroadcastq {{.*#+}} ymm1 = [18446744071562067968,18446744071562067968,18446744071562067968,18446744071562067968] +; AVX512F-NEXT: vpmaxsq %zmm1, %zmm0, %zmm0 +; AVX512F-NEXT: vpmovqd %zmm0, %ymm0 +; AVX512F-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0 +; AVX512F-NEXT: vzeroupper +; AVX512F-NEXT: retq +; +; AVX512VL-LABEL: trunc_ssat_v4i64_v4i32: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsqd %ymm0, %xmm0 +; AVX512VL-NEXT: vzeroupper +; AVX512VL-NEXT: retq +; +; AVX512BW-LABEL: trunc_ssat_v4i64_v4i32: +; AVX512BW: # %bb.0: +; AVX512BW-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0 +; AVX512BW-NEXT: vpbroadcastq {{.*#+}} ymm1 = [2147483647,2147483647,2147483647,2147483647] +; AVX512BW-NEXT: vpminsq %zmm1, %zmm0, %zmm0 +; AVX512BW-NEXT: vpbroadcastq {{.*#+}} ymm1 = [18446744071562067968,18446744071562067968,18446744071562067968,18446744071562067968] +; AVX512BW-NEXT: vpmaxsq %zmm1, %zmm0, %zmm0 +; AVX512BW-NEXT: vpmovqd %zmm0, %ymm0 +; AVX512BW-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0 +; AVX512BW-NEXT: vzeroupper +; AVX512BW-NEXT: retq +; +; AVX512BWVL-LABEL: trunc_ssat_v4i64_v4i32: +; AVX512BWVL: # %bb.0: +; AVX512BWVL-NEXT: vpmovsqd %ymm0, %xmm0 +; AVX512BWVL-NEXT: vzeroupper +; AVX512BWVL-NEXT: retq + %1 = icmp slt <4 x i64> %a0, <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647> + %2 = select <4 x i1> %1, <4 x i64> %a0, <4 x i64> <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647> + %3 = icmp sgt <4 x i64> %2, <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648> + %4 = select <4 x i1> %3, <4 x i64> %2, <4 x i64> <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648> + %5 = trunc <4 x i64> %4 to <4 x i32> + ret <4 x i32> %5 +} + + +define <8 x i32> @trunc_ssat_v8i64_v8i32(<8 x i64> %a0) { +; SSE2-LABEL: trunc_ssat_v8i64_v8i32: +; SSE2: # %bb.0: +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [2147483647,2147483647] +; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648] +; SSE2-NEXT: movdqa %xmm0, %xmm5 +; SSE2-NEXT: pxor %xmm4, %xmm5 +; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [4294967295,4294967295] +; SSE2-NEXT: movdqa %xmm9, %xmm7 +; SSE2-NEXT: pcmpgtd %xmm5, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm5 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm5 +; SSE2-NEXT: pand %xmm5, %xmm0 +; SSE2-NEXT: pandn %xmm8, %xmm5 +; SSE2-NEXT: por %xmm0, %xmm5 +; SSE2-NEXT: movdqa %xmm1, %xmm0 +; SSE2-NEXT: pxor %xmm4, %xmm0 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm0 +; SSE2-NEXT: pand %xmm0, %xmm1 +; SSE2-NEXT: pandn %xmm8, %xmm0 +; SSE2-NEXT: por %xmm1, %xmm0 +; SSE2-NEXT: movdqa %xmm2, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm1, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm7 +; SSE2-NEXT: pand %xmm7, %xmm2 +; SSE2-NEXT: pandn %xmm8, %xmm7 +; SSE2-NEXT: por %xmm2, %xmm7 +; SSE2-NEXT: movdqa %xmm3, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm9, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm1, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm6 +; SSE2-NEXT: pand %xmm6, %xmm3 +; SSE2-NEXT: pandn %xmm8, %xmm6 +; SSE2-NEXT: por %xmm3, %xmm6 +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [18446744071562067968,18446744071562067968] +; SSE2-NEXT: movdqa %xmm6, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [18446744069414584320,18446744069414584320] +; SSE2-NEXT: movdqa %xmm1, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm3, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm6 +; SSE2-NEXT: pandn %xmm8, %xmm2 +; SSE2-NEXT: por %xmm6, %xmm2 +; SSE2-NEXT: movdqa %xmm7, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm3 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm3[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm3[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm1 +; SSE2-NEXT: pand %xmm1, %xmm7 +; SSE2-NEXT: pandn %xmm8, %xmm1 +; SSE2-NEXT: por %xmm7, %xmm1 +; SSE2-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2] +; SSE2-NEXT: movdqa %xmm0, %xmm2 +; SSE2-NEXT: pxor %xmm4, %xmm2 +; SSE2-NEXT: movdqa %xmm2, %xmm3 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm3[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSE2-NEXT: por %xmm2, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm0 +; SSE2-NEXT: pandn %xmm8, %xmm3 +; SSE2-NEXT: por %xmm0, %xmm3 +; SSE2-NEXT: pxor %xmm5, %xmm4 +; SSE2-NEXT: movdqa %xmm4, %xmm0 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSE2-NEXT: pand %xmm2, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSE2-NEXT: por %xmm4, %xmm0 +; SSE2-NEXT: pand %xmm0, %xmm5 +; SSE2-NEXT: pandn %xmm8, %xmm0 +; SSE2-NEXT: por %xmm5, %xmm0 +; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm3[0,2] +; SSE2-NEXT: retq +; +; SSSE3-LABEL: trunc_ssat_v8i64_v8i32: +; SSSE3: # %bb.0: +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [2147483647,2147483647] +; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648] +; SSSE3-NEXT: movdqa %xmm0, %xmm5 +; SSSE3-NEXT: pxor %xmm4, %xmm5 +; SSSE3-NEXT: movdqa {{.*#+}} xmm9 = [4294967295,4294967295] +; SSSE3-NEXT: movdqa %xmm9, %xmm7 +; SSSE3-NEXT: pcmpgtd %xmm5, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm5 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm5 +; SSSE3-NEXT: pand %xmm5, %xmm0 +; SSSE3-NEXT: pandn %xmm8, %xmm5 +; SSSE3-NEXT: por %xmm0, %xmm5 +; SSSE3-NEXT: movdqa %xmm1, %xmm0 +; SSSE3-NEXT: pxor %xmm4, %xmm0 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm0, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm0 +; SSSE3-NEXT: pand %xmm0, %xmm1 +; SSSE3-NEXT: pandn %xmm8, %xmm0 +; SSSE3-NEXT: por %xmm1, %xmm0 +; SSSE3-NEXT: movdqa %xmm2, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm1, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm7 +; SSSE3-NEXT: pand %xmm7, %xmm2 +; SSSE3-NEXT: pandn %xmm8, %xmm7 +; SSSE3-NEXT: por %xmm2, %xmm7 +; SSSE3-NEXT: movdqa %xmm3, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm9, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm1, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm6 +; SSSE3-NEXT: pand %xmm6, %xmm3 +; SSSE3-NEXT: pandn %xmm8, %xmm6 +; SSSE3-NEXT: por %xmm3, %xmm6 +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [18446744071562067968,18446744071562067968] +; SSSE3-NEXT: movdqa %xmm6, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa {{.*#+}} xmm9 = [18446744069414584320,18446744069414584320] +; SSSE3-NEXT: movdqa %xmm1, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm3, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm6 +; SSSE3-NEXT: pandn %xmm8, %xmm2 +; SSSE3-NEXT: por %xmm6, %xmm2 +; SSSE3-NEXT: movdqa %xmm7, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm3 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm3[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm3[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm1 +; SSSE3-NEXT: pand %xmm1, %xmm7 +; SSSE3-NEXT: pandn %xmm8, %xmm1 +; SSSE3-NEXT: por %xmm7, %xmm1 +; SSSE3-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2] +; SSSE3-NEXT: movdqa %xmm0, %xmm2 +; SSSE3-NEXT: pxor %xmm4, %xmm2 +; SSSE3-NEXT: movdqa %xmm2, %xmm3 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm3[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSSE3-NEXT: por %xmm2, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm0 +; SSSE3-NEXT: pandn %xmm8, %xmm3 +; SSSE3-NEXT: por %xmm0, %xmm3 +; SSSE3-NEXT: pxor %xmm5, %xmm4 +; SSSE3-NEXT: movdqa %xmm4, %xmm0 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm0[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSSE3-NEXT: pand %xmm2, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSSE3-NEXT: por %xmm4, %xmm0 +; SSSE3-NEXT: pand %xmm0, %xmm5 +; SSSE3-NEXT: pandn %xmm8, %xmm0 +; SSSE3-NEXT: por %xmm5, %xmm0 +; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm3[0,2] +; SSSE3-NEXT: retq +; +; SSE41-LABEL: trunc_ssat_v8i64_v8i32: +; SSE41: # %bb.0: +; SSE41-NEXT: movdqa %xmm0, %xmm9 +; SSE41-NEXT: movapd {{.*#+}} xmm10 = [2147483647,2147483647] +; SSE41-NEXT: movdqa {{.*#+}} xmm5 = [2147483648,2147483648] +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [4294967295,4294967295] +; SSE41-NEXT: movdqa %xmm4, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm7 +; SSE41-NEXT: movdqa %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm10, %xmm8 +; SSE41-NEXT: blendvpd %xmm0, %xmm9, %xmm8 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm4, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm6 +; SSE41-NEXT: movdqa %xmm4, %xmm7 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm7 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm7[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm7, %xmm0 +; SSE41-NEXT: movapd %xmm10, %xmm9 +; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm9 +; SSE41-NEXT: movdqa %xmm2, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm4, %xmm1 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm1 +; SSE41-NEXT: movdqa %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm1, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm10, %xmm6 +; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm6 +; SSE41-NEXT: movdqa %xmm3, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm4, %xmm1 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm1 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm1, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm3, %xmm10 +; SSE41-NEXT: movapd {{.*#+}} xmm2 = [18446744071562067968,18446744071562067968] +; SSE41-NEXT: movapd %xmm10, %xmm1 +; SSE41-NEXT: xorpd %xmm5, %xmm1 +; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [18446744069414584320,18446744069414584320] +; SSE41-NEXT: movapd %xmm1, %xmm4 +; SSE41-NEXT: pcmpeqd %xmm3, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm3, %xmm1 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,2,2] +; SSE41-NEXT: pand %xmm4, %xmm0 +; SSE41-NEXT: por %xmm1, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm4 +; SSE41-NEXT: blendvpd %xmm0, %xmm10, %xmm4 +; SSE41-NEXT: movapd %xmm6, %xmm1 +; SSE41-NEXT: xorpd %xmm5, %xmm1 +; SSE41-NEXT: movapd %xmm1, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm3, %xmm7 +; SSE41-NEXT: pcmpgtd %xmm3, %xmm1 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm1, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm1 +; SSE41-NEXT: blendvpd %xmm0, %xmm6, %xmm1 +; SSE41-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,2],xmm4[0,2] +; SSE41-NEXT: movapd %xmm9, %xmm4 +; SSE41-NEXT: xorpd %xmm5, %xmm4 +; SSE41-NEXT: movapd %xmm4, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm3, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm3, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm4 +; SSE41-NEXT: blendvpd %xmm0, %xmm9, %xmm4 +; SSE41-NEXT: xorpd %xmm8, %xmm5 +; SSE41-NEXT: movapd %xmm5, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm3, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm3, %xmm5 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm5[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm5, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm8, %xmm2 +; SSE41-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm4[0,2] +; SSE41-NEXT: movaps %xmm2, %xmm0 +; SSE41-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v8i64_v8i32: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [2147483647,2147483647] +; AVX1-NEXT: vpcmpgtq %xmm2, %xmm3, %xmm8 +; AVX1-NEXT: vpcmpgtq %xmm1, %xmm3, %xmm5 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm6 +; AVX1-NEXT: vpcmpgtq %xmm6, %xmm3, %xmm7 +; AVX1-NEXT: vpcmpgtq %xmm0, %xmm3, %xmm4 +; AVX1-NEXT: vblendvpd %xmm4, %xmm0, %xmm3, %xmm0 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [18446744071562067968,18446744071562067968] +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm0, %xmm9 +; AVX1-NEXT: vblendvpd %xmm7, %xmm6, %xmm3, %xmm6 +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm6, %xmm7 +; AVX1-NEXT: vblendvpd %xmm5, %xmm1, %xmm3, %xmm1 +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm1, %xmm5 +; AVX1-NEXT: vblendvpd %xmm8, %xmm2, %xmm3, %xmm2 +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm2, %xmm3 +; AVX1-NEXT: vblendvpd %xmm3, %xmm2, %xmm4, %xmm2 +; AVX1-NEXT: vblendvpd %xmm5, %xmm1, %xmm4, %xmm1 +; AVX1-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2] +; AVX1-NEXT: vblendvpd %xmm7, %xmm6, %xmm4, %xmm2 +; AVX1-NEXT: vblendvpd %xmm9, %xmm0, %xmm4, %xmm0 +; AVX1-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2] +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-SLOW-LABEL: trunc_ssat_v8i64_v8i32: +; AVX2-SLOW: # %bb.0: +; AVX2-SLOW-NEXT: vpbroadcastq {{.*#+}} ymm2 = [2147483647,2147483647,2147483647,2147483647] +; AVX2-SLOW-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3 +; AVX2-SLOW-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-SLOW-NEXT: vpcmpgtq %ymm0, %ymm2, %ymm3 +; AVX2-SLOW-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-SLOW-NEXT: vpbroadcastq {{.*#+}} ymm2 = [18446744071562067968,18446744071562067968,18446744071562067968,18446744071562067968] +; AVX2-SLOW-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm3 +; AVX2-SLOW-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-SLOW-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm3 +; AVX2-SLOW-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2] +; AVX2-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm2 +; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2] +; AVX2-SLOW-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX2-SLOW-NEXT: retq +; +; AVX2-FAST-LABEL: trunc_ssat_v8i64_v8i32: +; AVX2-FAST: # %bb.0: +; AVX2-FAST-NEXT: vpbroadcastq {{.*#+}} ymm2 = [2147483647,2147483647,2147483647,2147483647] +; AVX2-FAST-NEXT: vpcmpgtq %ymm0, %ymm2, %ymm3 +; AVX2-FAST-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-FAST-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3 +; AVX2-FAST-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-FAST-NEXT: vpbroadcastq {{.*#+}} ymm2 = [18446744071562067968,18446744071562067968,18446744071562067968,18446744071562067968] +; AVX2-FAST-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm3 +; AVX2-FAST-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-FAST-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm3 +; AVX2-FAST-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-FAST-NEXT: vmovapd {{.*#+}} ymm2 = [0,2,4,6,4,6,6,7] +; AVX2-FAST-NEXT: vpermps %ymm0, %ymm2, %ymm0 +; AVX2-FAST-NEXT: vpermps %ymm1, %ymm2, %ymm1 +; AVX2-FAST-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX2-FAST-NEXT: retq +; +; AVX512-LABEL: trunc_ssat_v8i64_v8i32: +; AVX512: # %bb.0: +; AVX512-NEXT: vpmovsqd %zmm0, %ymm0 +; AVX512-NEXT: retq + %1 = icmp slt <8 x i64> %a0, <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647> + %2 = select <8 x i1> %1, <8 x i64> %a0, <8 x i64> <i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647, i64 2147483647> + %3 = icmp sgt <8 x i64> %2, <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648> + %4 = select <8 x i1> %3, <8 x i64> %2, <8 x i64> <i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648, i64 -2147483648> + %5 = trunc <8 x i64> %4 to <8 x i32> + ret <8 x i32> %5 +} + +; +; Signed saturation truncation to vXi16 +; + +define <8 x i16> @trunc_ssat_v8i64_v8i16(<8 x i64> %a0) { +; SSE2-LABEL: trunc_ssat_v8i64_v8i16: +; SSE2: # %bb.0: +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [32767,32767] +; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648] +; SSE2-NEXT: movdqa %xmm2, %xmm5 +; SSE2-NEXT: pxor %xmm4, %xmm5 +; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [2147516415,2147516415] +; SSE2-NEXT: movdqa %xmm9, %xmm7 +; SSE2-NEXT: pcmpgtd %xmm5, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm5 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm5 +; SSE2-NEXT: pand %xmm5, %xmm2 +; SSE2-NEXT: pandn %xmm8, %xmm5 +; SSE2-NEXT: por %xmm2, %xmm5 +; SSE2-NEXT: movdqa %xmm3, %xmm2 +; SSE2-NEXT: pxor %xmm4, %xmm2 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm2, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm2[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm3 +; SSE2-NEXT: pandn %xmm8, %xmm2 +; SSE2-NEXT: por %xmm3, %xmm2 +; SSE2-NEXT: movdqa %xmm0, %xmm3 +; SSE2-NEXT: pxor %xmm4, %xmm3 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm3, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm0 +; SSE2-NEXT: pandn %xmm8, %xmm3 +; SSE2-NEXT: por %xmm0, %xmm3 +; SSE2-NEXT: movdqa %xmm1, %xmm0 +; SSE2-NEXT: pxor %xmm4, %xmm0 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm0, %xmm7 +; SSE2-NEXT: pand %xmm7, %xmm1 +; SSE2-NEXT: pandn %xmm8, %xmm7 +; SSE2-NEXT: por %xmm1, %xmm7 +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [18446744073709518848,18446744073709518848] +; SSE2-NEXT: movdqa %xmm7, %xmm0 +; SSE2-NEXT: pxor %xmm4, %xmm0 +; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [18446744071562035200,18446744071562035200] +; SSE2-NEXT: movdqa %xmm0, %xmm1 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm1[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: por %xmm0, %xmm1 +; SSE2-NEXT: pand %xmm1, %xmm7 +; SSE2-NEXT: pandn %xmm8, %xmm1 +; SSE2-NEXT: por %xmm7, %xmm1 +; SSE2-NEXT: movdqa %xmm3, %xmm0 +; SSE2-NEXT: pxor %xmm4, %xmm0 +; SSE2-NEXT: movdqa %xmm0, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm0 +; SSE2-NEXT: pand %xmm0, %xmm3 +; SSE2-NEXT: pandn %xmm8, %xmm0 +; SSE2-NEXT: por %xmm3, %xmm0 +; SSE2-NEXT: packssdw %xmm1, %xmm0 +; SSE2-NEXT: movdqa %xmm2, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm3 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm3[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm2 +; SSE2-NEXT: pandn %xmm8, %xmm3 +; SSE2-NEXT: por %xmm2, %xmm3 +; SSE2-NEXT: pxor %xmm5, %xmm4 +; SSE2-NEXT: movdqa %xmm4, %xmm1 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSE2-NEXT: pand %xmm2, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: por %xmm4, %xmm1 +; SSE2-NEXT: pand %xmm1, %xmm5 +; SSE2-NEXT: pandn %xmm8, %xmm1 +; SSE2-NEXT: por %xmm5, %xmm1 +; SSE2-NEXT: packssdw %xmm3, %xmm1 +; SSE2-NEXT: packssdw %xmm1, %xmm0 +; SSE2-NEXT: retq +; +; SSSE3-LABEL: trunc_ssat_v8i64_v8i16: +; SSSE3: # %bb.0: +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [32767,32767] +; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648] +; SSSE3-NEXT: movdqa %xmm2, %xmm5 +; SSSE3-NEXT: pxor %xmm4, %xmm5 +; SSSE3-NEXT: movdqa {{.*#+}} xmm9 = [2147516415,2147516415] +; SSSE3-NEXT: movdqa %xmm9, %xmm7 +; SSSE3-NEXT: pcmpgtd %xmm5, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm5 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm5 +; SSSE3-NEXT: pand %xmm5, %xmm2 +; SSSE3-NEXT: pandn %xmm8, %xmm5 +; SSSE3-NEXT: por %xmm2, %xmm5 +; SSSE3-NEXT: movdqa %xmm3, %xmm2 +; SSSE3-NEXT: pxor %xmm4, %xmm2 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm2, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm2[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm3 +; SSSE3-NEXT: pandn %xmm8, %xmm2 +; SSSE3-NEXT: por %xmm3, %xmm2 +; SSSE3-NEXT: movdqa %xmm0, %xmm3 +; SSSE3-NEXT: pxor %xmm4, %xmm3 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm3, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm0 +; SSSE3-NEXT: pandn %xmm8, %xmm3 +; SSSE3-NEXT: por %xmm0, %xmm3 +; SSSE3-NEXT: movdqa %xmm1, %xmm0 +; SSSE3-NEXT: pxor %xmm4, %xmm0 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm0, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm0, %xmm7 +; SSSE3-NEXT: pand %xmm7, %xmm1 +; SSSE3-NEXT: pandn %xmm8, %xmm7 +; SSSE3-NEXT: por %xmm1, %xmm7 +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [18446744073709518848,18446744073709518848] +; SSSE3-NEXT: movdqa %xmm7, %xmm0 +; SSSE3-NEXT: pxor %xmm4, %xmm0 +; SSSE3-NEXT: movdqa {{.*#+}} xmm9 = [18446744071562035200,18446744071562035200] +; SSSE3-NEXT: movdqa %xmm0, %xmm1 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm1[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: por %xmm0, %xmm1 +; SSSE3-NEXT: pand %xmm1, %xmm7 +; SSSE3-NEXT: pandn %xmm8, %xmm1 +; SSSE3-NEXT: por %xmm7, %xmm1 +; SSSE3-NEXT: movdqa %xmm3, %xmm0 +; SSSE3-NEXT: pxor %xmm4, %xmm0 +; SSSE3-NEXT: movdqa %xmm0, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm0 +; SSSE3-NEXT: pand %xmm0, %xmm3 +; SSSE3-NEXT: pandn %xmm8, %xmm0 +; SSSE3-NEXT: por %xmm3, %xmm0 +; SSSE3-NEXT: packssdw %xmm1, %xmm0 +; SSSE3-NEXT: movdqa %xmm2, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm3 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm3[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm2 +; SSSE3-NEXT: pandn %xmm8, %xmm3 +; SSSE3-NEXT: por %xmm2, %xmm3 +; SSSE3-NEXT: pxor %xmm5, %xmm4 +; SSSE3-NEXT: movdqa %xmm4, %xmm1 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSSE3-NEXT: pand %xmm2, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: por %xmm4, %xmm1 +; SSSE3-NEXT: pand %xmm1, %xmm5 +; SSSE3-NEXT: pandn %xmm8, %xmm1 +; SSSE3-NEXT: por %xmm5, %xmm1 +; SSSE3-NEXT: packssdw %xmm3, %xmm1 +; SSSE3-NEXT: packssdw %xmm1, %xmm0 +; SSSE3-NEXT: retq +; +; SSE41-LABEL: trunc_ssat_v8i64_v8i16: +; SSE41: # %bb.0: +; SSE41-NEXT: movdqa %xmm0, %xmm10 +; SSE41-NEXT: movapd {{.*#+}} xmm11 = [32767,32767] +; SSE41-NEXT: movdqa {{.*#+}} xmm5 = [2147483648,2147483648] +; SSE41-NEXT: movdqa %xmm2, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [2147516415,2147516415] +; SSE41-NEXT: movdqa %xmm4, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm7 +; SSE41-NEXT: movdqa %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm8 +; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm8 +; SSE41-NEXT: movdqa %xmm3, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm4, %xmm2 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm2 +; SSE41-NEXT: movdqa %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm2, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm9 +; SSE41-NEXT: blendvpd %xmm0, %xmm3, %xmm9 +; SSE41-NEXT: movdqa %xmm10, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm4, %xmm2 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm2 +; SSE41-NEXT: movdqa %xmm4, %xmm3 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm2, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm2 +; SSE41-NEXT: blendvpd %xmm0, %xmm10, %xmm2 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm4, %xmm3 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm3 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm3, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm11 +; SSE41-NEXT: movapd {{.*#+}} xmm3 = [18446744073709518848,18446744073709518848] +; SSE41-NEXT: movapd %xmm11, %xmm1 +; SSE41-NEXT: xorpd %xmm5, %xmm1 +; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [18446744071562035200,18446744071562035200] +; SSE41-NEXT: movapd %xmm1, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm1 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm1, %xmm0 +; SSE41-NEXT: movapd %xmm3, %xmm6 +; SSE41-NEXT: blendvpd %xmm0, %xmm11, %xmm6 +; SSE41-NEXT: movapd %xmm2, %xmm1 +; SSE41-NEXT: xorpd %xmm5, %xmm1 +; SSE41-NEXT: movapd %xmm1, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm7 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm1 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm1, %xmm0 +; SSE41-NEXT: movapd %xmm3, %xmm1 +; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm1 +; SSE41-NEXT: packssdw %xmm6, %xmm1 +; SSE41-NEXT: movapd %xmm9, %xmm2 +; SSE41-NEXT: xorpd %xmm5, %xmm2 +; SSE41-NEXT: movapd %xmm2, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm2 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm2, %xmm0 +; SSE41-NEXT: movapd %xmm3, %xmm2 +; SSE41-NEXT: blendvpd %xmm0, %xmm9, %xmm2 +; SSE41-NEXT: xorpd %xmm8, %xmm5 +; SSE41-NEXT: movapd %xmm5, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm5 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm5[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm5, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm8, %xmm3 +; SSE41-NEXT: packssdw %xmm2, %xmm3 +; SSE41-NEXT: packssdw %xmm3, %xmm1 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v8i64_v8i16: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [32767,32767] +; AVX1-NEXT: vpcmpgtq %xmm2, %xmm3, %xmm8 +; AVX1-NEXT: vpcmpgtq %xmm1, %xmm3, %xmm5 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm6 +; AVX1-NEXT: vpcmpgtq %xmm6, %xmm3, %xmm7 +; AVX1-NEXT: vpcmpgtq %xmm0, %xmm3, %xmm4 +; AVX1-NEXT: vblendvpd %xmm4, %xmm0, %xmm3, %xmm0 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [18446744073709518848,18446744073709518848] +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm0, %xmm9 +; AVX1-NEXT: vblendvpd %xmm7, %xmm6, %xmm3, %xmm6 +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm6, %xmm7 +; AVX1-NEXT: vblendvpd %xmm5, %xmm1, %xmm3, %xmm1 +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm1, %xmm5 +; AVX1-NEXT: vblendvpd %xmm8, %xmm2, %xmm3, %xmm2 +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm2, %xmm3 +; AVX1-NEXT: vblendvpd %xmm3, %xmm2, %xmm4, %xmm2 +; AVX1-NEXT: vblendvpd %xmm5, %xmm1, %xmm4, %xmm1 +; AVX1-NEXT: vpackssdw %xmm2, %xmm1, %xmm1 +; AVX1-NEXT: vblendvpd %xmm7, %xmm6, %xmm4, %xmm2 +; AVX1-NEXT: vblendvpd %xmm9, %xmm0, %xmm4, %xmm0 +; AVX1-NEXT: vpackssdw %xmm2, %xmm0, %xmm0 +; AVX1-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v8i64_v8i16: +; AVX2: # %bb.0: +; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [32767,32767,32767,32767] +; AVX2-NEXT: vpcmpgtq %ymm0, %ymm2, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [18446744073709518848,18446744073709518848,18446744073709518848,18446744073709518848] +; AVX2-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-NEXT: vpackssdw %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3] +; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512-LABEL: trunc_ssat_v8i64_v8i16: +; AVX512: # %bb.0: +; AVX512-NEXT: vpmovsqw %zmm0, %xmm0 +; AVX512-NEXT: vzeroupper +; AVX512-NEXT: retq + %1 = icmp slt <8 x i64> %a0, <i64 32767, i64 32767, i64 32767, i64 32767, i64 32767, i64 32767, i64 32767, i64 32767> + %2 = select <8 x i1> %1, <8 x i64> %a0, <8 x i64> <i64 32767, i64 32767, i64 32767, i64 32767, i64 32767, i64 32767, i64 32767, i64 32767> + %3 = icmp sgt <8 x i64> %2, <i64 -32768, i64 -32768, i64 -32768, i64 -32768, i64 -32768, i64 -32768, i64 -32768, i64 -32768> + %4 = select <8 x i1> %3, <8 x i64> %2, <8 x i64> <i64 -32768, i64 -32768, i64 -32768, i64 -32768, i64 -32768, i64 -32768, i64 -32768, i64 -32768> + %5 = trunc <8 x i64> %4 to <8 x i16> + ret <8 x i16> %5 +} + +define <8 x i16> @trunc_ssat_v8i32_v8i16(<8 x i32> %a0) { +; SSE-LABEL: trunc_ssat_v8i32_v8i16: +; SSE: # %bb.0: +; SSE-NEXT: packssdw %xmm1, %xmm0 +; SSE-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v8i32_v8i16: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v8i32_v8i16: +; AVX2: # %bb.0: +; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512F-LABEL: trunc_ssat_v8i32_v8i16: +; AVX512F: # %bb.0: +; AVX512F-NEXT: vpbroadcastd {{.*#+}} ymm1 = [32767,32767,32767,32767,32767,32767,32767,32767] +; AVX512F-NEXT: vpminsd %ymm1, %ymm0, %ymm0 +; AVX512F-NEXT: vpbroadcastd {{.*#+}} ymm1 = [4294934528,4294934528,4294934528,4294934528,4294934528,4294934528,4294934528,4294934528] +; AVX512F-NEXT: vpmaxsd %ymm1, %ymm0, %ymm0 +; AVX512F-NEXT: vpmovdw %zmm0, %ymm0 +; AVX512F-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0 +; AVX512F-NEXT: vzeroupper +; AVX512F-NEXT: retq +; +; AVX512VL-LABEL: trunc_ssat_v8i32_v8i16: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsdw %ymm0, %xmm0 +; AVX512VL-NEXT: vzeroupper +; AVX512VL-NEXT: retq +; +; AVX512BW-LABEL: trunc_ssat_v8i32_v8i16: +; AVX512BW: # %bb.0: +; AVX512BW-NEXT: vpbroadcastd {{.*#+}} ymm1 = [32767,32767,32767,32767,32767,32767,32767,32767] +; AVX512BW-NEXT: vpminsd %ymm1, %ymm0, %ymm0 +; AVX512BW-NEXT: vpbroadcastd {{.*#+}} ymm1 = [4294934528,4294934528,4294934528,4294934528,4294934528,4294934528,4294934528,4294934528] +; AVX512BW-NEXT: vpmaxsd %ymm1, %ymm0, %ymm0 +; AVX512BW-NEXT: vpmovdw %zmm0, %ymm0 +; AVX512BW-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0 +; AVX512BW-NEXT: vzeroupper +; AVX512BW-NEXT: retq +; +; AVX512BWVL-LABEL: trunc_ssat_v8i32_v8i16: +; AVX512BWVL: # %bb.0: +; AVX512BWVL-NEXT: vpmovsdw %ymm0, %xmm0 +; AVX512BWVL-NEXT: vzeroupper +; AVX512BWVL-NEXT: retq + %1 = icmp slt <8 x i32> %a0, <i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767> + %2 = select <8 x i1> %1, <8 x i32> %a0, <8 x i32> <i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767> + %3 = icmp sgt <8 x i32> %2, <i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768> + %4 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> <i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768> + %5 = trunc <8 x i32> %4 to <8 x i16> + ret <8 x i16> %5 +} + +define <16 x i16> @trunc_ssat_v16i32_v16i16(<16 x i32> %a0) { +; SSE-LABEL: trunc_ssat_v16i32_v16i16: +; SSE: # %bb.0: +; SSE-NEXT: packssdw %xmm1, %xmm0 +; SSE-NEXT: packssdw %xmm3, %xmm2 +; SSE-NEXT: movdqa %xmm2, %xmm1 +; SSE-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v16i32_v16i16: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-NEXT: vpackssdw %xmm2, %xmm1, %xmm1 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 +; AVX1-NEXT: vpackssdw %xmm2, %xmm0, %xmm0 +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v16i32_v16i16: +; AVX2: # %bb.0: +; AVX2-NEXT: vpackssdw %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3] +; AVX2-NEXT: retq +; +; AVX512-LABEL: trunc_ssat_v16i32_v16i16: +; AVX512: # %bb.0: +; AVX512-NEXT: vpmovsdw %zmm0, %ymm0 +; AVX512-NEXT: retq + %1 = icmp slt <16 x i32> %a0, <i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767> + %2 = select <16 x i1> %1, <16 x i32> %a0, <16 x i32> <i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767, i32 32767> + %3 = icmp sgt <16 x i32> %2, <i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768> + %4 = select <16 x i1> %3, <16 x i32> %2, <16 x i32> <i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768, i32 -32768> + %5 = trunc <16 x i32> %4 to <16 x i16> + ret <16 x i16> %5 +} + +; +; Signed saturation truncation to v16i8 +; + +define <8 x i8> @trunc_ssat_v8i64_v8i8(<8 x i64> %a0) { +; SSE2-LABEL: trunc_ssat_v8i64_v8i8: +; SSE2: # %bb.0: +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [127,127] +; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648] +; SSE2-NEXT: movdqa %xmm3, %xmm5 +; SSE2-NEXT: pxor %xmm4, %xmm5 +; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [2147483775,2147483775] +; SSE2-NEXT: movdqa %xmm9, %xmm7 +; SSE2-NEXT: pcmpgtd %xmm5, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm5 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm5 +; SSE2-NEXT: pand %xmm5, %xmm3 +; SSE2-NEXT: pandn %xmm8, %xmm5 +; SSE2-NEXT: por %xmm3, %xmm5 +; SSE2-NEXT: movdqa %xmm2, %xmm3 +; SSE2-NEXT: pxor %xmm4, %xmm3 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm3, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm2 +; SSE2-NEXT: pandn %xmm8, %xmm3 +; SSE2-NEXT: por %xmm2, %xmm3 +; SSE2-NEXT: movdqa %xmm1, %xmm2 +; SSE2-NEXT: pxor %xmm4, %xmm2 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm2, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm2[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm1 +; SSE2-NEXT: pandn %xmm8, %xmm2 +; SSE2-NEXT: por %xmm1, %xmm2 +; SSE2-NEXT: movdqa %xmm0, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm1, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm7 +; SSE2-NEXT: pand %xmm7, %xmm0 +; SSE2-NEXT: pandn %xmm8, %xmm7 +; SSE2-NEXT: por %xmm0, %xmm7 +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [18446744073709551488,18446744073709551488] +; SSE2-NEXT: movdqa %xmm7, %xmm0 +; SSE2-NEXT: pxor %xmm4, %xmm0 +; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [18446744071562067840,18446744071562067840] +; SSE2-NEXT: movdqa %xmm0, %xmm1 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm1[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm0 +; SSE2-NEXT: pand %xmm0, %xmm7 +; SSE2-NEXT: pandn %xmm8, %xmm0 +; SSE2-NEXT: por %xmm7, %xmm0 +; SSE2-NEXT: movdqa %xmm2, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm7 +; SSE2-NEXT: pand %xmm7, %xmm2 +; SSE2-NEXT: pandn %xmm8, %xmm7 +; SSE2-NEXT: por %xmm2, %xmm7 +; SSE2-NEXT: movdqa %xmm3, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm3 +; SSE2-NEXT: pandn %xmm8, %xmm2 +; SSE2-NEXT: por %xmm3, %xmm2 +; SSE2-NEXT: pxor %xmm5, %xmm4 +; SSE2-NEXT: movdqa %xmm4, %xmm1 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm1[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSE2-NEXT: pand %xmm3, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: por %xmm4, %xmm1 +; SSE2-NEXT: pand %xmm1, %xmm5 +; SSE2-NEXT: pandn %xmm8, %xmm1 +; SSE2-NEXT: por %xmm5, %xmm1 +; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0] +; SSE2-NEXT: pand %xmm3, %xmm1 +; SSE2-NEXT: pand %xmm3, %xmm2 +; SSE2-NEXT: packuswb %xmm1, %xmm2 +; SSE2-NEXT: pand %xmm3, %xmm7 +; SSE2-NEXT: pand %xmm3, %xmm0 +; SSE2-NEXT: packuswb %xmm7, %xmm0 +; SSE2-NEXT: packuswb %xmm2, %xmm0 +; SSE2-NEXT: packuswb %xmm0, %xmm0 +; SSE2-NEXT: retq +; +; SSSE3-LABEL: trunc_ssat_v8i64_v8i8: +; SSSE3: # %bb.0: +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [127,127] +; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648] +; SSSE3-NEXT: movdqa %xmm3, %xmm5 +; SSSE3-NEXT: pxor %xmm4, %xmm5 +; SSSE3-NEXT: movdqa {{.*#+}} xmm9 = [2147483775,2147483775] +; SSSE3-NEXT: movdqa %xmm9, %xmm7 +; SSSE3-NEXT: pcmpgtd %xmm5, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm5 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm5 +; SSSE3-NEXT: pand %xmm5, %xmm3 +; SSSE3-NEXT: pandn %xmm8, %xmm5 +; SSSE3-NEXT: por %xmm3, %xmm5 +; SSSE3-NEXT: movdqa %xmm2, %xmm3 +; SSSE3-NEXT: pxor %xmm4, %xmm3 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm3, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm2 +; SSSE3-NEXT: pandn %xmm8, %xmm3 +; SSSE3-NEXT: por %xmm2, %xmm3 +; SSSE3-NEXT: movdqa %xmm1, %xmm2 +; SSSE3-NEXT: pxor %xmm4, %xmm2 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm2, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm2[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm1 +; SSSE3-NEXT: pandn %xmm8, %xmm2 +; SSSE3-NEXT: por %xmm1, %xmm2 +; SSSE3-NEXT: movdqa %xmm0, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm1, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm7 +; SSSE3-NEXT: pand %xmm7, %xmm0 +; SSSE3-NEXT: pandn %xmm8, %xmm7 +; SSSE3-NEXT: por %xmm0, %xmm7 +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [18446744073709551488,18446744073709551488] +; SSSE3-NEXT: movdqa %xmm7, %xmm0 +; SSSE3-NEXT: pxor %xmm4, %xmm0 +; SSSE3-NEXT: movdqa {{.*#+}} xmm9 = [18446744071562067840,18446744071562067840] +; SSSE3-NEXT: movdqa %xmm0, %xmm1 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm1[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm0 +; SSSE3-NEXT: pand %xmm0, %xmm7 +; SSSE3-NEXT: pandn %xmm8, %xmm0 +; SSSE3-NEXT: por %xmm7, %xmm0 +; SSSE3-NEXT: movdqa %xmm2, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm7 +; SSSE3-NEXT: pand %xmm7, %xmm2 +; SSSE3-NEXT: pandn %xmm8, %xmm7 +; SSSE3-NEXT: por %xmm2, %xmm7 +; SSSE3-NEXT: movdqa %xmm3, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm3 +; SSSE3-NEXT: pandn %xmm8, %xmm2 +; SSSE3-NEXT: por %xmm3, %xmm2 +; SSSE3-NEXT: pxor %xmm5, %xmm4 +; SSSE3-NEXT: movdqa %xmm4, %xmm1 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm1[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSSE3-NEXT: pand %xmm3, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: por %xmm4, %xmm1 +; SSSE3-NEXT: pand %xmm1, %xmm5 +; SSSE3-NEXT: pandn %xmm8, %xmm1 +; SSSE3-NEXT: por %xmm5, %xmm1 +; SSSE3-NEXT: movdqa {{.*#+}} xmm3 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0] +; SSSE3-NEXT: pand %xmm3, %xmm1 +; SSSE3-NEXT: pand %xmm3, %xmm2 +; SSSE3-NEXT: packuswb %xmm1, %xmm2 +; SSSE3-NEXT: pand %xmm3, %xmm7 +; SSSE3-NEXT: pand %xmm3, %xmm0 +; SSSE3-NEXT: packuswb %xmm7, %xmm0 +; SSSE3-NEXT: packuswb %xmm2, %xmm0 +; SSSE3-NEXT: packuswb %xmm0, %xmm0 +; SSSE3-NEXT: retq +; +; SSE41-LABEL: trunc_ssat_v8i64_v8i8: +; SSE41: # %bb.0: +; SSE41-NEXT: movdqa %xmm0, %xmm8 +; SSE41-NEXT: movapd {{.*#+}} xmm11 = [127,127] +; SSE41-NEXT: movdqa {{.*#+}} xmm5 = [2147483648,2147483648] +; SSE41-NEXT: movdqa %xmm3, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa {{.*#+}} xmm6 = [2147483775,2147483775] +; SSE41-NEXT: movdqa %xmm6, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm7 +; SSE41-NEXT: movdqa %xmm6, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm9 +; SSE41-NEXT: blendvpd %xmm0, %xmm3, %xmm9 +; SSE41-NEXT: movdqa %xmm2, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm6, %xmm3 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm3 +; SSE41-NEXT: movdqa %xmm6, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm3, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm10 +; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm10 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm6, %xmm2 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm2 +; SSE41-NEXT: movdqa %xmm6, %xmm3 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm2, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm3 +; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm3 +; SSE41-NEXT: movdqa %xmm8, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm6, %xmm1 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm1 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm1, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm8, %xmm11 +; SSE41-NEXT: movapd {{.*#+}} xmm2 = [18446744073709551488,18446744073709551488] +; SSE41-NEXT: movapd %xmm11, %xmm1 +; SSE41-NEXT: xorpd %xmm5, %xmm1 +; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [18446744071562067840,18446744071562067840] +; SSE41-NEXT: movapd %xmm1, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm1 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm1, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm1 +; SSE41-NEXT: blendvpd %xmm0, %xmm11, %xmm1 +; SSE41-NEXT: movapd %xmm3, %xmm6 +; SSE41-NEXT: xorpd %xmm5, %xmm6 +; SSE41-NEXT: movapd %xmm6, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm7 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm7 +; SSE41-NEXT: blendvpd %xmm0, %xmm3, %xmm7 +; SSE41-NEXT: movapd %xmm10, %xmm3 +; SSE41-NEXT: xorpd %xmm5, %xmm3 +; SSE41-NEXT: movapd %xmm3, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm3 +; SSE41-NEXT: blendvpd %xmm0, %xmm10, %xmm3 +; SSE41-NEXT: xorpd %xmm9, %xmm5 +; SSE41-NEXT: movapd %xmm5, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm5 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm5[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm5, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm9, %xmm2 +; SSE41-NEXT: movapd {{.*#+}} xmm0 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0] +; SSE41-NEXT: andpd %xmm0, %xmm2 +; SSE41-NEXT: andpd %xmm0, %xmm3 +; SSE41-NEXT: packusdw %xmm2, %xmm3 +; SSE41-NEXT: andpd %xmm0, %xmm7 +; SSE41-NEXT: andpd %xmm0, %xmm1 +; SSE41-NEXT: packusdw %xmm7, %xmm1 +; SSE41-NEXT: packusdw %xmm3, %xmm1 +; SSE41-NEXT: packuswb %xmm1, %xmm1 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v8i64_v8i8: +; AVX1: # %bb.0: +; AVX1-NEXT: vmovapd {{.*#+}} ymm8 = [127,127,127,127] +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [127,127] +; AVX1-NEXT: vpcmpgtq %xmm3, %xmm4, %xmm5 +; AVX1-NEXT: vpcmpgtq %xmm1, %xmm4, %xmm6 +; AVX1-NEXT: vinsertf128 $1, %xmm5, %ymm6, %ymm7 +; AVX1-NEXT: vblendvpd %ymm7, %ymm1, %ymm8, %ymm9 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 +; AVX1-NEXT: vpcmpgtq %xmm2, %xmm4, %xmm7 +; AVX1-NEXT: vpcmpgtq %xmm0, %xmm4, %xmm10 +; AVX1-NEXT: vinsertf128 $1, %xmm7, %ymm10, %ymm11 +; AVX1-NEXT: vblendvpd %ymm11, %ymm0, %ymm8, %ymm8 +; AVX1-NEXT: vmovapd {{.*#+}} ymm11 = [18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488] +; AVX1-NEXT: vblendvpd %xmm7, %xmm2, %xmm4, %xmm2 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm7 = [18446744073709551488,18446744073709551488] +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm2, %xmm2 +; AVX1-NEXT: vblendvpd %xmm10, %xmm0, %xmm4, %xmm0 +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm0, %xmm0 +; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0 +; AVX1-NEXT: vblendvpd %ymm0, %ymm8, %ymm11, %ymm0 +; AVX1-NEXT: vblendvpd %xmm5, %xmm3, %xmm4, %xmm2 +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm2, %xmm2 +; AVX1-NEXT: vblendvpd %xmm6, %xmm1, %xmm4, %xmm1 +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm1, %xmm1 +; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm1, %ymm1 +; AVX1-NEXT: vblendvpd %ymm1, %ymm9, %ymm11, %ymm1 +; AVX1-NEXT: vmovapd {{.*#+}} ymm2 = [255,255,255,255] +; AVX1-NEXT: vandpd %ymm2, %ymm1, %ymm1 +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3 +; AVX1-NEXT: vpackusdw %xmm3, %xmm1, %xmm1 +; AVX1-NEXT: vandpd %ymm2, %ymm0, %ymm0 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 +; AVX1-NEXT: vpackusdw %xmm2, %xmm0, %xmm0 +; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vpackuswb %xmm0, %xmm0, %xmm0 +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v8i64_v8i8: +; AVX2: # %bb.0: +; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [127,127,127,127] +; AVX2-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-NEXT: vpcmpgtq %ymm0, %ymm2, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488] +; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-NEXT: vextractf128 $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: vextractf128 $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: vzeroupper +; AVX2-NEXT: retq +; +; AVX512-LABEL: trunc_ssat_v8i64_v8i8: +; AVX512: # %bb.0: +; AVX512-NEXT: vpminsq {{.*}}(%rip){1to8}, %zmm0, %zmm0 +; AVX512-NEXT: vpmaxsq {{.*}}(%rip){1to8}, %zmm0, %zmm0 +; AVX512-NEXT: vpmovqb %zmm0, %xmm0 +; AVX512-NEXT: vzeroupper +; AVX512-NEXT: retq + %1 = icmp slt <8 x i64> %a0, <i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127> + %2 = select <8 x i1> %1, <8 x i64> %a0, <8 x i64> <i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127> + %3 = icmp sgt <8 x i64> %2, <i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128> + %4 = select <8 x i1> %3, <8 x i64> %2, <8 x i64> <i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128> + %5 = trunc <8 x i64> %4 to <8 x i8> + ret <8 x i8> %5 +} + +; TODO: The AVX1 codegen shows a missed opportunity to narrow blendv+logic to 128-bit. + +define void @trunc_ssat_v8i64_v8i8_store(<8 x i64> %a0, <8 x i8> *%p1) { +; SSE2-LABEL: trunc_ssat_v8i64_v8i8_store: +; SSE2: # %bb.0: +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [127,127] +; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648] +; SSE2-NEXT: movdqa %xmm3, %xmm5 +; SSE2-NEXT: pxor %xmm4, %xmm5 +; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [2147483775,2147483775] +; SSE2-NEXT: movdqa %xmm9, %xmm7 +; SSE2-NEXT: pcmpgtd %xmm5, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm5 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm5 +; SSE2-NEXT: pand %xmm5, %xmm3 +; SSE2-NEXT: pandn %xmm8, %xmm5 +; SSE2-NEXT: por %xmm3, %xmm5 +; SSE2-NEXT: movdqa %xmm2, %xmm3 +; SSE2-NEXT: pxor %xmm4, %xmm3 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm3, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm2 +; SSE2-NEXT: pandn %xmm8, %xmm3 +; SSE2-NEXT: por %xmm2, %xmm3 +; SSE2-NEXT: movdqa %xmm1, %xmm2 +; SSE2-NEXT: pxor %xmm4, %xmm2 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm2, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm2[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm1 +; SSE2-NEXT: pandn %xmm8, %xmm2 +; SSE2-NEXT: por %xmm1, %xmm2 +; SSE2-NEXT: movdqa %xmm0, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm9, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm1, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm7 +; SSE2-NEXT: pand %xmm7, %xmm0 +; SSE2-NEXT: pandn %xmm8, %xmm7 +; SSE2-NEXT: por %xmm0, %xmm7 +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [18446744073709551488,18446744073709551488] +; SSE2-NEXT: movdqa %xmm7, %xmm0 +; SSE2-NEXT: pxor %xmm4, %xmm0 +; SSE2-NEXT: movdqa {{.*#+}} xmm9 = [18446744071562067840,18446744071562067840] +; SSE2-NEXT: movdqa %xmm0, %xmm1 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm10 = xmm1[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm10, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm0 +; SSE2-NEXT: pand %xmm0, %xmm7 +; SSE2-NEXT: pandn %xmm8, %xmm0 +; SSE2-NEXT: por %xmm7, %xmm0 +; SSE2-NEXT: movdqa %xmm2, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm7 +; SSE2-NEXT: pand %xmm7, %xmm2 +; SSE2-NEXT: pandn %xmm8, %xmm7 +; SSE2-NEXT: por %xmm2, %xmm7 +; SSE2-NEXT: movdqa %xmm3, %xmm1 +; SSE2-NEXT: pxor %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm3 +; SSE2-NEXT: pandn %xmm8, %xmm2 +; SSE2-NEXT: por %xmm3, %xmm2 +; SSE2-NEXT: pxor %xmm5, %xmm4 +; SSE2-NEXT: movdqa %xmm4, %xmm1 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm1[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm9, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSE2-NEXT: pand %xmm3, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: por %xmm4, %xmm1 +; SSE2-NEXT: pand %xmm1, %xmm5 +; SSE2-NEXT: pandn %xmm8, %xmm1 +; SSE2-NEXT: por %xmm5, %xmm1 +; SSE2-NEXT: movdqa {{.*#+}} xmm3 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0] +; SSE2-NEXT: pand %xmm3, %xmm1 +; SSE2-NEXT: pand %xmm3, %xmm2 +; SSE2-NEXT: packuswb %xmm1, %xmm2 +; SSE2-NEXT: pand %xmm3, %xmm7 +; SSE2-NEXT: pand %xmm3, %xmm0 +; SSE2-NEXT: packuswb %xmm7, %xmm0 +; SSE2-NEXT: packuswb %xmm2, %xmm0 +; SSE2-NEXT: packuswb %xmm0, %xmm0 +; SSE2-NEXT: movq %xmm0, (%rdi) +; SSE2-NEXT: retq +; +; SSSE3-LABEL: trunc_ssat_v8i64_v8i8_store: +; SSSE3: # %bb.0: +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [127,127] +; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648] +; SSSE3-NEXT: movdqa %xmm3, %xmm5 +; SSSE3-NEXT: pxor %xmm4, %xmm5 +; SSSE3-NEXT: movdqa {{.*#+}} xmm9 = [2147483775,2147483775] +; SSSE3-NEXT: movdqa %xmm9, %xmm7 +; SSSE3-NEXT: pcmpgtd %xmm5, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm7[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm5 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm5[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm5 +; SSSE3-NEXT: pand %xmm5, %xmm3 +; SSSE3-NEXT: pandn %xmm8, %xmm5 +; SSSE3-NEXT: por %xmm3, %xmm5 +; SSSE3-NEXT: movdqa %xmm2, %xmm3 +; SSSE3-NEXT: pxor %xmm4, %xmm3 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm3, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm3[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm2 +; SSSE3-NEXT: pandn %xmm8, %xmm3 +; SSSE3-NEXT: por %xmm2, %xmm3 +; SSSE3-NEXT: movdqa %xmm1, %xmm2 +; SSSE3-NEXT: pxor %xmm4, %xmm2 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm2, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm2[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm1 +; SSSE3-NEXT: pandn %xmm8, %xmm2 +; SSSE3-NEXT: por %xmm1, %xmm2 +; SSSE3-NEXT: movdqa %xmm0, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm9, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm1, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm7 +; SSSE3-NEXT: pand %xmm7, %xmm0 +; SSSE3-NEXT: pandn %xmm8, %xmm7 +; SSSE3-NEXT: por %xmm0, %xmm7 +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [18446744073709551488,18446744073709551488] +; SSSE3-NEXT: movdqa %xmm7, %xmm0 +; SSSE3-NEXT: pxor %xmm4, %xmm0 +; SSSE3-NEXT: movdqa {{.*#+}} xmm9 = [18446744071562067840,18446744071562067840] +; SSSE3-NEXT: movdqa %xmm0, %xmm1 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm10 = xmm1[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm10, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm0 +; SSSE3-NEXT: pand %xmm0, %xmm7 +; SSSE3-NEXT: pandn %xmm8, %xmm0 +; SSSE3-NEXT: por %xmm7, %xmm0 +; SSSE3-NEXT: movdqa %xmm2, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm7 +; SSSE3-NEXT: pand %xmm7, %xmm2 +; SSSE3-NEXT: pandn %xmm8, %xmm7 +; SSSE3-NEXT: por %xmm2, %xmm7 +; SSSE3-NEXT: movdqa %xmm3, %xmm1 +; SSSE3-NEXT: pxor %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm3 +; SSSE3-NEXT: pandn %xmm8, %xmm2 +; SSSE3-NEXT: por %xmm3, %xmm2 +; SSSE3-NEXT: pxor %xmm5, %xmm4 +; SSSE3-NEXT: movdqa %xmm4, %xmm1 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm1[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm9, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSSE3-NEXT: pand %xmm3, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: por %xmm4, %xmm1 +; SSSE3-NEXT: pand %xmm1, %xmm5 +; SSSE3-NEXT: pandn %xmm8, %xmm1 +; SSSE3-NEXT: por %xmm5, %xmm1 +; SSSE3-NEXT: movdqa {{.*#+}} xmm3 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0] +; SSSE3-NEXT: pand %xmm3, %xmm1 +; SSSE3-NEXT: pand %xmm3, %xmm2 +; SSSE3-NEXT: packuswb %xmm1, %xmm2 +; SSSE3-NEXT: pand %xmm3, %xmm7 +; SSSE3-NEXT: pand %xmm3, %xmm0 +; SSSE3-NEXT: packuswb %xmm7, %xmm0 +; SSSE3-NEXT: packuswb %xmm2, %xmm0 +; SSSE3-NEXT: packuswb %xmm0, %xmm0 +; SSSE3-NEXT: movq %xmm0, (%rdi) +; SSSE3-NEXT: retq +; +; SSE41-LABEL: trunc_ssat_v8i64_v8i8_store: +; SSE41: # %bb.0: +; SSE41-NEXT: movdqa %xmm0, %xmm8 +; SSE41-NEXT: movapd {{.*#+}} xmm11 = [127,127] +; SSE41-NEXT: movdqa {{.*#+}} xmm5 = [2147483648,2147483648] +; SSE41-NEXT: movdqa %xmm3, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa {{.*#+}} xmm6 = [2147483775,2147483775] +; SSE41-NEXT: movdqa %xmm6, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm7 +; SSE41-NEXT: movdqa %xmm6, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm9 +; SSE41-NEXT: blendvpd %xmm0, %xmm3, %xmm9 +; SSE41-NEXT: movdqa %xmm2, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm6, %xmm3 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm3 +; SSE41-NEXT: movdqa %xmm6, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm3, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm10 +; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm10 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm6, %xmm2 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm2 +; SSE41-NEXT: movdqa %xmm6, %xmm3 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm2, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm3 +; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm3 +; SSE41-NEXT: movdqa %xmm8, %xmm0 +; SSE41-NEXT: pxor %xmm5, %xmm0 +; SSE41-NEXT: movdqa %xmm6, %xmm1 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm1 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm1, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm8, %xmm11 +; SSE41-NEXT: movapd {{.*#+}} xmm1 = [18446744073709551488,18446744073709551488] +; SSE41-NEXT: movapd %xmm11, %xmm2 +; SSE41-NEXT: xorpd %xmm5, %xmm2 +; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [18446744071562067840,18446744071562067840] +; SSE41-NEXT: movapd %xmm2, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm2 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm2, %xmm0 +; SSE41-NEXT: movapd %xmm1, %xmm2 +; SSE41-NEXT: blendvpd %xmm0, %xmm11, %xmm2 +; SSE41-NEXT: movapd %xmm3, %xmm6 +; SSE41-NEXT: xorpd %xmm5, %xmm6 +; SSE41-NEXT: movapd %xmm6, %xmm7 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm7 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm7, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm1, %xmm7 +; SSE41-NEXT: blendvpd %xmm0, %xmm3, %xmm7 +; SSE41-NEXT: movapd %xmm10, %xmm3 +; SSE41-NEXT: xorpd %xmm5, %xmm3 +; SSE41-NEXT: movapd %xmm3, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm1, %xmm3 +; SSE41-NEXT: blendvpd %xmm0, %xmm10, %xmm3 +; SSE41-NEXT: xorpd %xmm9, %xmm5 +; SSE41-NEXT: movapd %xmm5, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm4, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm4, %xmm5 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm5[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm5, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm9, %xmm1 +; SSE41-NEXT: movapd {{.*#+}} xmm0 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0] +; SSE41-NEXT: andpd %xmm0, %xmm1 +; SSE41-NEXT: andpd %xmm0, %xmm3 +; SSE41-NEXT: packusdw %xmm1, %xmm3 +; SSE41-NEXT: andpd %xmm0, %xmm7 +; SSE41-NEXT: andpd %xmm0, %xmm2 +; SSE41-NEXT: packusdw %xmm7, %xmm2 +; SSE41-NEXT: packusdw %xmm3, %xmm2 +; SSE41-NEXT: packuswb %xmm0, %xmm2 +; SSE41-NEXT: movq %xmm2, (%rdi) +; SSE41-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v8i64_v8i8_store: +; AVX1: # %bb.0: +; AVX1-NEXT: vmovapd {{.*#+}} ymm8 = [127,127,127,127] +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [127,127] +; AVX1-NEXT: vpcmpgtq %xmm3, %xmm4, %xmm5 +; AVX1-NEXT: vpcmpgtq %xmm1, %xmm4, %xmm6 +; AVX1-NEXT: vinsertf128 $1, %xmm5, %ymm6, %ymm7 +; AVX1-NEXT: vblendvpd %ymm7, %ymm1, %ymm8, %ymm9 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 +; AVX1-NEXT: vpcmpgtq %xmm2, %xmm4, %xmm7 +; AVX1-NEXT: vpcmpgtq %xmm0, %xmm4, %xmm10 +; AVX1-NEXT: vinsertf128 $1, %xmm7, %ymm10, %ymm11 +; AVX1-NEXT: vblendvpd %ymm11, %ymm0, %ymm8, %ymm8 +; AVX1-NEXT: vmovapd {{.*#+}} ymm11 = [18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488] +; AVX1-NEXT: vblendvpd %xmm7, %xmm2, %xmm4, %xmm2 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm7 = [18446744073709551488,18446744073709551488] +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm2, %xmm2 +; AVX1-NEXT: vblendvpd %xmm10, %xmm0, %xmm4, %xmm0 +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm0, %xmm0 +; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0 +; AVX1-NEXT: vblendvpd %ymm0, %ymm8, %ymm11, %ymm0 +; AVX1-NEXT: vblendvpd %xmm5, %xmm3, %xmm4, %xmm2 +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm2, %xmm2 +; AVX1-NEXT: vblendvpd %xmm6, %xmm1, %xmm4, %xmm1 +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm1, %xmm1 +; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm1, %ymm1 +; AVX1-NEXT: vblendvpd %ymm1, %ymm9, %ymm11, %ymm1 +; AVX1-NEXT: vmovapd {{.*#+}} ymm2 = [255,255,255,255] +; AVX1-NEXT: vandpd %ymm2, %ymm1, %ymm1 +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3 +; AVX1-NEXT: vpackusdw %xmm3, %xmm1, %xmm1 +; AVX1-NEXT: vandpd %ymm2, %ymm0, %ymm0 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 +; AVX1-NEXT: vpackusdw %xmm2, %xmm0, %xmm0 +; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vpackuswb %xmm0, %xmm0, %xmm0 +; AVX1-NEXT: vmovq %xmm0, (%rdi) +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v8i64_v8i8_store: +; AVX2: # %bb.0: +; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [127,127,127,127] +; AVX2-NEXT: vpcmpgtq %ymm1, %ymm2, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-NEXT: vpcmpgtq %ymm0, %ymm2, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488] +; AVX2-NEXT: vpcmpgtq %ymm2, %ymm0, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm2, %ymm0 +; AVX2-NEXT: vpcmpgtq %ymm2, %ymm1, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm1, %ymm2, %ymm1 +; AVX2-NEXT: vextractf128 $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: vextractf128 $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, (%rdi) +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512-LABEL: trunc_ssat_v8i64_v8i8_store: +; AVX512: # %bb.0: +; AVX512-NEXT: vpmovsqb %zmm0, (%rdi) +; AVX512-NEXT: vzeroupper +; AVX512-NEXT: retq + %1 = icmp slt <8 x i64> %a0, <i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127> + %2 = select <8 x i1> %1, <8 x i64> %a0, <8 x i64> <i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127> + %3 = icmp sgt <8 x i64> %2, <i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128> + %4 = select <8 x i1> %3, <8 x i64> %2, <8 x i64> <i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128> + %5 = trunc <8 x i64> %4 to <8 x i8> + store <8 x i8> %5, <8 x i8> *%p1 + ret void +} + +define <16 x i8> @trunc_ssat_v16i64_v16i8(<16 x i64> %a0) { +; SSE2-LABEL: trunc_ssat_v16i64_v16i8: +; SSE2: # %bb.0: +; SSE2-NEXT: movdqa {{.*#+}} xmm10 = [127,127] +; SSE2-NEXT: movdqa {{.*#+}} xmm8 = [2147483648,2147483648] +; SSE2-NEXT: movdqa %xmm6, %xmm9 +; SSE2-NEXT: pxor %xmm8, %xmm9 +; SSE2-NEXT: movdqa {{.*#+}} xmm11 = [2147483775,2147483775] +; SSE2-NEXT: movdqa %xmm11, %xmm12 +; SSE2-NEXT: pcmpgtd %xmm9, %xmm12 +; SSE2-NEXT: pshufd {{.*#+}} xmm13 = xmm12[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm9 +; SSE2-NEXT: pshufd {{.*#+}} xmm14 = xmm9[1,1,3,3] +; SSE2-NEXT: pand %xmm13, %xmm14 +; SSE2-NEXT: pshufd {{.*#+}} xmm9 = xmm12[1,1,3,3] +; SSE2-NEXT: por %xmm14, %xmm9 +; SSE2-NEXT: pand %xmm9, %xmm6 +; SSE2-NEXT: pandn %xmm10, %xmm9 +; SSE2-NEXT: por %xmm6, %xmm9 +; SSE2-NEXT: movdqa %xmm7, %xmm6 +; SSE2-NEXT: pxor %xmm8, %xmm6 +; SSE2-NEXT: movdqa %xmm11, %xmm12 +; SSE2-NEXT: pcmpgtd %xmm6, %xmm12 +; SSE2-NEXT: pshufd {{.*#+}} xmm13 = xmm12[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3] +; SSE2-NEXT: pand %xmm13, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm12 = xmm12[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm12 +; SSE2-NEXT: pand %xmm12, %xmm7 +; SSE2-NEXT: pandn %xmm10, %xmm12 +; SSE2-NEXT: por %xmm7, %xmm12 +; SSE2-NEXT: movdqa %xmm4, %xmm6 +; SSE2-NEXT: pxor %xmm8, %xmm6 +; SSE2-NEXT: movdqa %xmm11, %xmm7 +; SSE2-NEXT: pcmpgtd %xmm6, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm13 = xmm7[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3] +; SSE2-NEXT: pand %xmm13, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm13 = xmm7[1,1,3,3] +; SSE2-NEXT: por %xmm6, %xmm13 +; SSE2-NEXT: pand %xmm13, %xmm4 +; SSE2-NEXT: pandn %xmm10, %xmm13 +; SSE2-NEXT: por %xmm4, %xmm13 +; SSE2-NEXT: movdqa %xmm5, %xmm4 +; SSE2-NEXT: pxor %xmm8, %xmm4 +; SSE2-NEXT: movdqa %xmm11, %xmm6 +; SSE2-NEXT: pcmpgtd %xmm4, %xmm6 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSE2-NEXT: pand %xmm7, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm14 = xmm6[1,1,3,3] +; SSE2-NEXT: por %xmm4, %xmm14 +; SSE2-NEXT: pand %xmm14, %xmm5 +; SSE2-NEXT: pandn %xmm10, %xmm14 +; SSE2-NEXT: por %xmm5, %xmm14 +; SSE2-NEXT: movdqa %xmm2, %xmm4 +; SSE2-NEXT: pxor %xmm8, %xmm4 +; SSE2-NEXT: movdqa %xmm11, %xmm5 +; SSE2-NEXT: pcmpgtd %xmm4, %xmm5 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3] +; SSE2-NEXT: por %xmm4, %xmm5 +; SSE2-NEXT: pand %xmm5, %xmm2 +; SSE2-NEXT: pandn %xmm10, %xmm5 +; SSE2-NEXT: por %xmm2, %xmm5 +; SSE2-NEXT: movdqa %xmm3, %xmm2 +; SSE2-NEXT: pxor %xmm8, %xmm2 +; SSE2-NEXT: movdqa %xmm11, %xmm4 +; SSE2-NEXT: pcmpgtd %xmm2, %xmm4 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: pand %xmm6, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm4[1,1,3,3] +; SSE2-NEXT: por %xmm2, %xmm6 +; SSE2-NEXT: pand %xmm6, %xmm3 +; SSE2-NEXT: pandn %xmm10, %xmm6 +; SSE2-NEXT: por %xmm3, %xmm6 +; SSE2-NEXT: movdqa %xmm0, %xmm2 +; SSE2-NEXT: pxor %xmm8, %xmm2 +; SSE2-NEXT: movdqa %xmm11, %xmm3 +; SSE2-NEXT: pcmpgtd %xmm2, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: pand %xmm4, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSE2-NEXT: por %xmm2, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm0 +; SSE2-NEXT: pandn %xmm10, %xmm3 +; SSE2-NEXT: por %xmm0, %xmm3 +; SSE2-NEXT: movdqa %xmm1, %xmm0 +; SSE2-NEXT: pxor %xmm8, %xmm0 +; SSE2-NEXT: movdqa %xmm11, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm0, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm4, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm0, %xmm4 +; SSE2-NEXT: pand %xmm4, %xmm1 +; SSE2-NEXT: pandn %xmm10, %xmm4 +; SSE2-NEXT: por %xmm1, %xmm4 +; SSE2-NEXT: movdqa {{.*#+}} xmm10 = [18446744073709551488,18446744073709551488] +; SSE2-NEXT: movdqa %xmm4, %xmm0 +; SSE2-NEXT: pxor %xmm8, %xmm0 +; SSE2-NEXT: movdqa {{.*#+}} xmm11 = [18446744071562067840,18446744071562067840] +; SSE2-NEXT: movdqa %xmm0, %xmm1 +; SSE2-NEXT: pcmpgtd %xmm11, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm2, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: por %xmm0, %xmm1 +; SSE2-NEXT: pand %xmm1, %xmm4 +; SSE2-NEXT: pandn %xmm10, %xmm1 +; SSE2-NEXT: por %xmm4, %xmm1 +; SSE2-NEXT: movdqa %xmm3, %xmm0 +; SSE2-NEXT: pxor %xmm8, %xmm0 +; SSE2-NEXT: movdqa %xmm0, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm11, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm0[1,1,3,3] +; SSE2-NEXT: pand %xmm4, %xmm7 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm7, %xmm0 +; SSE2-NEXT: pand %xmm0, %xmm3 +; SSE2-NEXT: pandn %xmm10, %xmm0 +; SSE2-NEXT: por %xmm3, %xmm0 +; SSE2-NEXT: packssdw %xmm1, %xmm0 +; SSE2-NEXT: movdqa %xmm6, %xmm1 +; SSE2-NEXT: pxor %xmm8, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm11, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm3, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm6 +; SSE2-NEXT: pandn %xmm10, %xmm2 +; SSE2-NEXT: por %xmm6, %xmm2 +; SSE2-NEXT: movdqa %xmm5, %xmm1 +; SSE2-NEXT: pxor %xmm8, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm3 +; SSE2-NEXT: pcmpgtd %xmm11, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm4, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm5 +; SSE2-NEXT: pandn %xmm10, %xmm3 +; SSE2-NEXT: por %xmm5, %xmm3 +; SSE2-NEXT: packssdw %xmm2, %xmm3 +; SSE2-NEXT: packssdw %xmm3, %xmm0 +; SSE2-NEXT: movdqa %xmm14, %xmm1 +; SSE2-NEXT: pxor %xmm8, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm11, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm3, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm14 +; SSE2-NEXT: pandn %xmm10, %xmm2 +; SSE2-NEXT: por %xmm14, %xmm2 +; SSE2-NEXT: movdqa %xmm13, %xmm1 +; SSE2-NEXT: pxor %xmm8, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm3 +; SSE2-NEXT: pcmpgtd %xmm11, %xmm3 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm4, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm3 +; SSE2-NEXT: pand %xmm3, %xmm13 +; SSE2-NEXT: pandn %xmm10, %xmm3 +; SSE2-NEXT: por %xmm13, %xmm3 +; SSE2-NEXT: packssdw %xmm2, %xmm3 +; SSE2-NEXT: movdqa %xmm12, %xmm1 +; SSE2-NEXT: pxor %xmm8, %xmm1 +; SSE2-NEXT: movdqa %xmm1, %xmm2 +; SSE2-NEXT: pcmpgtd %xmm11, %xmm2 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pand %xmm4, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSE2-NEXT: por %xmm1, %xmm2 +; SSE2-NEXT: pand %xmm2, %xmm12 +; SSE2-NEXT: pandn %xmm10, %xmm2 +; SSE2-NEXT: por %xmm12, %xmm2 +; SSE2-NEXT: pxor %xmm9, %xmm8 +; SSE2-NEXT: movdqa %xmm8, %xmm1 +; SSE2-NEXT: pcmpgtd %xmm11, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm1[0,0,2,2] +; SSE2-NEXT: pcmpeqd %xmm11, %xmm8 +; SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm8[1,1,3,3] +; SSE2-NEXT: pand %xmm4, %xmm5 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: por %xmm5, %xmm1 +; SSE2-NEXT: pand %xmm1, %xmm9 +; SSE2-NEXT: pandn %xmm10, %xmm1 +; SSE2-NEXT: por %xmm9, %xmm1 +; SSE2-NEXT: packssdw %xmm2, %xmm1 +; SSE2-NEXT: packssdw %xmm1, %xmm3 +; SSE2-NEXT: packsswb %xmm3, %xmm0 +; SSE2-NEXT: retq +; +; SSSE3-LABEL: trunc_ssat_v16i64_v16i8: +; SSSE3: # %bb.0: +; SSSE3-NEXT: movdqa {{.*#+}} xmm10 = [127,127] +; SSSE3-NEXT: movdqa {{.*#+}} xmm8 = [2147483648,2147483648] +; SSSE3-NEXT: movdqa %xmm6, %xmm9 +; SSSE3-NEXT: pxor %xmm8, %xmm9 +; SSSE3-NEXT: movdqa {{.*#+}} xmm11 = [2147483775,2147483775] +; SSSE3-NEXT: movdqa %xmm11, %xmm12 +; SSSE3-NEXT: pcmpgtd %xmm9, %xmm12 +; SSSE3-NEXT: pshufd {{.*#+}} xmm13 = xmm12[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm9 +; SSSE3-NEXT: pshufd {{.*#+}} xmm14 = xmm9[1,1,3,3] +; SSSE3-NEXT: pand %xmm13, %xmm14 +; SSSE3-NEXT: pshufd {{.*#+}} xmm9 = xmm12[1,1,3,3] +; SSSE3-NEXT: por %xmm14, %xmm9 +; SSSE3-NEXT: pand %xmm9, %xmm6 +; SSSE3-NEXT: pandn %xmm10, %xmm9 +; SSSE3-NEXT: por %xmm6, %xmm9 +; SSSE3-NEXT: movdqa %xmm7, %xmm6 +; SSSE3-NEXT: pxor %xmm8, %xmm6 +; SSSE3-NEXT: movdqa %xmm11, %xmm12 +; SSSE3-NEXT: pcmpgtd %xmm6, %xmm12 +; SSSE3-NEXT: pshufd {{.*#+}} xmm13 = xmm12[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3] +; SSSE3-NEXT: pand %xmm13, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm12 = xmm12[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm12 +; SSSE3-NEXT: pand %xmm12, %xmm7 +; SSSE3-NEXT: pandn %xmm10, %xmm12 +; SSSE3-NEXT: por %xmm7, %xmm12 +; SSSE3-NEXT: movdqa %xmm4, %xmm6 +; SSSE3-NEXT: pxor %xmm8, %xmm6 +; SSSE3-NEXT: movdqa %xmm11, %xmm7 +; SSSE3-NEXT: pcmpgtd %xmm6, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm13 = xmm7[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3] +; SSSE3-NEXT: pand %xmm13, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm13 = xmm7[1,1,3,3] +; SSSE3-NEXT: por %xmm6, %xmm13 +; SSSE3-NEXT: pand %xmm13, %xmm4 +; SSSE3-NEXT: pandn %xmm10, %xmm13 +; SSSE3-NEXT: por %xmm4, %xmm13 +; SSSE3-NEXT: movdqa %xmm5, %xmm4 +; SSSE3-NEXT: pxor %xmm8, %xmm4 +; SSSE3-NEXT: movdqa %xmm11, %xmm6 +; SSSE3-NEXT: pcmpgtd %xmm4, %xmm6 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSSE3-NEXT: pand %xmm7, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm14 = xmm6[1,1,3,3] +; SSSE3-NEXT: por %xmm4, %xmm14 +; SSSE3-NEXT: pand %xmm14, %xmm5 +; SSSE3-NEXT: pandn %xmm10, %xmm14 +; SSSE3-NEXT: por %xmm5, %xmm14 +; SSSE3-NEXT: movdqa %xmm2, %xmm4 +; SSSE3-NEXT: pxor %xmm8, %xmm4 +; SSSE3-NEXT: movdqa %xmm11, %xmm5 +; SSSE3-NEXT: pcmpgtd %xmm4, %xmm5 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3] +; SSSE3-NEXT: por %xmm4, %xmm5 +; SSSE3-NEXT: pand %xmm5, %xmm2 +; SSSE3-NEXT: pandn %xmm10, %xmm5 +; SSSE3-NEXT: por %xmm2, %xmm5 +; SSSE3-NEXT: movdqa %xmm3, %xmm2 +; SSSE3-NEXT: pxor %xmm8, %xmm2 +; SSSE3-NEXT: movdqa %xmm11, %xmm4 +; SSSE3-NEXT: pcmpgtd %xmm2, %xmm4 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: pand %xmm6, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm6 = xmm4[1,1,3,3] +; SSSE3-NEXT: por %xmm2, %xmm6 +; SSSE3-NEXT: pand %xmm6, %xmm3 +; SSSE3-NEXT: pandn %xmm10, %xmm6 +; SSSE3-NEXT: por %xmm3, %xmm6 +; SSSE3-NEXT: movdqa %xmm0, %xmm2 +; SSSE3-NEXT: pxor %xmm8, %xmm2 +; SSSE3-NEXT: movdqa %xmm11, %xmm3 +; SSSE3-NEXT: pcmpgtd %xmm2, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: pand %xmm4, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSSE3-NEXT: por %xmm2, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm0 +; SSSE3-NEXT: pandn %xmm10, %xmm3 +; SSSE3-NEXT: por %xmm0, %xmm3 +; SSSE3-NEXT: movdqa %xmm1, %xmm0 +; SSSE3-NEXT: pxor %xmm8, %xmm0 +; SSSE3-NEXT: movdqa %xmm11, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm0, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm4, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm0, %xmm4 +; SSSE3-NEXT: pand %xmm4, %xmm1 +; SSSE3-NEXT: pandn %xmm10, %xmm4 +; SSSE3-NEXT: por %xmm1, %xmm4 +; SSSE3-NEXT: movdqa {{.*#+}} xmm10 = [18446744073709551488,18446744073709551488] +; SSSE3-NEXT: movdqa %xmm4, %xmm0 +; SSSE3-NEXT: pxor %xmm8, %xmm0 +; SSSE3-NEXT: movdqa {{.*#+}} xmm11 = [18446744071562067840,18446744071562067840] +; SSSE3-NEXT: movdqa %xmm0, %xmm1 +; SSSE3-NEXT: pcmpgtd %xmm11, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm2, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: por %xmm0, %xmm1 +; SSSE3-NEXT: pand %xmm1, %xmm4 +; SSSE3-NEXT: pandn %xmm10, %xmm1 +; SSSE3-NEXT: por %xmm4, %xmm1 +; SSSE3-NEXT: movdqa %xmm3, %xmm0 +; SSSE3-NEXT: pxor %xmm8, %xmm0 +; SSSE3-NEXT: movdqa %xmm0, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm11, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm0 +; SSSE3-NEXT: pshufd {{.*#+}} xmm7 = xmm0[1,1,3,3] +; SSSE3-NEXT: pand %xmm4, %xmm7 +; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm7, %xmm0 +; SSSE3-NEXT: pand %xmm0, %xmm3 +; SSSE3-NEXT: pandn %xmm10, %xmm0 +; SSSE3-NEXT: por %xmm3, %xmm0 +; SSSE3-NEXT: packssdw %xmm1, %xmm0 +; SSSE3-NEXT: movdqa %xmm6, %xmm1 +; SSSE3-NEXT: pxor %xmm8, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm11, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm3, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm6 +; SSSE3-NEXT: pandn %xmm10, %xmm2 +; SSSE3-NEXT: por %xmm6, %xmm2 +; SSSE3-NEXT: movdqa %xmm5, %xmm1 +; SSSE3-NEXT: pxor %xmm8, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm3 +; SSSE3-NEXT: pcmpgtd %xmm11, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm4, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm5 +; SSSE3-NEXT: pandn %xmm10, %xmm3 +; SSSE3-NEXT: por %xmm5, %xmm3 +; SSSE3-NEXT: packssdw %xmm2, %xmm3 +; SSSE3-NEXT: packssdw %xmm3, %xmm0 +; SSSE3-NEXT: movdqa %xmm14, %xmm1 +; SSSE3-NEXT: pxor %xmm8, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm11, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm3, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm14 +; SSSE3-NEXT: pandn %xmm10, %xmm2 +; SSSE3-NEXT: por %xmm14, %xmm2 +; SSSE3-NEXT: movdqa %xmm13, %xmm1 +; SSSE3-NEXT: pxor %xmm8, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm3 +; SSSE3-NEXT: pcmpgtd %xmm11, %xmm3 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm4, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm3 +; SSSE3-NEXT: pand %xmm3, %xmm13 +; SSSE3-NEXT: pandn %xmm10, %xmm3 +; SSSE3-NEXT: por %xmm13, %xmm3 +; SSSE3-NEXT: packssdw %xmm2, %xmm3 +; SSSE3-NEXT: movdqa %xmm12, %xmm1 +; SSSE3-NEXT: pxor %xmm8, %xmm1 +; SSSE3-NEXT: movdqa %xmm1, %xmm2 +; SSSE3-NEXT: pcmpgtd %xmm11, %xmm2 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: pand %xmm4, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3] +; SSSE3-NEXT: por %xmm1, %xmm2 +; SSSE3-NEXT: pand %xmm2, %xmm12 +; SSSE3-NEXT: pandn %xmm10, %xmm2 +; SSSE3-NEXT: por %xmm12, %xmm2 +; SSSE3-NEXT: pxor %xmm9, %xmm8 +; SSSE3-NEXT: movdqa %xmm8, %xmm1 +; SSSE3-NEXT: pcmpgtd %xmm11, %xmm1 +; SSSE3-NEXT: pshufd {{.*#+}} xmm4 = xmm1[0,0,2,2] +; SSSE3-NEXT: pcmpeqd %xmm11, %xmm8 +; SSSE3-NEXT: pshufd {{.*#+}} xmm5 = xmm8[1,1,3,3] +; SSSE3-NEXT: pand %xmm4, %xmm5 +; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSSE3-NEXT: por %xmm5, %xmm1 +; SSSE3-NEXT: pand %xmm1, %xmm9 +; SSSE3-NEXT: pandn %xmm10, %xmm1 +; SSSE3-NEXT: por %xmm9, %xmm1 +; SSSE3-NEXT: packssdw %xmm2, %xmm1 +; SSSE3-NEXT: packssdw %xmm1, %xmm3 +; SSSE3-NEXT: packsswb %xmm3, %xmm0 +; SSSE3-NEXT: retq +; +; SSE41-LABEL: trunc_ssat_v16i64_v16i8: +; SSE41: # %bb.0: +; SSE41-NEXT: movdqa %xmm0, %xmm8 +; SSE41-NEXT: movapd {{.*#+}} xmm11 = [127,127] +; SSE41-NEXT: movdqa {{.*#+}} xmm9 = [2147483648,2147483648] +; SSE41-NEXT: movdqa %xmm6, %xmm0 +; SSE41-NEXT: pxor %xmm9, %xmm0 +; SSE41-NEXT: movdqa {{.*#+}} xmm12 = [2147483775,2147483775] +; SSE41-NEXT: movdqa %xmm12, %xmm10 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm10 +; SSE41-NEXT: movdqa %xmm12, %xmm13 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm13 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm13[0,0,2,2] +; SSE41-NEXT: pand %xmm10, %xmm0 +; SSE41-NEXT: por %xmm13, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm10 +; SSE41-NEXT: blendvpd %xmm0, %xmm6, %xmm10 +; SSE41-NEXT: movdqa %xmm7, %xmm0 +; SSE41-NEXT: pxor %xmm9, %xmm0 +; SSE41-NEXT: movdqa %xmm12, %xmm13 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm13 +; SSE41-NEXT: movdqa %xmm12, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm13, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm13 +; SSE41-NEXT: blendvpd %xmm0, %xmm7, %xmm13 +; SSE41-NEXT: movdqa %xmm4, %xmm0 +; SSE41-NEXT: pxor %xmm9, %xmm0 +; SSE41-NEXT: movdqa %xmm12, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm6 +; SSE41-NEXT: movdqa %xmm12, %xmm7 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm7 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm7[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm7, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm14 +; SSE41-NEXT: blendvpd %xmm0, %xmm4, %xmm14 +; SSE41-NEXT: movdqa %xmm5, %xmm0 +; SSE41-NEXT: pxor %xmm9, %xmm0 +; SSE41-NEXT: movdqa %xmm12, %xmm4 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm4 +; SSE41-NEXT: movdqa %xmm12, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm4, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm15 +; SSE41-NEXT: blendvpd %xmm0, %xmm5, %xmm15 +; SSE41-NEXT: movdqa %xmm2, %xmm0 +; SSE41-NEXT: pxor %xmm9, %xmm0 +; SSE41-NEXT: movdqa %xmm12, %xmm5 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm5 +; SSE41-NEXT: movdqa %xmm12, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm5, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm5 +; SSE41-NEXT: blendvpd %xmm0, %xmm2, %xmm5 +; SSE41-NEXT: movdqa %xmm3, %xmm0 +; SSE41-NEXT: pxor %xmm9, %xmm0 +; SSE41-NEXT: movdqa %xmm12, %xmm2 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm2 +; SSE41-NEXT: movdqa %xmm12, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm6 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm6[0,0,2,2] +; SSE41-NEXT: pand %xmm2, %xmm0 +; SSE41-NEXT: por %xmm6, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm6 +; SSE41-NEXT: blendvpd %xmm0, %xmm3, %xmm6 +; SSE41-NEXT: movdqa %xmm8, %xmm0 +; SSE41-NEXT: pxor %xmm9, %xmm0 +; SSE41-NEXT: movdqa %xmm12, %xmm2 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm2 +; SSE41-NEXT: movdqa %xmm12, %xmm3 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm2, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm11, %xmm7 +; SSE41-NEXT: blendvpd %xmm0, %xmm8, %xmm7 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: pxor %xmm9, %xmm0 +; SSE41-NEXT: movdqa %xmm12, %xmm2 +; SSE41-NEXT: pcmpeqd %xmm0, %xmm2 +; SSE41-NEXT: pcmpgtd %xmm0, %xmm12 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm12[0,0,2,2] +; SSE41-NEXT: pand %xmm2, %xmm0 +; SSE41-NEXT: por %xmm12, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm1, %xmm11 +; SSE41-NEXT: movapd {{.*#+}} xmm2 = [18446744073709551488,18446744073709551488] +; SSE41-NEXT: movapd %xmm11, %xmm1 +; SSE41-NEXT: xorpd %xmm9, %xmm1 +; SSE41-NEXT: movdqa {{.*#+}} xmm8 = [18446744071562067840,18446744071562067840] +; SSE41-NEXT: movapd %xmm1, %xmm4 +; SSE41-NEXT: pcmpeqd %xmm8, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm8, %xmm1 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,2,2] +; SSE41-NEXT: pand %xmm4, %xmm0 +; SSE41-NEXT: por %xmm1, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm4 +; SSE41-NEXT: blendvpd %xmm0, %xmm11, %xmm4 +; SSE41-NEXT: movapd %xmm7, %xmm1 +; SSE41-NEXT: xorpd %xmm9, %xmm1 +; SSE41-NEXT: movapd %xmm1, %xmm3 +; SSE41-NEXT: pcmpeqd %xmm8, %xmm3 +; SSE41-NEXT: pcmpgtd %xmm8, %xmm1 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,0,2,2] +; SSE41-NEXT: pand %xmm3, %xmm0 +; SSE41-NEXT: por %xmm1, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm1 +; SSE41-NEXT: blendvpd %xmm0, %xmm7, %xmm1 +; SSE41-NEXT: packssdw %xmm4, %xmm1 +; SSE41-NEXT: movapd %xmm6, %xmm3 +; SSE41-NEXT: xorpd %xmm9, %xmm3 +; SSE41-NEXT: movapd %xmm3, %xmm4 +; SSE41-NEXT: pcmpeqd %xmm8, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm8, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm4, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm3 +; SSE41-NEXT: blendvpd %xmm0, %xmm6, %xmm3 +; SSE41-NEXT: movapd %xmm5, %xmm4 +; SSE41-NEXT: xorpd %xmm9, %xmm4 +; SSE41-NEXT: movapd %xmm4, %xmm6 +; SSE41-NEXT: pcmpeqd %xmm8, %xmm6 +; SSE41-NEXT: pcmpgtd %xmm8, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm6, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm4 +; SSE41-NEXT: blendvpd %xmm0, %xmm5, %xmm4 +; SSE41-NEXT: packssdw %xmm3, %xmm4 +; SSE41-NEXT: packssdw %xmm4, %xmm1 +; SSE41-NEXT: movapd %xmm15, %xmm3 +; SSE41-NEXT: xorpd %xmm9, %xmm3 +; SSE41-NEXT: movapd %xmm3, %xmm4 +; SSE41-NEXT: pcmpeqd %xmm8, %xmm4 +; SSE41-NEXT: pcmpgtd %xmm8, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm4, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm3 +; SSE41-NEXT: blendvpd %xmm0, %xmm15, %xmm3 +; SSE41-NEXT: movapd %xmm14, %xmm4 +; SSE41-NEXT: xorpd %xmm9, %xmm4 +; SSE41-NEXT: movapd %xmm4, %xmm5 +; SSE41-NEXT: pcmpeqd %xmm8, %xmm5 +; SSE41-NEXT: pcmpgtd %xmm8, %xmm4 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm4[0,0,2,2] +; SSE41-NEXT: pand %xmm5, %xmm0 +; SSE41-NEXT: por %xmm4, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm4 +; SSE41-NEXT: blendvpd %xmm0, %xmm14, %xmm4 +; SSE41-NEXT: packssdw %xmm3, %xmm4 +; SSE41-NEXT: movapd %xmm13, %xmm3 +; SSE41-NEXT: xorpd %xmm9, %xmm3 +; SSE41-NEXT: movapd %xmm3, %xmm5 +; SSE41-NEXT: pcmpeqd %xmm8, %xmm5 +; SSE41-NEXT: pcmpgtd %xmm8, %xmm3 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2] +; SSE41-NEXT: pand %xmm5, %xmm0 +; SSE41-NEXT: por %xmm3, %xmm0 +; SSE41-NEXT: movapd %xmm2, %xmm3 +; SSE41-NEXT: blendvpd %xmm0, %xmm13, %xmm3 +; SSE41-NEXT: xorpd %xmm10, %xmm9 +; SSE41-NEXT: movapd %xmm9, %xmm5 +; SSE41-NEXT: pcmpeqd %xmm8, %xmm5 +; SSE41-NEXT: pcmpgtd %xmm8, %xmm9 +; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm9[0,0,2,2] +; SSE41-NEXT: pand %xmm5, %xmm0 +; SSE41-NEXT: por %xmm9, %xmm0 +; SSE41-NEXT: blendvpd %xmm0, %xmm10, %xmm2 +; SSE41-NEXT: packssdw %xmm3, %xmm2 +; SSE41-NEXT: packssdw %xmm2, %xmm4 +; SSE41-NEXT: packsswb %xmm4, %xmm1 +; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v16i64_v16i8: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm3, %xmm8 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [127,127] +; AVX1-NEXT: vextractf128 $1, %ymm2, %xmm9 +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm7 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm4 +; AVX1-NEXT: vpcmpgtq %xmm0, %xmm5, %xmm6 +; AVX1-NEXT: vblendvpd %xmm6, %xmm0, %xmm5, %xmm10 +; AVX1-NEXT: vpcmpgtq %xmm4, %xmm5, %xmm6 +; AVX1-NEXT: vblendvpd %xmm6, %xmm4, %xmm5, %xmm11 +; AVX1-NEXT: vpcmpgtq %xmm1, %xmm5, %xmm6 +; AVX1-NEXT: vblendvpd %xmm6, %xmm1, %xmm5, %xmm1 +; AVX1-NEXT: vpcmpgtq %xmm7, %xmm5, %xmm6 +; AVX1-NEXT: vblendvpd %xmm6, %xmm7, %xmm5, %xmm6 +; AVX1-NEXT: vpcmpgtq %xmm2, %xmm5, %xmm7 +; AVX1-NEXT: vblendvpd %xmm7, %xmm2, %xmm5, %xmm2 +; AVX1-NEXT: vpcmpgtq %xmm9, %xmm5, %xmm7 +; AVX1-NEXT: vblendvpd %xmm7, %xmm9, %xmm5, %xmm7 +; AVX1-NEXT: vpcmpgtq %xmm3, %xmm5, %xmm0 +; AVX1-NEXT: vblendvpd %xmm0, %xmm3, %xmm5, %xmm0 +; AVX1-NEXT: vpcmpgtq %xmm8, %xmm5, %xmm3 +; AVX1-NEXT: vblendvpd %xmm3, %xmm8, %xmm5, %xmm3 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [18446744073709551488,18446744073709551488] +; AVX1-NEXT: vpcmpgtq %xmm5, %xmm3, %xmm4 +; AVX1-NEXT: vblendvpd %xmm4, %xmm3, %xmm5, %xmm8 +; AVX1-NEXT: vpcmpgtq %xmm5, %xmm0, %xmm4 +; AVX1-NEXT: vblendvpd %xmm4, %xmm0, %xmm5, %xmm0 +; AVX1-NEXT: vpcmpgtq %xmm5, %xmm7, %xmm4 +; AVX1-NEXT: vblendvpd %xmm4, %xmm7, %xmm5, %xmm4 +; AVX1-NEXT: vpcmpgtq %xmm5, %xmm2, %xmm7 +; AVX1-NEXT: vblendvpd %xmm7, %xmm2, %xmm5, %xmm2 +; AVX1-NEXT: vpcmpgtq %xmm5, %xmm6, %xmm7 +; AVX1-NEXT: vblendvpd %xmm7, %xmm6, %xmm5, %xmm6 +; AVX1-NEXT: vpcmpgtq %xmm5, %xmm1, %xmm7 +; AVX1-NEXT: vblendvpd %xmm7, %xmm1, %xmm5, %xmm1 +; AVX1-NEXT: vpcmpgtq %xmm5, %xmm11, %xmm7 +; AVX1-NEXT: vblendvpd %xmm7, %xmm11, %xmm5, %xmm7 +; AVX1-NEXT: vpcmpgtq %xmm5, %xmm10, %xmm3 +; AVX1-NEXT: vblendvpd %xmm3, %xmm10, %xmm5, %xmm3 +; AVX1-NEXT: vpackssdw %xmm8, %xmm0, %xmm0 +; AVX1-NEXT: vpackssdw %xmm4, %xmm2, %xmm2 +; AVX1-NEXT: vpackssdw %xmm0, %xmm2, %xmm0 +; AVX1-NEXT: vpackssdw %xmm6, %xmm1, %xmm1 +; AVX1-NEXT: vpackssdw %xmm7, %xmm3, %xmm2 +; AVX1-NEXT: vpackssdw %xmm1, %xmm2, %xmm1 +; AVX1-NEXT: vpacksswb %xmm0, %xmm1, %xmm0 +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v16i64_v16i8: +; AVX2: # %bb.0: +; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm4 = [127,127,127,127] +; AVX2-NEXT: vpcmpgtq %ymm2, %ymm4, %ymm5 +; AVX2-NEXT: vblendvpd %ymm5, %ymm2, %ymm4, %ymm2 +; AVX2-NEXT: vpcmpgtq %ymm3, %ymm4, %ymm5 +; AVX2-NEXT: vblendvpd %ymm5, %ymm3, %ymm4, %ymm3 +; AVX2-NEXT: vpcmpgtq %ymm0, %ymm4, %ymm5 +; AVX2-NEXT: vblendvpd %ymm5, %ymm0, %ymm4, %ymm0 +; AVX2-NEXT: vpcmpgtq %ymm1, %ymm4, %ymm5 +; AVX2-NEXT: vblendvpd %ymm5, %ymm1, %ymm4, %ymm1 +; AVX2-NEXT: vpbroadcastq {{.*#+}} ymm4 = [18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488] +; AVX2-NEXT: vpcmpgtq %ymm4, %ymm1, %ymm5 +; AVX2-NEXT: vblendvpd %ymm5, %ymm1, %ymm4, %ymm1 +; AVX2-NEXT: vpcmpgtq %ymm4, %ymm0, %ymm5 +; AVX2-NEXT: vblendvpd %ymm5, %ymm0, %ymm4, %ymm0 +; AVX2-NEXT: vpackssdw %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vpcmpgtq %ymm4, %ymm3, %ymm1 +; AVX2-NEXT: vblendvpd %ymm1, %ymm3, %ymm4, %ymm1 +; AVX2-NEXT: vpcmpgtq %ymm4, %ymm2, %ymm3 +; AVX2-NEXT: vblendvpd %ymm3, %ymm2, %ymm4, %ymm2 +; AVX2-NEXT: vpackssdw %ymm1, %ymm2, %ymm1 +; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,2,1,3] +; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3] +; AVX2-NEXT: vpackssdw %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3] +; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vpacksswb %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512-LABEL: trunc_ssat_v16i64_v16i8: +; AVX512: # %bb.0: +; AVX512-NEXT: vpbroadcastq {{.*#+}} zmm2 = [127,127,127,127,127,127,127,127] +; AVX512-NEXT: vpminsq %zmm2, %zmm0, %zmm0 +; AVX512-NEXT: vpminsq %zmm2, %zmm1, %zmm1 +; AVX512-NEXT: vpbroadcastq {{.*#+}} zmm2 = [18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488,18446744073709551488] +; AVX512-NEXT: vpmaxsq %zmm2, %zmm1, %zmm1 +; AVX512-NEXT: vpmaxsq %zmm2, %zmm0, %zmm0 +; AVX512-NEXT: vpmovqd %zmm0, %ymm0 +; AVX512-NEXT: vpmovqd %zmm1, %ymm1 +; AVX512-NEXT: vinserti64x4 $1, %ymm1, %zmm0, %zmm0 +; AVX512-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512-NEXT: vzeroupper +; AVX512-NEXT: retq + %1 = icmp slt <16 x i64> %a0, <i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127> + %2 = select <16 x i1> %1, <16 x i64> %a0, <16 x i64> <i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127, i64 127> + %3 = icmp sgt <16 x i64> %2, <i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128> + %4 = select <16 x i1> %3, <16 x i64> %2, <16 x i64> <i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128, i64 -128> + %5 = trunc <16 x i64> %4 to <16 x i8> + ret <16 x i8> %5 +} + +define <8 x i8> @trunc_ssat_v8i32_v8i8(<8 x i32> %a0) { +; SSE-LABEL: trunc_ssat_v8i32_v8i8: +; SSE: # %bb.0: +; SSE-NEXT: packssdw %xmm1, %xmm0 +; SSE-NEXT: packsswb %xmm0, %xmm0 +; SSE-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v8i32_v8i8: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vpacksswb %xmm0, %xmm0, %xmm0 +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v8i32_v8i8: +; AVX2: # %bb.0: +; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpacksswb %xmm0, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512F-LABEL: trunc_ssat_v8i32_v8i8: +; AVX512F: # %bb.0: +; AVX512F-NEXT: vpbroadcastd {{.*#+}} ymm1 = [127,127,127,127,127,127,127,127] +; AVX512F-NEXT: vpminsd %ymm1, %ymm0, %ymm0 +; AVX512F-NEXT: vpbroadcastd {{.*#+}} ymm1 = [4294967168,4294967168,4294967168,4294967168,4294967168,4294967168,4294967168,4294967168] +; AVX512F-NEXT: vpmaxsd %ymm1, %ymm0, %ymm0 +; AVX512F-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512F-NEXT: vzeroupper +; AVX512F-NEXT: retq +; +; AVX512VL-LABEL: trunc_ssat_v8i32_v8i8: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpminsd {{.*}}(%rip){1to8}, %ymm0, %ymm0 +; AVX512VL-NEXT: vpmaxsd {{.*}}(%rip){1to8}, %ymm0, %ymm0 +; AVX512VL-NEXT: vpmovdb %ymm0, %xmm0 +; AVX512VL-NEXT: vzeroupper +; AVX512VL-NEXT: retq +; +; AVX512BW-LABEL: trunc_ssat_v8i32_v8i8: +; AVX512BW: # %bb.0: +; AVX512BW-NEXT: vpbroadcastd {{.*#+}} ymm1 = [127,127,127,127,127,127,127,127] +; AVX512BW-NEXT: vpminsd %ymm1, %ymm0, %ymm0 +; AVX512BW-NEXT: vpbroadcastd {{.*#+}} ymm1 = [4294967168,4294967168,4294967168,4294967168,4294967168,4294967168,4294967168,4294967168] +; AVX512BW-NEXT: vpmaxsd %ymm1, %ymm0, %ymm0 +; AVX512BW-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512BW-NEXT: vzeroupper +; AVX512BW-NEXT: retq +; +; AVX512BWVL-LABEL: trunc_ssat_v8i32_v8i8: +; AVX512BWVL: # %bb.0: +; AVX512BWVL-NEXT: vpminsd {{.*}}(%rip){1to8}, %ymm0, %ymm0 +; AVX512BWVL-NEXT: vpmaxsd {{.*}}(%rip){1to8}, %ymm0, %ymm0 +; AVX512BWVL-NEXT: vpmovdb %ymm0, %xmm0 +; AVX512BWVL-NEXT: vzeroupper +; AVX512BWVL-NEXT: retq + %1 = icmp slt <8 x i32> %a0, <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127> + %2 = select <8 x i1> %1, <8 x i32> %a0, <8 x i32> <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127> + %3 = icmp sgt <8 x i32> %2, <i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128> + %4 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> <i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128> + %5 = trunc <8 x i32> %4 to <8 x i8> + ret <8 x i8> %5 +} + +define void @trunc_ssat_v8i32_v8i8_store(<8 x i32> %a0, <8 x i8> *%p1) { +; SSE-LABEL: trunc_ssat_v8i32_v8i8_store: +; SSE: # %bb.0: +; SSE-NEXT: packssdw %xmm1, %xmm0 +; SSE-NEXT: packsswb %xmm0, %xmm0 +; SSE-NEXT: movq %xmm0, (%rdi) +; SSE-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v8i32_v8i8_store: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vpacksswb %xmm0, %xmm0, %xmm0 +; AVX1-NEXT: vmovq %xmm0, (%rdi) +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v8i32_v8i8_store: +; AVX2: # %bb.0: +; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpacksswb %xmm0, %xmm0, %xmm0 +; AVX2-NEXT: vmovq %xmm0, (%rdi) +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512F-LABEL: trunc_ssat_v8i32_v8i8_store: +; AVX512F: # %bb.0: +; AVX512F-NEXT: vpbroadcastd {{.*#+}} ymm1 = [127,127,127,127,127,127,127,127] +; AVX512F-NEXT: vpminsd %ymm1, %ymm0, %ymm0 +; AVX512F-NEXT: vpbroadcastd {{.*#+}} ymm1 = [4294967168,4294967168,4294967168,4294967168,4294967168,4294967168,4294967168,4294967168] +; AVX512F-NEXT: vpmaxsd %ymm1, %ymm0, %ymm0 +; AVX512F-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512F-NEXT: vmovq %xmm0, (%rdi) +; AVX512F-NEXT: vzeroupper +; AVX512F-NEXT: retq +; +; AVX512VL-LABEL: trunc_ssat_v8i32_v8i8_store: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsdb %ymm0, (%rdi) +; AVX512VL-NEXT: vzeroupper +; AVX512VL-NEXT: retq +; +; AVX512BW-LABEL: trunc_ssat_v8i32_v8i8_store: +; AVX512BW: # %bb.0: +; AVX512BW-NEXT: vpbroadcastd {{.*#+}} ymm1 = [127,127,127,127,127,127,127,127] +; AVX512BW-NEXT: vpminsd %ymm1, %ymm0, %ymm0 +; AVX512BW-NEXT: vpbroadcastd {{.*#+}} ymm1 = [4294967168,4294967168,4294967168,4294967168,4294967168,4294967168,4294967168,4294967168] +; AVX512BW-NEXT: vpmaxsd %ymm1, %ymm0, %ymm0 +; AVX512BW-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512BW-NEXT: vmovq %xmm0, (%rdi) +; AVX512BW-NEXT: vzeroupper +; AVX512BW-NEXT: retq +; +; AVX512BWVL-LABEL: trunc_ssat_v8i32_v8i8_store: +; AVX512BWVL: # %bb.0: +; AVX512BWVL-NEXT: vpmovsdb %ymm0, (%rdi) +; AVX512BWVL-NEXT: vzeroupper +; AVX512BWVL-NEXT: retq + %1 = icmp slt <8 x i32> %a0, <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127> + %2 = select <8 x i1> %1, <8 x i32> %a0, <8 x i32> <i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127, i32 127> + %3 = icmp sgt <8 x i32> %2, <i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128> + %4 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> <i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128, i32 -128> + %5 = trunc <8 x i32> %4 to <8 x i8> + store <8 x i8> %5, <8 x i8> *%p1 + ret void +} + +define <16 x i8> @trunc_ssat_v16i32_v16i8(<16 x i32> %a0) { +; SSE-LABEL: trunc_ssat_v16i32_v16i8: +; SSE: # %bb.0: +; SSE-NEXT: packssdw %xmm3, %xmm2 +; SSE-NEXT: packssdw %xmm1, %xmm0 +; SSE-NEXT: packsswb %xmm2, %xmm0 +; SSE-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v16i32_v16i8: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-NEXT: vpackssdw %xmm2, %xmm1, %xmm1 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 +; AVX1-NEXT: vpackssdw %xmm2, %xmm0, %xmm0 +; AVX1-NEXT: vpacksswb %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v16i32_v16i8: +; AVX2: # %bb.0: +; AVX2-NEXT: vpackssdw %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3] +; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vpacksswb %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512-LABEL: trunc_ssat_v16i32_v16i8: +; AVX512: # %bb.0: +; AVX512-NEXT: vpmovsdb %zmm0, %xmm0 +; AVX512-NEXT: vzeroupper +; AVX512-NEXT: retq + %1 = icmp slt <16 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> + %2 = select <16 x i1> %1, <16 x i32> %a0, <16 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> + %3 = icmp sgt <16 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> + %4 = select <16 x i1> %3, <16 x i32> %2, <16 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> + %5 = trunc <16 x i32> %4 to <16 x i8> + ret <16 x i8> %5 +} + +define <16 x i8> @trunc_ssat_v16i16_v16i8(<16 x i16> %a0) { +; SSE-LABEL: trunc_ssat_v16i16_v16i8: +; SSE: # %bb.0: +; SSE-NEXT: packsswb %xmm1, %xmm0 +; SSE-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v16i16_v16i8: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 +; AVX1-NEXT: vpacksswb %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vzeroupper +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v16i16_v16i8: +; AVX2: # %bb.0: +; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1 +; AVX2-NEXT: vpacksswb %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512F-LABEL: trunc_ssat_v16i16_v16i8: +; AVX512F: # %bb.0: +; AVX512F-NEXT: vpminsw {{.*}}(%rip), %ymm0, %ymm0 +; AVX512F-NEXT: vpmaxsw {{.*}}(%rip), %ymm0, %ymm0 +; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero +; AVX512F-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512F-NEXT: vzeroupper +; AVX512F-NEXT: retq +; +; AVX512VL-LABEL: trunc_ssat_v16i16_v16i8: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpminsw {{.*}}(%rip), %ymm0, %ymm0 +; AVX512VL-NEXT: vpmaxsw {{.*}}(%rip), %ymm0, %ymm0 +; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero +; AVX512VL-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512VL-NEXT: vzeroupper +; AVX512VL-NEXT: retq +; +; AVX512BW-LABEL: trunc_ssat_v16i16_v16i8: +; AVX512BW: # %bb.0: +; AVX512BW-NEXT: vpminsw {{.*}}(%rip), %ymm0, %ymm0 +; AVX512BW-NEXT: vpmaxsw {{.*}}(%rip), %ymm0, %ymm0 +; AVX512BW-NEXT: vpmovwb %zmm0, %ymm0 +; AVX512BW-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0 +; AVX512BW-NEXT: vzeroupper +; AVX512BW-NEXT: retq +; +; AVX512BWVL-LABEL: trunc_ssat_v16i16_v16i8: +; AVX512BWVL: # %bb.0: +; AVX512BWVL-NEXT: vpmovswb %ymm0, %xmm0 +; AVX512BWVL-NEXT: vzeroupper +; AVX512BWVL-NEXT: retq + %1 = icmp slt <16 x i16> %a0, <i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127> + %2 = select <16 x i1> %1, <16 x i16> %a0, <16 x i16> <i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127> + %3 = icmp sgt <16 x i16> %2, <i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128> + %4 = select <16 x i1> %3, <16 x i16> %2, <16 x i16> <i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128> + %5 = trunc <16 x i16> %4 to <16 x i8> + ret <16 x i8> %5 +} + +define <32 x i8> @trunc_ssat_v32i16_v32i8(<32 x i16> %a0) { +; SSE-LABEL: trunc_ssat_v32i16_v32i8: +; SSE: # %bb.0: +; SSE-NEXT: packsswb %xmm1, %xmm0 +; SSE-NEXT: packsswb %xmm3, %xmm2 +; SSE-NEXT: movdqa %xmm2, %xmm1 +; SSE-NEXT: retq +; +; AVX1-LABEL: trunc_ssat_v32i16_v32i8: +; AVX1: # %bb.0: +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-NEXT: vpacksswb %xmm2, %xmm1, %xmm1 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 +; AVX1-NEXT: vpacksswb %xmm2, %xmm0, %xmm0 +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: trunc_ssat_v32i16_v32i8: +; AVX2: # %bb.0: +; AVX2-NEXT: vpacksswb %ymm1, %ymm0, %ymm0 +; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3] +; AVX2-NEXT: retq +; +; AVX512F-LABEL: trunc_ssat_v32i16_v32i8: +; AVX512F: # %bb.0: +; AVX512F-NEXT: vmovdqa {{.*#+}} ymm2 = [127,127,127,127,127,127,127,127,127,127,127,127,127,127,127,127] +; AVX512F-NEXT: vpminsw %ymm2, %ymm0, %ymm0 +; AVX512F-NEXT: vpminsw %ymm2, %ymm1, %ymm1 +; AVX512F-NEXT: vmovdqa {{.*#+}} ymm2 = [65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408] +; AVX512F-NEXT: vpmaxsw %ymm2, %ymm1, %ymm1 +; AVX512F-NEXT: vpmaxsw %ymm2, %ymm0, %ymm0 +; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero +; AVX512F-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero +; AVX512F-NEXT: vpmovdb %zmm1, %xmm1 +; AVX512F-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0 +; AVX512F-NEXT: retq +; +; AVX512VL-LABEL: trunc_ssat_v32i16_v32i8: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm2 = [127,127,127,127,127,127,127,127,127,127,127,127,127,127,127,127] +; AVX512VL-NEXT: vpminsw %ymm2, %ymm0, %ymm0 +; AVX512VL-NEXT: vpminsw %ymm2, %ymm1, %ymm1 +; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm2 = [65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408,65408] +; AVX512VL-NEXT: vpmaxsw %ymm2, %ymm1, %ymm1 +; AVX512VL-NEXT: vpmaxsw %ymm2, %ymm0, %ymm0 +; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero +; AVX512VL-NEXT: vpmovdb %zmm0, %xmm0 +; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero +; AVX512VL-NEXT: vpmovdb %zmm1, %xmm1 +; AVX512VL-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0 +; AVX512VL-NEXT: retq +; +; AVX512BW-LABEL: trunc_ssat_v32i16_v32i8: +; AVX512BW: # %bb.0: +; AVX512BW-NEXT: vpmovswb %zmm0, %ymm0 +; AVX512BW-NEXT: retq +; +; AVX512BWVL-LABEL: trunc_ssat_v32i16_v32i8: +; AVX512BWVL: # %bb.0: +; AVX512BWVL-NEXT: vpmovswb %zmm0, %ymm0 +; AVX512BWVL-NEXT: retq + %1 = icmp slt <32 x i16> %a0, <i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127> + %2 = select <32 x i1> %1, <32 x i16> %a0, <32 x i16> <i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127, i16 127> + %3 = icmp sgt <32 x i16> %2, <i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128> + %4 = select <32 x i1> %3, <32 x i16> %2, <32 x i16> <i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128, i16 -128> + %5 = trunc <32 x i16> %4 to <32 x i8> + ret <32 x i8> %5 +} |
