Index: llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp =================================================================== --- llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp +++ llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp @@ -1312,9 +1312,16 @@ } SDValue SelectionDAG::getAnyExtOrTrunc(SDValue Op, const SDLoc &DL, EVT VT) { - return VT.bitsGT(Op.getValueType()) ? - getNode(ISD::ANY_EXTEND, DL, VT, Op) : - getNode(ISD::TRUNCATE, DL, VT, Op); + if (VT.bitsGT(Op.getValueType())) { + ISD::NodeType ExtendKind = ISD::ANY_EXTEND; + if (ConstantSDNode *C = isConstOrConstSplat(Op)) { + if (C->getAPIntValue().isNegative()) + ExtendKind = ISD::SIGN_EXTEND; + } + return getNode(ExtendKind, DL, VT, Op); + } + + return getNode(ISD::TRUNCATE, DL, VT, Op); } SDValue SelectionDAG::getSExtOrTrunc(SDValue Op, const SDLoc &DL, EVT VT) { Index: llvm/test/CodeGen/AArch64/sve-vector-splat.ll =================================================================== --- llvm/test/CodeGen/AArch64/sve-vector-splat.ll +++ llvm/test/CodeGen/AArch64/sve-vector-splat.ll @@ -116,6 +116,16 @@ ret %splat } +define @sve_splat_8xi8_imm() { +; CHECK-LABEL: sve_splat_8xi8_imm: +; CHECK: // %bb.0: +; CHECK-NEXT: mov z0.h, #-1 // =0xffffffffffffffff +; CHECK-NEXT: ret + %ins = insertelement undef, i8 -1, i32 0 + %splat = shufflevector %ins, undef, zeroinitializer + ret %splat +} + define @sve_splat_2xi16(i16 %val) { ; CHECK-LABEL: sve_splat_2xi16: ; CHECK: // %bb.0: @@ -137,6 +147,16 @@ ret %splat } +define @sve_splat_4xi16_imm() { +; CHECK-LABEL: sve_splat_4xi16_imm: +; CHECK: // %bb.0: +; CHECK-NEXT: mov z0.s, #-1 // =0xffffffffffffffff +; CHECK-NEXT: ret + %ins = insertelement undef, i16 -1, i32 0 + %splat = shufflevector %ins, undef, zeroinitializer + ret %splat +} + define @sve_splat_2xi32(i32 %val) { ; CHECK-LABEL: sve_splat_2xi32: ; CHECK: // %bb.0: @@ -148,6 +168,16 @@ ret %splat } +define @sve_splat_2xi32_imm() { +; CHECK-LABEL: sve_splat_2xi32_imm: +; CHECK: // %bb.0: +; CHECK-NEXT: mov z0.d, #-1 // =0xffffffffffffffff +; CHECK-NEXT: ret + %ins = insertelement undef, i32 -1, i32 0 + %splat = shufflevector %ins, undef, zeroinitializer + ret %splat +} + ;; Widen/split splats of wide vector types. define @sve_splat_1xi32(i32 %val) { @@ -497,9 +527,9 @@ define @splat_nxv2f64_imm_out_of_range() { ; CHECK-LABEL: splat_nxv2f64_imm_out_of_range: ; CHECK: // %bb.0: -; CHECK-NEXT: adrp x8, .LCPI47_0 +; CHECK-NEXT: adrp x8, .LCPI50_0 ; CHECK-NEXT: ptrue p0.d -; CHECK-NEXT: add x8, x8, :lo12:.LCPI47_0 +; CHECK-NEXT: add x8, x8, :lo12:.LCPI50_0 ; CHECK-NEXT: ld1rd { z0.d }, p0/z, [x8] ; CHECK-NEXT: ret %1 = insertelement undef, double 3.33, i32 0 Index: llvm/test/CodeGen/PowerPC/and-mask.ll =================================================================== --- llvm/test/CodeGen/PowerPC/and-mask.ll +++ llvm/test/CodeGen/PowerPC/and-mask.ll @@ -6,7 +6,7 @@ define i32 @test1(i32 %a) { ; CHECK-LABEL: test1: ; CHECK: # %bb.0: -; CHECK-NEXT: rlwinm 3, 3, 0, 0, 30 +; CHECK-NEXT: rldicr 3, 3, 0, 62 ; CHECK-NEXT: blr %and = and i32 %a, -2 ret i32 %and Index: llvm/test/CodeGen/PowerPC/bool-math.ll =================================================================== --- llvm/test/CodeGen/PowerPC/bool-math.ll +++ llvm/test/CodeGen/PowerPC/bool-math.ll @@ -5,9 +5,9 @@ define i32 @sub_zext_cmp_mask_same_size_result(i32 %x) { ; CHECK-LABEL: sub_zext_cmp_mask_same_size_result: ; CHECK: # %bb.0: +; CHECK-NEXT: li 4, -28 ; CHECK-NEXT: clrldi 3, 3, 63 -; CHECK-NEXT: ori 3, 3, 65508 -; CHECK-NEXT: oris 3, 3, 65535 +; CHECK-NEXT: or 3, 3, 4 ; CHECK-NEXT: blr %a = and i32 %x, 1 %c = icmp eq i32 %a, 0 Index: llvm/test/CodeGen/PowerPC/combine_ext_trunc.ll =================================================================== --- llvm/test/CodeGen/PowerPC/combine_ext_trunc.ll +++ llvm/test/CodeGen/PowerPC/combine_ext_trunc.ll @@ -8,9 +8,6 @@ define i32 @pattern1(i32 %x, i32 %y){ ; CHECK-LABEL: pattern1: ; CHECK: # %bb.0: -; CHECK-NEXT: xori 5, 4, 65535 -; CHECK-NEXT: xoris 5, 5, 65535 -; CHECK-NEXT: and 3, 3, 5 ; CHECK-NEXT: or 3, 3, 4 ; CHECK-NEXT: blr %a = xor i32 %y, -1 @@ -23,9 +20,6 @@ define i32 @pattern2(i32 %x, i32 %y){ ; CHECK-LABEL: pattern2: ; CHECK: # %bb.0: -; CHECK-NEXT: xori 5, 4, 65535 -; CHECK-NEXT: xoris 5, 5, 65535 -; CHECK-NEXT: and 3, 5, 3 ; CHECK-NEXT: or 3, 3, 4 ; CHECK-NEXT: blr %a = xor i32 %y, -1 @@ -38,10 +32,8 @@ define i32 @pattern3(i1 %cond, i32 %x) { ; CHECK-LABEL: pattern3: ; CHECK: # %bb.0: -; CHECK-NEXT: li 5, -1 -; CHECK-NEXT: andi. 3, 3, 1 -; CHECK-NEXT: rldic 3, 5, 0, 32 -; CHECK-NEXT: iselgt 3, 0, 3 +; CHECK-NEXT: clrldi 3, 3, 63 +; CHECK-NEXT: addi 3, 3, -1 ; CHECK-NEXT: and 3, 3, 4 ; CHECK-NEXT: blr %sel = select i1 %cond, i32 0, i32 -1 @@ -53,11 +45,8 @@ define i32 @pattern4(i1 %cond, i32 %x) { ; CHECK-LABEL: pattern4: ; CHECK: # %bb.0: -; CHECK-NEXT: li 5, -1 -; CHECK-NEXT: andi. 3, 3, 1 -; CHECK-NEXT: rldic 3, 5, 0, 32 -; CHECK-NEXT: li 5, 0 -; CHECK-NEXT: iselgt 3, 3, 5 +; CHECK-NEXT: clrldi 3, 3, 63 +; CHECK-NEXT: neg 3, 3 ; CHECK-NEXT: or 3, 4, 3 ; CHECK-NEXT: blr %sel = select i1 %cond, i32 -1, i32 0 Index: llvm/test/CodeGen/PowerPC/shift-cmp.ll =================================================================== --- llvm/test/CodeGen/PowerPC/shift-cmp.ll +++ llvm/test/CodeGen/PowerPC/shift-cmp.ll @@ -28,7 +28,7 @@ ; CHECK-LABEL: and_ncmp_variable_power_of_two: ; CHECK: # %bb.0: ; CHECK-NEXT: srw 3, 3, 4 -; CHECK-NEXT: xori 3, 3, 1 +; CHECK-NEXT: not 3, 3 ; CHECK-NEXT: blr %shl = shl i32 1, %y %and = and i32 %x, %shl @@ -40,7 +40,7 @@ ; CHECK-LABEL: and_ncmp_variable_power_of_two_64: ; CHECK: # %bb.0: ; CHECK-NEXT: srd 3, 3, 4 -; CHECK-NEXT: xori 3, 3, 1 +; CHECK-NEXT: not 3, 3 ; CHECK-NEXT: blr %shl = shl i64 1, %y %and = and i64 %x, %shl Index: llvm/test/CodeGen/X86/min-legal-vector-width.ll =================================================================== --- llvm/test/CodeGen/X86/min-legal-vector-width.ll +++ llvm/test/CodeGen/X86/min-legal-vector-width.ll @@ -1720,11 +1720,11 @@ ; CHECK-AVX512-NEXT: vpsrlw $8, %ymm1, %ymm1 ; CHECK-AVX512-NEXT: vpackuswb %ymm2, %ymm1, %ymm1 ; CHECK-AVX512-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; CHECK-AVX512-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2 +; CHECK-AVX512-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2 ; CHECK-AVX512-NEXT: vmovdqa {{.*#+}} ymm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; CHECK-AVX512-NEXT: vpand %ymm3, %ymm2, %ymm2 ; CHECK-AVX512-NEXT: vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; CHECK-AVX512-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0 +; CHECK-AVX512-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0 ; CHECK-AVX512-NEXT: vpand %ymm3, %ymm0, %ymm0 ; CHECK-AVX512-NEXT: vpackuswb %ymm2, %ymm0, %ymm0 ; CHECK-AVX512-NEXT: vpor %ymm1, %ymm0, %ymm0 @@ -1733,9 +1733,9 @@ ; CHECK-VBMI-LABEL: constant_rotate_v32i8: ; CHECK-VBMI: # %bb.0: ; CHECK-VBMI-NEXT: vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; CHECK-VBMI-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm1, %ymm1 +; CHECK-VBMI-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm1, %ymm1 ; CHECK-VBMI-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; CHECK-VBMI-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2 +; CHECK-VBMI-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2 ; CHECK-VBMI-NEXT: vmovdqa {{.*#+}} ymm3 = [0,2,4,6,8,10,12,14,32,34,36,38,40,42,44,46,16,18,20,22,24,26,28,30,48,50,52,54,56,58,60,62] ; CHECK-VBMI-NEXT: vpermi2b %ymm1, %ymm2, %ymm3 ; CHECK-VBMI-NEXT: vpxor %xmm1, %xmm1, %xmm1 Index: llvm/test/CodeGen/X86/vector-fshl-256.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshl-256.ll +++ llvm/test/CodeGen/X86/vector-fshl-256.ll @@ -2018,12 +2018,12 @@ ; AVX1: # %bb.0: ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15] -; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,128,64,32,16,8,4,2] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,65408,64,32,16,8,4,2] ; AVX1-NEXT: vpmullw %xmm4, %xmm3, %xmm3 ; AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [255,255,255,255,255,255,255,255] ; AVX1-NEXT: vpand %xmm5, %xmm3, %xmm3 ; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm2 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero -; AVX1-NEXT: vmovdqa {{.*#+}} xmm6 = [1,2,4,8,16,32,64,128] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm6 = [1,2,4,8,16,32,64,65408] ; AVX1-NEXT: vpmullw %xmm6, %xmm2, %xmm2 ; AVX1-NEXT: vpand %xmm5, %xmm2, %xmm2 ; AVX1-NEXT: vpackuswb %xmm3, %xmm2, %xmm2 Index: llvm/test/CodeGen/X86/vector-fshl-512.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshl-512.ll +++ llvm/test/CodeGen/X86/vector-fshl-512.ll @@ -1052,13 +1052,13 @@ ; AVX512F: # %bb.0: ; AVX512F-NEXT: vextracti64x4 $1, %zmm0, %ymm2 ; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [1,128,64,32,16,8,4,2,1,128,64,32,16,8,4,2] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [1,65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2] ; AVX512F-NEXT: # ymm4 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm4, %ymm3, %ymm3 ; AVX512F-NEXT: vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512F-NEXT: vpand %ymm5, %ymm3, %ymm3 ; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512F-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm6, %ymm2, %ymm2 ; AVX512F-NEXT: vpand %ymm5, %ymm2, %ymm2 @@ -1104,13 +1104,13 @@ ; AVX512VL: # %bb.0: ; AVX512VL-NEXT: vextracti64x4 $1, %zmm0, %ymm2 ; AVX512VL-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [1,128,64,32,16,8,4,2,1,128,64,32,16,8,4,2] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [1,65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2] ; AVX512VL-NEXT: # ymm4 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm4, %ymm3, %ymm3 ; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VL-NEXT: vpand %ymm5, %ymm3, %ymm3 ; AVX512VL-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512VL-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm6, %ymm2, %ymm2 ; AVX512VL-NEXT: vpand %ymm5, %ymm2, %ymm2 @@ -1155,11 +1155,11 @@ ; AVX512BW-LABEL: constant_funnnel_v64i8: ; AVX512BW: # %bb.0: ; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512BW-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512BW-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512BW-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512BW-NEXT: vpsrlw $1, %zmm1, %zmm1 @@ -1178,11 +1178,11 @@ ; AVX512VBMI2-LABEL: constant_funnnel_v64i8: ; AVX512VBMI2: # %bb.0: ; AVX512VBMI2-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VBMI2-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VBMI2-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VBMI2-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpsrlw $1, %zmm1, %zmm1 @@ -1201,11 +1201,11 @@ ; AVX512VLBW-LABEL: constant_funnnel_v64i8: ; AVX512VLBW: # %bb.0: ; AVX512VLBW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VLBW-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLBW-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VLBW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpsrlw $1, %zmm1, %zmm1 @@ -1224,11 +1224,11 @@ ; AVX512VLVBMI2-LABEL: constant_funnnel_v64i8: ; AVX512VLVBMI2: # %bb.0: ; AVX512VLVBMI2-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VLVBMI2-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLVBMI2-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VLVBMI2-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpsrlw $1, %zmm1, %zmm1 Index: llvm/test/CodeGen/X86/vector-fshl-rot-128.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshl-rot-128.ll +++ llvm/test/CodeGen/X86/vector-fshl-rot-128.ll @@ -1622,7 +1622,7 @@ ; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255] ; SSE41-NEXT: pand %xmm3, %xmm2 ; SSE41-NEXT: pmovzxbw {{.*#+}} xmm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero -; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [1,2,4,8,16,32,64,128] +; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [1,2,4,8,16,32,64,65408] ; SSE41-NEXT: pmullw %xmm1, %xmm4 ; SSE41-NEXT: pand %xmm3, %xmm4 ; SSE41-NEXT: packuswb %xmm2, %xmm4 Index: llvm/test/CodeGen/X86/vector-fshl-rot-256.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshl-rot-256.ll +++ llvm/test/CodeGen/X86/vector-fshl-rot-256.ll @@ -1310,11 +1310,11 @@ ; AVX1-NEXT: vpsrlw $8, %xmm7, %xmm7 ; AVX1-NEXT: vpackuswb %xmm3, %xmm7, %xmm3 ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15] -; AVX1-NEXT: vmovdqa {{.*#+}} xmm7 = [1,128,64,32,16,8,4,2] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm7 = [1,65408,64,32,16,8,4,2] ; AVX1-NEXT: vpmullw %xmm7, %xmm1, %xmm1 ; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255] ; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1 -; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,2,4,8,16,32,64,128] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,2,4,8,16,32,64,65408] ; AVX1-NEXT: vpmullw %xmm4, %xmm5, %xmm5 ; AVX1-NEXT: vpand %xmm2, %xmm5, %xmm5 ; AVX1-NEXT: vpackuswb %xmm1, %xmm5, %xmm1 Index: llvm/test/CodeGen/X86/vector-fshl-rot-512.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshl-rot-512.ll +++ llvm/test/CodeGen/X86/vector-fshl-rot-512.ll @@ -790,13 +790,13 @@ ; AVX512F-NEXT: vpackuswb %ymm4, %ymm2, %ymm2 ; AVX512F-NEXT: vinserti64x4 $1, %ymm3, %zmm2, %zmm2 ; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [1,128,64,32,16,8,4,2,1,128,64,32,16,8,4,2] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [1,65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2] ; AVX512F-NEXT: # ymm4 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm4, %ymm3, %ymm3 ; AVX512F-NEXT: vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512F-NEXT: vpand %ymm5, %ymm3, %ymm3 ; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512F-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm6, %ymm1, %ymm1 ; AVX512F-NEXT: vpand %ymm5, %ymm1, %ymm1 @@ -836,13 +836,13 @@ ; AVX512VL-NEXT: vpackuswb %ymm4, %ymm2, %ymm2 ; AVX512VL-NEXT: vinserti64x4 $1, %ymm3, %zmm2, %zmm2 ; AVX512VL-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [1,128,64,32,16,8,4,2,1,128,64,32,16,8,4,2] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [1,65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2] ; AVX512VL-NEXT: # ymm4 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm4, %ymm3, %ymm3 ; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VL-NEXT: vpand %ymm5, %ymm3, %ymm3 ; AVX512VL-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512VL-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm6, %ymm1, %ymm1 ; AVX512VL-NEXT: vpand %ymm5, %ymm1, %ymm1 @@ -869,11 +869,11 @@ ; AVX512BW-NEXT: vpsrlw $8, %zmm1, %zmm1 ; AVX512BW-NEXT: vpackuswb %zmm2, %zmm1, %zmm1 ; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512BW-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512BW-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512BW-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512BW-NEXT: vporq %zmm1, %zmm0, %zmm0 @@ -890,11 +890,11 @@ ; AVX512VLBW-NEXT: vpsrlw $8, %zmm1, %zmm1 ; AVX512VLBW-NEXT: vpackuswb %zmm2, %zmm1, %zmm1 ; AVX512VLBW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VLBW-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLBW-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VLBW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vporq %zmm1, %zmm0, %zmm0 @@ -911,11 +911,11 @@ ; AVX512VBMI2-NEXT: vpsrlw $8, %zmm1, %zmm1 ; AVX512VBMI2-NEXT: vpackuswb %zmm2, %zmm1, %zmm1 ; AVX512VBMI2-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VBMI2-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VBMI2-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VBMI2-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vporq %zmm1, %zmm0, %zmm0 @@ -932,11 +932,11 @@ ; AVX512VLVBMI2-NEXT: vpsrlw $8, %zmm1, %zmm1 ; AVX512VLVBMI2-NEXT: vpackuswb %zmm2, %zmm1, %zmm1 ; AVX512VLVBMI2-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VLVBMI2-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLVBMI2-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VLVBMI2-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vporq %zmm1, %zmm0, %zmm0 Index: llvm/test/CodeGen/X86/vector-fshr-256.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshr-256.ll +++ llvm/test/CodeGen/X86/vector-fshr-256.ll @@ -1728,12 +1728,12 @@ ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2 ; AVX1-NEXT: vpaddb %xmm2, %xmm2, %xmm2 ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15] -; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [128,1,2,4,8,16,32,64] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [65408,1,2,4,8,16,32,64] ; AVX1-NEXT: vpmullw %xmm4, %xmm3, %xmm3 ; AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [255,255,255,255,255,255,255,255] ; AVX1-NEXT: vpand %xmm5, %xmm3, %xmm3 ; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm2 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero -; AVX1-NEXT: vmovdqa {{.*#+}} xmm6 = [128,64,32,16,8,4,2,1] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm6 = [65408,64,32,16,8,4,2,1] ; AVX1-NEXT: vpmullw %xmm6, %xmm2, %xmm2 ; AVX1-NEXT: vpand %xmm5, %xmm2, %xmm2 ; AVX1-NEXT: vpackuswb %xmm3, %xmm2, %xmm2 Index: llvm/test/CodeGen/X86/vector-fshr-512.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshr-512.ll +++ llvm/test/CodeGen/X86/vector-fshr-512.ll @@ -1066,13 +1066,13 @@ ; AVX512F-NEXT: vextracti64x4 $1, %zmm0, %ymm2 ; AVX512F-NEXT: vpaddb %ymm2, %ymm2, %ymm2 ; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [128,1,2,4,8,16,32,64,128,1,2,4,8,16,32,64] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [65408,1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64] ; AVX512F-NEXT: # ymm4 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm4, %ymm3, %ymm3 ; AVX512F-NEXT: vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512F-NEXT: vpand %ymm5, %ymm3, %ymm3 ; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [128,64,32,16,8,4,2,1,128,64,32,16,8,4,2,1] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2,1] ; AVX512F-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm6, %ymm2, %ymm2 ; AVX512F-NEXT: vpand %ymm5, %ymm2, %ymm2 @@ -1115,13 +1115,13 @@ ; AVX512VL-NEXT: vextracti64x4 $1, %zmm0, %ymm2 ; AVX512VL-NEXT: vpaddb %ymm2, %ymm2, %ymm2 ; AVX512VL-NEXT: vpunpckhbw {{.*#+}} ymm3 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [128,1,2,4,8,16,32,64,128,1,2,4,8,16,32,64] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm4 = [65408,1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64] ; AVX512VL-NEXT: # ymm4 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm4, %ymm3, %ymm3 ; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VL-NEXT: vpand %ymm5, %ymm3, %ymm3 ; AVX512VL-NEXT: vpunpcklbw {{.*#+}} ymm2 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [128,64,32,16,8,4,2,1,128,64,32,16,8,4,2,1] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2,1] ; AVX512VL-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm6, %ymm2, %ymm2 ; AVX512VL-NEXT: vpand %ymm5, %ymm2, %ymm2 @@ -1150,11 +1150,11 @@ ; AVX512BW-NEXT: vpackuswb %zmm3, %zmm1, %zmm1 ; AVX512BW-NEXT: vpaddb %zmm0, %zmm0, %zmm0 ; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512BW-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512BW-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512BW-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512BW-NEXT: vporq %zmm1, %zmm0, %zmm0 @@ -1172,11 +1172,11 @@ ; AVX512VBMI2-NEXT: vpackuswb %zmm3, %zmm1, %zmm1 ; AVX512VBMI2-NEXT: vpaddb %zmm0, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VBMI2-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VBMI2-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VBMI2-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vporq %zmm1, %zmm0, %zmm0 @@ -1194,11 +1194,11 @@ ; AVX512VLBW-NEXT: vpackuswb %zmm3, %zmm1, %zmm1 ; AVX512VLBW-NEXT: vpaddb %zmm0, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VLBW-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLBW-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VLBW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vporq %zmm1, %zmm0, %zmm0 @@ -1216,11 +1216,11 @@ ; AVX512VLVBMI2-NEXT: vpackuswb %zmm3, %zmm1, %zmm1 ; AVX512VLVBMI2-NEXT: vpaddb %zmm0, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpunpckhbw {{.*#+}} zmm2 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm2 ; AVX512VLVBMI2-NEXT: vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLVBMI2-NEXT: vpandq %zmm3, %zmm2, %zmm2 ; AVX512VLVBMI2-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpandq %zmm3, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpackuswb %zmm2, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vporq %zmm1, %zmm0, %zmm0 Index: llvm/test/CodeGen/X86/vector-fshr-rot-128.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshr-rot-128.ll +++ llvm/test/CodeGen/X86/vector-fshr-rot-128.ll @@ -1709,7 +1709,7 @@ ; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255] ; SSE41-NEXT: pand %xmm3, %xmm2 ; SSE41-NEXT: pmovzxbw {{.*#+}} xmm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero -; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [1,128,64,32,16,8,4,2] +; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [1,65408,64,32,16,8,4,2] ; SSE41-NEXT: pmullw %xmm1, %xmm4 ; SSE41-NEXT: pand %xmm3, %xmm4 ; SSE41-NEXT: packuswb %xmm2, %xmm4 Index: llvm/test/CodeGen/X86/vector-fshr-rot-256.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshr-rot-256.ll +++ llvm/test/CodeGen/X86/vector-fshr-rot-256.ll @@ -1405,11 +1405,11 @@ ; AVX1-NEXT: vpsrlw $8, %xmm7, %xmm7 ; AVX1-NEXT: vpackuswb %xmm3, %xmm7, %xmm3 ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15] -; AVX1-NEXT: vmovdqa {{.*#+}} xmm7 = [1,2,4,8,16,32,64,128] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm7 = [1,2,4,8,16,32,64,65408] ; AVX1-NEXT: vpmullw %xmm7, %xmm1, %xmm1 ; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255] ; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1 -; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,128,64,32,16,8,4,2] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,65408,64,32,16,8,4,2] ; AVX1-NEXT: vpmullw %xmm4, %xmm5, %xmm5 ; AVX1-NEXT: vpand %xmm2, %xmm5, %xmm5 ; AVX1-NEXT: vpackuswb %xmm1, %xmm5, %xmm1 Index: llvm/test/CodeGen/X86/vector-fshr-rot-512.ll =================================================================== --- llvm/test/CodeGen/X86/vector-fshr-rot-512.ll +++ llvm/test/CodeGen/X86/vector-fshr-rot-512.ll @@ -794,13 +794,13 @@ ; AVX512F: # %bb.0: ; AVX512F-NEXT: vextracti64x4 $1, %zmm0, %ymm1 ; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512F-NEXT: # ymm3 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm3, %ymm2, %ymm2 ; AVX512F-NEXT: vmovdqa {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512F-NEXT: vpand %ymm4, %ymm2, %ymm2 ; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm5 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,128,64,32,16,8,4,2,1,128,64,32,16,8,4,2] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2] ; AVX512F-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm6, %ymm5, %ymm5 ; AVX512F-NEXT: vpand %ymm4, %ymm5, %ymm5 @@ -840,13 +840,13 @@ ; AVX512VL: # %bb.0: ; AVX512VL-NEXT: vextracti64x4 $1, %zmm0, %ymm1 ; AVX512VL-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512VL-NEXT: # ymm3 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm3, %ymm2, %ymm2 ; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VL-NEXT: vpand %ymm4, %ymm2, %ymm2 ; AVX512VL-NEXT: vpunpcklbw {{.*#+}} ymm5 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,128,64,32,16,8,4,2,1,128,64,32,16,8,4,2] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2] ; AVX512VL-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm6, %ymm5, %ymm5 ; AVX512VL-NEXT: vpand %ymm4, %ymm5, %ymm5 @@ -885,11 +885,11 @@ ; AVX512BW-LABEL: constant_funnnel_v64i8: ; AVX512BW: # %bb.0: ; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 ; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512BW-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512BW-NEXT: vpandq %zmm2, %zmm3, %zmm2 ; AVX512BW-NEXT: vpackuswb %zmm1, %zmm2, %zmm1 ; AVX512BW-NEXT: vpxor %xmm2, %xmm2, %xmm2 @@ -906,11 +906,11 @@ ; AVX512VLBW-LABEL: constant_funnnel_v64i8: ; AVX512VLBW: # %bb.0: ; AVX512VLBW-NEXT: vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 ; AVX512VLBW-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLBW-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512VLBW-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VLBW-NEXT: vpandq %zmm2, %zmm3, %zmm2 ; AVX512VLBW-NEXT: vpackuswb %zmm1, %zmm2, %zmm1 ; AVX512VLBW-NEXT: vpxor %xmm2, %xmm2, %xmm2 @@ -927,11 +927,11 @@ ; AVX512VBMI2-LABEL: constant_funnnel_v64i8: ; AVX512VBMI2: # %bb.0: ; AVX512VBMI2-NEXT: vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 ; AVX512VBMI2-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VBMI2-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512VBMI2-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VBMI2-NEXT: vpandq %zmm2, %zmm3, %zmm2 ; AVX512VBMI2-NEXT: vpackuswb %zmm1, %zmm2, %zmm1 ; AVX512VBMI2-NEXT: vpxor %xmm2, %xmm2, %xmm2 @@ -948,11 +948,11 @@ ; AVX512VLVBMI2-LABEL: constant_funnnel_v64i8: ; AVX512VLVBMI2: # %bb.0: ; AVX512VLVBMI2-NEXT: vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 ; AVX512VLVBMI2-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLVBMI2-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512VLVBMI2-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VLVBMI2-NEXT: vpandq %zmm2, %zmm3, %zmm2 ; AVX512VLVBMI2-NEXT: vpackuswb %zmm1, %zmm2, %zmm1 ; AVX512VLVBMI2-NEXT: vpxor %xmm2, %xmm2, %xmm2 Index: llvm/test/CodeGen/X86/vector-rotate-128.ll =================================================================== --- llvm/test/CodeGen/X86/vector-rotate-128.ll +++ llvm/test/CodeGen/X86/vector-rotate-128.ll @@ -1595,7 +1595,7 @@ ; SSE41-NEXT: movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255] ; SSE41-NEXT: pand %xmm3, %xmm2 ; SSE41-NEXT: pmovzxbw {{.*#+}} xmm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero -; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [1,2,4,8,16,32,64,128] +; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [1,2,4,8,16,32,64,65408] ; SSE41-NEXT: pmullw %xmm1, %xmm4 ; SSE41-NEXT: pand %xmm3, %xmm4 ; SSE41-NEXT: packuswb %xmm2, %xmm4 Index: llvm/test/CodeGen/X86/vector-rotate-256.ll =================================================================== --- llvm/test/CodeGen/X86/vector-rotate-256.ll +++ llvm/test/CodeGen/X86/vector-rotate-256.ll @@ -1299,11 +1299,11 @@ ; AVX1-NEXT: vpsrlw $8, %xmm7, %xmm7 ; AVX1-NEXT: vpackuswb %xmm3, %xmm7, %xmm3 ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15] -; AVX1-NEXT: vmovdqa {{.*#+}} xmm7 = [1,128,64,32,16,8,4,2] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm7 = [1,65408,64,32,16,8,4,2] ; AVX1-NEXT: vpmullw %xmm7, %xmm1, %xmm1 ; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255] ; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1 -; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,2,4,8,16,32,64,128] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [1,2,4,8,16,32,64,65408] ; AVX1-NEXT: vpmullw %xmm4, %xmm5, %xmm5 ; AVX1-NEXT: vpand %xmm2, %xmm5, %xmm5 ; AVX1-NEXT: vpackuswb %xmm1, %xmm5, %xmm1 Index: llvm/test/CodeGen/X86/vector-rotate-512.ll =================================================================== --- llvm/test/CodeGen/X86/vector-rotate-512.ll +++ llvm/test/CodeGen/X86/vector-rotate-512.ll @@ -721,13 +721,13 @@ ; AVX512F: # %bb.0: ; AVX512F-NEXT: vextracti64x4 $1, %zmm0, %ymm1 ; AVX512F-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [0,128,64,32,16,8,4,2,0,128,64,32,16,8,4,2] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [0,65408,64,32,16,8,4,2,0,65408,64,32,16,8,4,2] ; AVX512F-NEXT: # ymm3 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm3, %ymm2, %ymm2 ; AVX512F-NEXT: vmovdqa {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512F-NEXT: vpand %ymm4, %ymm2, %ymm2 ; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm5 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512F-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm6, %ymm5, %ymm5 ; AVX512F-NEXT: vpand %ymm4, %ymm5, %ymm5 @@ -747,6 +747,8 @@ ; AVX512F-NEXT: vpmullw %ymm5, %ymm4, %ymm4 ; AVX512F-NEXT: vpsrlw $8, %ymm4, %ymm4 ; AVX512F-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm1[0],ymm3[0],ymm1[1],ymm3[1],ymm1[2],ymm3[2],ymm1[3],ymm3[3],ymm1[4],ymm3[4],ymm1[5],ymm3[5],ymm1[6],ymm3[6],ymm1[7],ymm3[7],ymm1[16],ymm3[16],ymm1[17],ymm3[17],ymm1[18],ymm3[18],ymm1[19],ymm3[19],ymm1[20],ymm3[20],ymm1[21],ymm3[21],ymm1[22],ymm3[22],ymm1[23],ymm3[23] +; AVX512F-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512F-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512F-NEXT: vpmullw %ymm6, %ymm1, %ymm1 ; AVX512F-NEXT: vpsrlw $8, %ymm1, %ymm1 ; AVX512F-NEXT: vpackuswb %ymm4, %ymm1, %ymm1 @@ -765,13 +767,13 @@ ; AVX512VL: # %bb.0: ; AVX512VL-NEXT: vextracti64x4 $1, %zmm0, %ymm1 ; AVX512VL-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [0,128,64,32,16,8,4,2,0,128,64,32,16,8,4,2] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [0,65408,64,32,16,8,4,2,0,65408,64,32,16,8,4,2] ; AVX512VL-NEXT: # ymm3 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm3, %ymm2, %ymm2 ; AVX512VL-NEXT: vmovdqa {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VL-NEXT: vpand %ymm4, %ymm2, %ymm2 ; AVX512VL-NEXT: vpunpcklbw {{.*#+}} ymm5 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512VL-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm6, %ymm5, %ymm5 ; AVX512VL-NEXT: vpand %ymm4, %ymm5, %ymm5 @@ -791,6 +793,8 @@ ; AVX512VL-NEXT: vpmullw %ymm5, %ymm4, %ymm4 ; AVX512VL-NEXT: vpsrlw $8, %ymm4, %ymm4 ; AVX512VL-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm1[0],ymm3[0],ymm1[1],ymm3[1],ymm1[2],ymm3[2],ymm1[3],ymm3[3],ymm1[4],ymm3[4],ymm1[5],ymm3[5],ymm1[6],ymm3[6],ymm1[7],ymm3[7],ymm1[16],ymm3[16],ymm1[17],ymm3[17],ymm1[18],ymm3[18],ymm1[19],ymm3[19],ymm1[20],ymm3[20],ymm1[21],ymm3[21],ymm1[22],ymm3[22],ymm1[23],ymm3[23] +; AVX512VL-NEXT: vbroadcasti128 {{.*#+}} ymm6 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512VL-NEXT: # ymm6 = mem[0,1,0,1] ; AVX512VL-NEXT: vpmullw %ymm6, %ymm1, %ymm1 ; AVX512VL-NEXT: vpsrlw $8, %ymm1, %ymm1 ; AVX512VL-NEXT: vpackuswb %ymm4, %ymm1, %ymm1 @@ -812,9 +816,7 @@ ; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512BW-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512BW-NEXT: vbroadcasti32x4 {{.*#+}} zmm4 = [0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7] -; AVX512BW-NEXT: # zmm4 = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3] -; AVX512BW-NEXT: vpsllvw %zmm4, %zmm3, %zmm3 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512BW-NEXT: vpandq %zmm2, %zmm3, %zmm2 ; AVX512BW-NEXT: vpackuswb %zmm1, %zmm2, %zmm1 ; AVX512BW-NEXT: vpxor %xmm2, %xmm2, %xmm2 @@ -822,7 +824,7 @@ ; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512BW-NEXT: vpsrlw $8, %zmm3, %zmm3 ; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0],zmm2[0],zmm0[1],zmm2[1],zmm0[2],zmm2[2],zmm0[3],zmm2[3],zmm0[4],zmm2[4],zmm0[5],zmm2[5],zmm0[6],zmm2[6],zmm0[7],zmm2[7],zmm0[16],zmm2[16],zmm0[17],zmm2[17],zmm0[18],zmm2[18],zmm0[19],zmm2[19],zmm0[20],zmm2[20],zmm0[21],zmm2[21],zmm0[22],zmm2[22],zmm0[23],zmm2[23],zmm0[32],zmm2[32],zmm0[33],zmm2[33],zmm0[34],zmm2[34],zmm0[35],zmm2[35],zmm0[36],zmm2[36],zmm0[37],zmm2[37],zmm0[38],zmm2[38],zmm0[39],zmm2[39],zmm0[48],zmm2[48],zmm0[49],zmm2[49],zmm0[50],zmm2[50],zmm0[51],zmm2[51],zmm0[52],zmm2[52],zmm0[53],zmm2[53],zmm0[54],zmm2[54],zmm0[55],zmm2[55] -; AVX512BW-NEXT: vpsllvw %zmm4, %zmm0, %zmm0 +; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512BW-NEXT: vpsrlw $8, %zmm0, %zmm0 ; AVX512BW-NEXT: vpackuswb %zmm3, %zmm0, %zmm0 ; AVX512BW-NEXT: vporq %zmm0, %zmm1, %zmm0 @@ -835,9 +837,7 @@ ; AVX512VLBW-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLBW-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512VLBW-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLBW-NEXT: vbroadcasti32x4 {{.*#+}} zmm4 = [0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7] -; AVX512VLBW-NEXT: # zmm4 = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3] -; AVX512VLBW-NEXT: vpsllvw %zmm4, %zmm3, %zmm3 +; AVX512VLBW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VLBW-NEXT: vpandq %zmm2, %zmm3, %zmm2 ; AVX512VLBW-NEXT: vpackuswb %zmm1, %zmm2, %zmm1 ; AVX512VLBW-NEXT: vpxor %xmm2, %xmm2, %xmm2 @@ -845,7 +845,7 @@ ; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VLBW-NEXT: vpsrlw $8, %zmm3, %zmm3 ; AVX512VLBW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0],zmm2[0],zmm0[1],zmm2[1],zmm0[2],zmm2[2],zmm0[3],zmm2[3],zmm0[4],zmm2[4],zmm0[5],zmm2[5],zmm0[6],zmm2[6],zmm0[7],zmm2[7],zmm0[16],zmm2[16],zmm0[17],zmm2[17],zmm0[18],zmm2[18],zmm0[19],zmm2[19],zmm0[20],zmm2[20],zmm0[21],zmm2[21],zmm0[22],zmm2[22],zmm0[23],zmm2[23],zmm0[32],zmm2[32],zmm0[33],zmm2[33],zmm0[34],zmm2[34],zmm0[35],zmm2[35],zmm0[36],zmm2[36],zmm0[37],zmm2[37],zmm0[38],zmm2[38],zmm0[39],zmm2[39],zmm0[48],zmm2[48],zmm0[49],zmm2[49],zmm0[50],zmm2[50],zmm0[51],zmm2[51],zmm0[52],zmm2[52],zmm0[53],zmm2[53],zmm0[54],zmm2[54],zmm0[55],zmm2[55] -; AVX512VLBW-NEXT: vpsllvw %zmm4, %zmm0, %zmm0 +; AVX512VLBW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpsrlw $8, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vpackuswb %zmm3, %zmm0, %zmm0 ; AVX512VLBW-NEXT: vporq %zmm0, %zmm1, %zmm0 @@ -858,9 +858,7 @@ ; AVX512VBMI2-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VBMI2-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512VBMI2-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VBMI2-NEXT: vbroadcasti32x4 {{.*#+}} zmm4 = [0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7] -; AVX512VBMI2-NEXT: # zmm4 = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3] -; AVX512VBMI2-NEXT: vpsllvw %zmm4, %zmm3, %zmm3 +; AVX512VBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VBMI2-NEXT: vpandq %zmm2, %zmm3, %zmm2 ; AVX512VBMI2-NEXT: vpackuswb %zmm1, %zmm2, %zmm1 ; AVX512VBMI2-NEXT: vpxor %xmm2, %xmm2, %xmm2 @@ -868,7 +866,7 @@ ; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VBMI2-NEXT: vpsrlw $8, %zmm3, %zmm3 ; AVX512VBMI2-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0],zmm2[0],zmm0[1],zmm2[1],zmm0[2],zmm2[2],zmm0[3],zmm2[3],zmm0[4],zmm2[4],zmm0[5],zmm2[5],zmm0[6],zmm2[6],zmm0[7],zmm2[7],zmm0[16],zmm2[16],zmm0[17],zmm2[17],zmm0[18],zmm2[18],zmm0[19],zmm2[19],zmm0[20],zmm2[20],zmm0[21],zmm2[21],zmm0[22],zmm2[22],zmm0[23],zmm2[23],zmm0[32],zmm2[32],zmm0[33],zmm2[33],zmm0[34],zmm2[34],zmm0[35],zmm2[35],zmm0[36],zmm2[36],zmm0[37],zmm2[37],zmm0[38],zmm2[38],zmm0[39],zmm2[39],zmm0[48],zmm2[48],zmm0[49],zmm2[49],zmm0[50],zmm2[50],zmm0[51],zmm2[51],zmm0[52],zmm2[52],zmm0[53],zmm2[53],zmm0[54],zmm2[54],zmm0[55],zmm2[55] -; AVX512VBMI2-NEXT: vpsllvw %zmm4, %zmm0, %zmm0 +; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpsrlw $8, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vpackuswb %zmm3, %zmm0, %zmm0 ; AVX512VBMI2-NEXT: vporq %zmm0, %zmm1, %zmm0 @@ -881,9 +879,7 @@ ; AVX512VLVBMI2-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512VLVBMI2-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512VLVBMI2-NEXT: vpunpcklbw {{.*#+}} zmm3 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512VLVBMI2-NEXT: vbroadcasti32x4 {{.*#+}} zmm4 = [0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7,0,1,2,3,4,5,6,7] -; AVX512VLVBMI2-NEXT: # zmm4 = mem[0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3] -; AVX512VLVBMI2-NEXT: vpsllvw %zmm4, %zmm3, %zmm3 +; AVX512VLVBMI2-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VLVBMI2-NEXT: vpandq %zmm2, %zmm3, %zmm2 ; AVX512VLVBMI2-NEXT: vpackuswb %zmm1, %zmm2, %zmm1 ; AVX512VLVBMI2-NEXT: vpxor %xmm2, %xmm2, %xmm2 @@ -891,7 +887,7 @@ ; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm3 ; AVX512VLVBMI2-NEXT: vpsrlw $8, %zmm3, %zmm3 ; AVX512VLVBMI2-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0],zmm2[0],zmm0[1],zmm2[1],zmm0[2],zmm2[2],zmm0[3],zmm2[3],zmm0[4],zmm2[4],zmm0[5],zmm2[5],zmm0[6],zmm2[6],zmm0[7],zmm2[7],zmm0[16],zmm2[16],zmm0[17],zmm2[17],zmm0[18],zmm2[18],zmm0[19],zmm2[19],zmm0[20],zmm2[20],zmm0[21],zmm2[21],zmm0[22],zmm2[22],zmm0[23],zmm2[23],zmm0[32],zmm2[32],zmm0[33],zmm2[33],zmm0[34],zmm2[34],zmm0[35],zmm2[35],zmm0[36],zmm2[36],zmm0[37],zmm2[37],zmm0[38],zmm2[38],zmm0[39],zmm2[39],zmm0[48],zmm2[48],zmm0[49],zmm2[49],zmm0[50],zmm2[50],zmm0[51],zmm2[51],zmm0[52],zmm2[52],zmm0[53],zmm2[53],zmm0[54],zmm2[54],zmm0[55],zmm2[55] -; AVX512VLVBMI2-NEXT: vpsllvw %zmm4, %zmm0, %zmm0 +; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpsrlw $8, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vpackuswb %zmm3, %zmm0, %zmm0 ; AVX512VLVBMI2-NEXT: vporq %zmm0, %zmm1, %zmm0 Index: llvm/test/CodeGen/X86/vector-shift-shl-256.ll =================================================================== --- llvm/test/CodeGen/X86/vector-shift-shl-256.ll +++ llvm/test/CodeGen/X86/vector-shift-shl-256.ll @@ -1264,12 +1264,12 @@ ; AVX1: # %bb.0: ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15] -; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [128,64,32,16,8,4,2,1] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [65408,64,32,16,8,4,2,1] ; AVX1-NEXT: vpmullw %xmm3, %xmm2, %xmm2 ; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [255,255,255,255,255,255,255,255] ; AVX1-NEXT: vpand %xmm4, %xmm2, %xmm2 ; AVX1-NEXT: vpmovzxbw {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero -; AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [1,2,4,8,16,32,64,128] +; AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [1,2,4,8,16,32,64,65408] ; AVX1-NEXT: vpmullw %xmm5, %xmm1, %xmm1 ; AVX1-NEXT: vpand %xmm4, %xmm1, %xmm1 ; AVX1-NEXT: vpackuswb %xmm2, %xmm1, %xmm1 @@ -1355,12 +1355,12 @@ ; X86-AVX1: # %bb.0: ; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1 ; X86-AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15] -; X86-AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [128,64,32,16,8,4,2,1] +; X86-AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [65408,64,32,16,8,4,2,1] ; X86-AVX1-NEXT: vpmullw %xmm3, %xmm2, %xmm2 ; X86-AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [255,255,255,255,255,255,255,255] ; X86-AVX1-NEXT: vpand %xmm4, %xmm2, %xmm2 ; X86-AVX1-NEXT: vpmovzxbw {{.*#+}} xmm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero -; X86-AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [1,2,4,8,16,32,64,128] +; X86-AVX1-NEXT: vmovdqa {{.*#+}} xmm5 = [1,2,4,8,16,32,64,65408] ; X86-AVX1-NEXT: vpmullw %xmm5, %xmm1, %xmm1 ; X86-AVX1-NEXT: vpand %xmm4, %xmm1, %xmm1 ; X86-AVX1-NEXT: vpackuswb %xmm2, %xmm1, %xmm1 Index: llvm/test/CodeGen/X86/vector-shift-shl-512.ll =================================================================== --- llvm/test/CodeGen/X86/vector-shift-shl-512.ll +++ llvm/test/CodeGen/X86/vector-shift-shl-512.ll @@ -303,13 +303,13 @@ ; AVX512DQ: # %bb.0: ; AVX512DQ-NEXT: vextracti64x4 $1, %zmm0, %ymm1 ; AVX512DQ-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31] -; AVX512DQ-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [128,64,32,16,8,4,2,1,128,64,32,16,8,4,2,1] +; AVX512DQ-NEXT: vbroadcasti128 {{.*#+}} ymm3 = [65408,64,32,16,8,4,2,1,65408,64,32,16,8,4,2,1] ; AVX512DQ-NEXT: # ymm3 = mem[0,1,0,1] ; AVX512DQ-NEXT: vpmullw %ymm3, %ymm2, %ymm2 ; AVX512DQ-NEXT: vmovdqa {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512DQ-NEXT: vpand %ymm4, %ymm2, %ymm2 ; AVX512DQ-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23] -; AVX512DQ-NEXT: vbroadcasti128 {{.*#+}} ymm5 = [1,2,4,8,16,32,64,128,1,2,4,8,16,32,64,128] +; AVX512DQ-NEXT: vbroadcasti128 {{.*#+}} ymm5 = [1,2,4,8,16,32,64,65408,1,2,4,8,16,32,64,65408] ; AVX512DQ-NEXT: # ymm5 = mem[0,1,0,1] ; AVX512DQ-NEXT: vpmullw %ymm5, %ymm1, %ymm1 ; AVX512DQ-NEXT: vpand %ymm4, %ymm1, %ymm1 @@ -327,11 +327,11 @@ ; AVX512BW-LABEL: constant_shift_v64i8: ; AVX512BW: # %bb.0: ; AVX512BW-NEXT: vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm1 ; AVX512BW-NEXT: vmovdqa64 {{.*#+}} zmm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX512BW-NEXT: vpandq %zmm2, %zmm1, %zmm1 ; AVX512BW-NEXT: vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55] -; AVX512BW-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 +; AVX512BW-NEXT: vpmullw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm0 ; AVX512BW-NEXT: vpandq %zmm2, %zmm0, %zmm0 ; AVX512BW-NEXT: vpackuswb %zmm1, %zmm0, %zmm0 ; AVX512BW-NEXT: retq