Index: llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp =================================================================== --- llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp +++ llvm/lib/Target/ARM/ARMTargetTransformInfo.cpp @@ -343,8 +343,8 @@ ValTy->isVectorTy() && ValTy->getScalarSizeInBits() <= 32) return 3; - if ((Opcode == Instruction::InsertElement || - Opcode == Instruction::ExtractElement)) { + if (ST->hasNEON() && (Opcode == Instruction::InsertElement || + Opcode == Instruction::ExtractElement)) { // Cross-class copies are expensive on many microarchitectures, // so assume they are expensive by default. if (ValTy->getVectorElementType()->isIntegerTy()) @@ -397,12 +397,15 @@ unsigned NumVectorInstToHideOverhead = 10; int MaxMergeDistance = 64; - if (Ty->isVectorTy() && SE && - !BaseT::isConstantStridedAccessLessThan(SE, Ptr, MaxMergeDistance + 1)) - return NumVectorInstToHideOverhead; + if (ST->hasNEON()) { + if (Ty->isVectorTy() && SE && + !BaseT::isConstantStridedAccessLessThan(SE, Ptr, MaxMergeDistance + 1)) + return NumVectorInstToHideOverhead; - // In many cases the address computation is not merged into the instruction - // addressing mode. + // In many cases the address computation is not merged into the instruction + // addressing mode. + return 1; + } return 1; } @@ -442,76 +445,73 @@ int ARMTTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index, Type *SubTp) { - if (Kind == TTI::SK_Broadcast) { - static const CostTblEntry NEONDupTbl[] = { - // VDUP handles these cases. - {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v4i16, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v8i8, 1}, - - {ISD::VECTOR_SHUFFLE, MVT::v4i32, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v4f32, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v8i16, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v16i8, 1}}; - - std::pair LT = TLI->getTypeLegalizationCost(DL, Tp); - - if (const auto *Entry = CostTableLookup(NEONDupTbl, ISD::VECTOR_SHUFFLE, - LT.second)) - return LT.first * Entry->Cost; - - return BaseT::getShuffleCost(Kind, Tp, Index, SubTp); - } - if (Kind == TTI::SK_Reverse) { - static const CostTblEntry NEONShuffleTbl[] = { - // Reverse shuffle cost one instruction if we are shuffling within a - // double word (vrev) or two if we shuffle a quad word (vrev, vext). - {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v4i16, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v8i8, 1}, - - {ISD::VECTOR_SHUFFLE, MVT::v4i32, 2}, - {ISD::VECTOR_SHUFFLE, MVT::v4f32, 2}, - {ISD::VECTOR_SHUFFLE, MVT::v8i16, 2}, - {ISD::VECTOR_SHUFFLE, MVT::v16i8, 2}}; - - std::pair LT = TLI->getTypeLegalizationCost(DL, Tp); - - if (const auto *Entry = CostTableLookup(NEONShuffleTbl, ISD::VECTOR_SHUFFLE, - LT.second)) - return LT.first * Entry->Cost; - - return BaseT::getShuffleCost(Kind, Tp, Index, SubTp); - } - if (Kind == TTI::SK_Select) { - static const CostTblEntry NEONSelShuffleTbl[] = { - // Select shuffle cost table for ARM. Cost is the number of instructions - // required to create the shuffled vector. + if (ST->hasNEON()) { + if (Kind == TTI::SK_Broadcast) { + static const CostTblEntry NEONDupTbl[] = { + // VDUP handles these cases. + {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v4i16, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v8i8, 1}, + + {ISD::VECTOR_SHUFFLE, MVT::v4i32, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v4f32, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v8i16, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v16i8, 1}}; + + std::pair LT = TLI->getTypeLegalizationCost(DL, Tp); + + if (const auto *Entry = CostTableLookup(NEONDupTbl, ISD::VECTOR_SHUFFLE, + LT.second)) + return LT.first * Entry->Cost; + } + if (Kind == TTI::SK_Reverse) { + static const CostTblEntry NEONShuffleTbl[] = { + // Reverse shuffle cost one instruction if we are shuffling within a + // double word (vrev) or two if we shuffle a quad word (vrev, vext). + {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v4i16, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v8i8, 1}, + + {ISD::VECTOR_SHUFFLE, MVT::v4i32, 2}, + {ISD::VECTOR_SHUFFLE, MVT::v4f32, 2}, + {ISD::VECTOR_SHUFFLE, MVT::v8i16, 2}, + {ISD::VECTOR_SHUFFLE, MVT::v16i8, 2}}; + + std::pair LT = TLI->getTypeLegalizationCost(DL, Tp); + + if (const auto *Entry = CostTableLookup(NEONShuffleTbl, ISD::VECTOR_SHUFFLE, + LT.second)) + return LT.first * Entry->Cost; + } + if (Kind == TTI::SK_Select) { + static const CostTblEntry NEONSelShuffleTbl[] = { + // Select shuffle cost table for ARM. Cost is the number of instructions + // required to create the shuffled vector. - {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2f32, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2i64, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2f64, 1}, + {ISD::VECTOR_SHUFFLE, MVT::v2i32, 1}, - {ISD::VECTOR_SHUFFLE, MVT::v4i32, 2}, - {ISD::VECTOR_SHUFFLE, MVT::v4f32, 2}, - {ISD::VECTOR_SHUFFLE, MVT::v4i16, 2}, + {ISD::VECTOR_SHUFFLE, MVT::v4i32, 2}, + {ISD::VECTOR_SHUFFLE, MVT::v4f32, 2}, + {ISD::VECTOR_SHUFFLE, MVT::v4i16, 2}, - {ISD::VECTOR_SHUFFLE, MVT::v8i16, 16}, + {ISD::VECTOR_SHUFFLE, MVT::v8i16, 16}, - {ISD::VECTOR_SHUFFLE, MVT::v16i8, 32}}; + {ISD::VECTOR_SHUFFLE, MVT::v16i8, 32}}; - std::pair LT = TLI->getTypeLegalizationCost(DL, Tp); - if (const auto *Entry = CostTableLookup(NEONSelShuffleTbl, - ISD::VECTOR_SHUFFLE, LT.second)) - return LT.first * Entry->Cost; - return BaseT::getShuffleCost(Kind, Tp, Index, SubTp); + std::pair LT = TLI->getTypeLegalizationCost(DL, Tp); + if (const auto *Entry = CostTableLookup(NEONSelShuffleTbl, + ISD::VECTOR_SHUFFLE, LT.second)) + return LT.first * Entry->Cost; + } } return BaseT::getShuffleCost(Kind, Tp, Index, SubTp); } @@ -592,7 +592,7 @@ unsigned AddressSpace, const Instruction *I) { std::pair LT = TLI->getTypeLegalizationCost(DL, Src); - if (Src->isVectorTy() && Alignment != 16 && + if (ST->hasNEON() && Src->isVectorTy() && Alignment != 16 && Src->getVectorElementType()->isDoubleTy()) { // Unaligned loads/stores are extremely inefficient. // We need 4 uops for vst.1/vld.1 vs 1uop for vldr/vstr. Index: llvm/test/Analysis/CostModel/ARM/cast.ll =================================================================== --- llvm/test/Analysis/CostModel/ARM/cast.ll +++ llvm/test/Analysis/CostModel/ARM/cast.ll @@ -351,26 +351,26 @@ ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r87 = fpext <4 x float> undef to <4 x double> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r88 = fpext <8 x float> undef to <8 x double> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r89 = fpext <16 x float> undef to <16 x double> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r90 = fptoui <2 x float> undef to <2 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r91 = fptosi <2 x float> undef to <2 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r92 = fptoui <2 x float> undef to <2 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r93 = fptosi <2 x float> undef to <2 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r94 = fptoui <2 x float> undef to <2 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r95 = fptosi <2 x float> undef to <2 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r96 = fptoui <2 x float> undef to <2 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r97 = fptosi <2 x float> undef to <2 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r98 = fptoui <2 x float> undef to <2 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r99 = fptosi <2 x float> undef to <2 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r100 = fptoui <2 x double> undef to <2 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r101 = fptosi <2 x double> undef to <2 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r102 = fptoui <2 x double> undef to <2 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r103 = fptosi <2 x double> undef to <2 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r104 = fptoui <2 x double> undef to <2 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r105 = fptosi <2 x double> undef to <2 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r106 = fptoui <2 x double> undef to <2 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r107 = fptosi <2 x double> undef to <2 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r108 = fptoui <2 x double> undef to <2 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %r109 = fptosi <2 x double> undef to <2 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r90 = fptoui <2 x float> undef to <2 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r91 = fptosi <2 x float> undef to <2 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r92 = fptoui <2 x float> undef to <2 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r93 = fptosi <2 x float> undef to <2 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r94 = fptoui <2 x float> undef to <2 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r95 = fptosi <2 x float> undef to <2 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r96 = fptoui <2 x float> undef to <2 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r97 = fptosi <2 x float> undef to <2 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %r98 = fptoui <2 x float> undef to <2 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %r99 = fptosi <2 x float> undef to <2 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r100 = fptoui <2 x double> undef to <2 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r101 = fptosi <2 x double> undef to <2 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r102 = fptoui <2 x double> undef to <2 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r103 = fptosi <2 x double> undef to <2 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r104 = fptoui <2 x double> undef to <2 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r105 = fptosi <2 x double> undef to <2 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r106 = fptoui <2 x double> undef to <2 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %r107 = fptosi <2 x double> undef to <2 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %r108 = fptoui <2 x double> undef to <2 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %r109 = fptosi <2 x double> undef to <2 x i64> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r110 = fptoui <4 x float> undef to <4 x i1> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r111 = fptosi <4 x float> undef to <4 x i1> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r112 = fptoui <4 x float> undef to <4 x i8> @@ -379,18 +379,18 @@ ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r115 = fptosi <4 x float> undef to <4 x i16> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r116 = fptoui <4 x float> undef to <4 x i32> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r117 = fptosi <4 x float> undef to <4 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r118 = fptoui <4 x float> undef to <4 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r119 = fptosi <4 x float> undef to <4 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r120 = fptoui <4 x double> undef to <4 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r121 = fptosi <4 x double> undef to <4 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r122 = fptoui <4 x double> undef to <4 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r123 = fptosi <4 x double> undef to <4 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r124 = fptoui <4 x double> undef to <4 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r125 = fptosi <4 x double> undef to <4 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r126 = fptoui <4 x double> undef to <4 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r127 = fptosi <4 x double> undef to <4 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r128 = fptoui <4 x double> undef to <4 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %r129 = fptosi <4 x double> undef to <4 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %r118 = fptoui <4 x float> undef to <4 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %r119 = fptosi <4 x float> undef to <4 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r120 = fptoui <4 x double> undef to <4 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r121 = fptosi <4 x double> undef to <4 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r122 = fptoui <4 x double> undef to <4 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r123 = fptosi <4 x double> undef to <4 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r124 = fptoui <4 x double> undef to <4 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r125 = fptosi <4 x double> undef to <4 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r126 = fptoui <4 x double> undef to <4 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %r127 = fptosi <4 x double> undef to <4 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %r128 = fptoui <4 x double> undef to <4 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %r129 = fptosi <4 x double> undef to <4 x i64> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %r130 = fptoui <8 x float> undef to <8 x i1> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %r131 = fptosi <8 x float> undef to <8 x i1> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %r132 = fptoui <8 x float> undef to <8 x i8> @@ -399,18 +399,18 @@ ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %r135 = fptosi <8 x float> undef to <8 x i16> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r136 = fptoui <8 x float> undef to <8 x i32> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r137 = fptosi <8 x float> undef to <8 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r138 = fptoui <8 x float> undef to <8 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r139 = fptosi <8 x float> undef to <8 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r140 = fptoui <8 x double> undef to <8 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r141 = fptosi <8 x double> undef to <8 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r142 = fptoui <8 x double> undef to <8 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r143 = fptosi <8 x double> undef to <8 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r144 = fptoui <8 x double> undef to <8 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r145 = fptosi <8 x double> undef to <8 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r146 = fptoui <8 x double> undef to <8 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r147 = fptosi <8 x double> undef to <8 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r148 = fptoui <8 x double> undef to <8 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 59 for instruction: %r149 = fptosi <8 x double> undef to <8 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %r138 = fptoui <8 x float> undef to <8 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %r139 = fptosi <8 x float> undef to <8 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r140 = fptoui <8 x double> undef to <8 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r141 = fptosi <8 x double> undef to <8 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r142 = fptoui <8 x double> undef to <8 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r143 = fptosi <8 x double> undef to <8 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r144 = fptoui <8 x double> undef to <8 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r145 = fptosi <8 x double> undef to <8 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r146 = fptoui <8 x double> undef to <8 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %r147 = fptosi <8 x double> undef to <8 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %r148 = fptoui <8 x double> undef to <8 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %r149 = fptosi <8 x double> undef to <8 x i64> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %r150 = fptoui <16 x float> undef to <16 x i1> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %r151 = fptosi <16 x float> undef to <16 x i1> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %r152 = fptoui <16 x float> undef to <16 x i8> @@ -419,18 +419,18 @@ ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %r155 = fptosi <16 x float> undef to <16 x i16> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r156 = fptoui <16 x float> undef to <16 x i32> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r157 = fptosi <16 x float> undef to <16 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r158 = fptoui <16 x float> undef to <16 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r159 = fptosi <16 x float> undef to <16 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r160 = fptoui <16 x double> undef to <16 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r161 = fptosi <16 x double> undef to <16 x i1> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r162 = fptoui <16 x double> undef to <16 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r163 = fptosi <16 x double> undef to <16 x i8> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r164 = fptoui <16 x double> undef to <16 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r165 = fptosi <16 x double> undef to <16 x i16> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r166 = fptoui <16 x double> undef to <16 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r167 = fptosi <16 x double> undef to <16 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r168 = fptoui <16 x double> undef to <16 x i64> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %r169 = fptosi <16 x double> undef to <16 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 87 for instruction: %r158 = fptoui <16 x float> undef to <16 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 87 for instruction: %r159 = fptosi <16 x float> undef to <16 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r160 = fptoui <16 x double> undef to <16 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r161 = fptosi <16 x double> undef to <16 x i1> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r162 = fptoui <16 x double> undef to <16 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r163 = fptosi <16 x double> undef to <16 x i8> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r164 = fptoui <16 x double> undef to <16 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r165 = fptosi <16 x double> undef to <16 x i16> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r166 = fptoui <16 x double> undef to <16 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 55 for instruction: %r167 = fptosi <16 x double> undef to <16 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 87 for instruction: %r168 = fptoui <16 x double> undef to <16 x i64> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 87 for instruction: %r169 = fptosi <16 x double> undef to <16 x i64> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r170 = uitofp <2 x i1> undef to <2 x float> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r171 = sitofp <2 x i1> undef to <2 x float> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %r172 = uitofp <2 x i8> undef to <2 x float> Index: llvm/test/Analysis/CostModel/ARM/divrem.ll =================================================================== --- llvm/test/Analysis/CostModel/ARM/divrem.ll +++ llvm/test/Analysis/CostModel/ARM/divrem.ll @@ -60,7 +60,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i8> %1 ; ; CHECK-MVE-LABEL: 'sdiv_v2_i8' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %1 = sdiv <2 x i8> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %1 = sdiv <2 x i8> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i8> %1 ; %1 = sdiv <2 x i8> %a, %b @@ -72,7 +72,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i16> %1 ; ; CHECK-MVE-LABEL: 'sdiv_v2_i16' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %1 = sdiv <2 x i16> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %1 = sdiv <2 x i16> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i16> %1 ; %1 = sdiv <2 x i16> %a, %b @@ -84,7 +84,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i32> %1 ; ; CHECK-MVE-LABEL: 'sdiv_v2_i32' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %1 = sdiv <2 x i32> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %1 = sdiv <2 x i32> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i32> %1 ; %1 = sdiv <2 x i32> %a, %b @@ -96,7 +96,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %1 ; ; CHECK-MVE-LABEL: 'sdiv_v2_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %1 = sdiv <2 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %1 = sdiv <2 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %1 ; %1 = sdiv <2 x i64> %a, %b @@ -144,7 +144,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1 ; ; CHECK-MVE-LABEL: 'sdiv_v4_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %1 = sdiv <4 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %1 = sdiv <4 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1 ; %1 = sdiv <4 x i64> %a, %b @@ -192,7 +192,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %1 ; ; CHECK-MVE-LABEL: 'sdiv_v8_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 80 for instruction: %1 = sdiv <8 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %1 = sdiv <8 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %1 ; %1 = sdiv <8 x i64> %a, %b @@ -240,7 +240,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i64> %1 ; ; CHECK-MVE-LABEL: 'sdiv_v16_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 160 for instruction: %1 = sdiv <16 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %1 = sdiv <16 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i64> %1 ; %1 = sdiv <16 x i64> %a, %b @@ -252,7 +252,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i8> %1 ; ; CHECK-MVE-LABEL: 'udiv_v2_i8' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %1 = udiv <2 x i8> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %1 = udiv <2 x i8> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i8> %1 ; %1 = udiv <2 x i8> %a, %b @@ -264,7 +264,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i16> %1 ; ; CHECK-MVE-LABEL: 'udiv_v2_i16' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %1 = udiv <2 x i16> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %1 = udiv <2 x i16> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i16> %1 ; %1 = udiv <2 x i16> %a, %b @@ -276,7 +276,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i32> %1 ; ; CHECK-MVE-LABEL: 'udiv_v2_i32' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %1 = udiv <2 x i32> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %1 = udiv <2 x i32> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i32> %1 ; %1 = udiv <2 x i32> %a, %b @@ -288,7 +288,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %1 ; ; CHECK-MVE-LABEL: 'udiv_v2_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %1 = udiv <2 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %1 = udiv <2 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %1 ; %1 = udiv <2 x i64> %a, %b @@ -336,7 +336,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1 ; ; CHECK-MVE-LABEL: 'udiv_v4_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %1 = udiv <4 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %1 = udiv <4 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1 ; %1 = udiv <4 x i64> %a, %b @@ -384,7 +384,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %1 ; ; CHECK-MVE-LABEL: 'udiv_v8_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 80 for instruction: %1 = udiv <8 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %1 = udiv <8 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %1 ; %1 = udiv <8 x i64> %a, %b @@ -432,7 +432,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i64> %1 ; ; CHECK-MVE-LABEL: 'udiv_v16_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 160 for instruction: %1 = udiv <16 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %1 = udiv <16 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i64> %1 ; %1 = udiv <16 x i64> %a, %b @@ -444,7 +444,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i8> %1 ; ; CHECK-MVE-LABEL: 'srem_v2_i8' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %1 = srem <2 x i8> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %1 = srem <2 x i8> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i8> %1 ; %1 = srem <2 x i8> %a, %b @@ -456,7 +456,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i16> %1 ; ; CHECK-MVE-LABEL: 'srem_v2_i16' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %1 = srem <2 x i16> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %1 = srem <2 x i16> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i16> %1 ; %1 = srem <2 x i16> %a, %b @@ -468,7 +468,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i32> %1 ; ; CHECK-MVE-LABEL: 'srem_v2_i32' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %1 = srem <2 x i32> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %1 = srem <2 x i32> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i32> %1 ; %1 = srem <2 x i32> %a, %b @@ -480,7 +480,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %1 ; ; CHECK-MVE-LABEL: 'srem_v2_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %1 = srem <2 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %1 = srem <2 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %1 ; %1 = srem <2 x i64> %a, %b @@ -528,7 +528,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1 ; ; CHECK-MVE-LABEL: 'srem_v4_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %1 = srem <4 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %1 = srem <4 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1 ; %1 = srem <4 x i64> %a, %b @@ -576,7 +576,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %1 ; ; CHECK-MVE-LABEL: 'srem_v8_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %1 = srem <8 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %1 = srem <8 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %1 ; %1 = srem <8 x i64> %a, %b @@ -624,7 +624,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i64> %1 ; ; CHECK-MVE-LABEL: 'srem_v16_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 112 for instruction: %1 = srem <16 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 80 for instruction: %1 = srem <16 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i64> %1 ; %1 = srem <16 x i64> %a, %b @@ -636,7 +636,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i8> %1 ; ; CHECK-MVE-LABEL: 'urem_v2_i8' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %1 = urem <2 x i8> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %1 = urem <2 x i8> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i8> %1 ; %1 = urem <2 x i8> %a, %b @@ -648,7 +648,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i16> %1 ; ; CHECK-MVE-LABEL: 'urem_v2_i16' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %1 = urem <2 x i16> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %1 = urem <2 x i16> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i16> %1 ; %1 = urem <2 x i16> %a, %b @@ -660,7 +660,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i32> %1 ; ; CHECK-MVE-LABEL: 'urem_v2_i32' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %1 = urem <2 x i32> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %1 = urem <2 x i32> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i32> %1 ; %1 = urem <2 x i32> %a, %b @@ -672,7 +672,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %1 ; ; CHECK-MVE-LABEL: 'urem_v2_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %1 = urem <2 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %1 = urem <2 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %1 ; %1 = urem <2 x i64> %a, %b @@ -720,7 +720,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1 ; ; CHECK-MVE-LABEL: 'urem_v4_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %1 = urem <4 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %1 = urem <4 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %1 ; %1 = urem <4 x i64> %a, %b @@ -768,7 +768,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %1 ; ; CHECK-MVE-LABEL: 'urem_v8_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %1 = urem <8 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %1 = urem <8 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %1 ; %1 = urem <8 x i64> %a, %b @@ -816,7 +816,7 @@ ; CHECK-NEON-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i64> %1 ; ; CHECK-MVE-LABEL: 'urem_v16_i64' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 112 for instruction: %1 = urem <16 x i64> %a, %b +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 80 for instruction: %1 = urem <16 x i64> %a, %b ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i64> %1 ; %1 = urem <16 x i64> %a, %b Index: llvm/test/Analysis/CostModel/ARM/load_store.ll =================================================================== --- llvm/test/Analysis/CostModel/ARM/load_store.ll +++ llvm/test/Analysis/CostModel/ARM/load_store.ll @@ -23,14 +23,14 @@ ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <8 x i16> undef, <8 x i16>* undef, align 2 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: store <16 x i8> undef, <16 x i8>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <4 x float> undef, <4 x float>* undef, align 4 -; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: store <4 x double> undef, <4 x double>* undef, align 4 +; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <4 x double> undef, <4 x double>* undef, align 4 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <2 x float> undef, <2 x float>* undef, align 4 -; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: store <2 x double> undef, <2 x double>* undef, align 4 +; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <2 x double> undef, <2 x double>* undef, align 4 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <2 x i64> undef, <2 x i64>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <4 x i32> undef, <4 x i32>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <8 x i16> undef, <8 x i16>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <4 x float> undef, <4 x float>* undef, align 1 -; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: store <2 x double> undef, <2 x double>* undef, align 1 +; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <2 x double> undef, <2 x double>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; CHECK-FP-LABEL: 'stores' @@ -49,14 +49,14 @@ ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <8 x i16> undef, <8 x i16>* undef, align 2 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 16 for instruction: store <16 x i8> undef, <16 x i8>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <4 x float> undef, <4 x float>* undef, align 4 -; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 16 for instruction: store <4 x double> undef, <4 x double>* undef, align 4 +; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <4 x double> undef, <4 x double>* undef, align 4 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <2 x float> undef, <2 x float>* undef, align 4 -; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <2 x double> undef, <2 x double>* undef, align 4 +; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <2 x double> undef, <2 x double>* undef, align 4 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <2 x i64> undef, <2 x i64>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <4 x i32> undef, <4 x i32>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <8 x i16> undef, <8 x i16>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <4 x float> undef, <4 x float>* undef, align 1 -; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <2 x double> undef, <2 x double>* undef, align 1 +; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <2 x double> undef, <2 x double>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; CHECK-MVE-LABEL: 'stores' @@ -75,14 +75,14 @@ ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <8 x i16> undef, <8 x i16>* undef, align 2 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <16 x i8> undef, <16 x i8>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <4 x float> undef, <4 x float>* undef, align 4 -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: store <4 x double> undef, <4 x double>* undef, align 4 +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: store <4 x double> undef, <4 x double>* undef, align 4 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <2 x float> undef, <2 x float>* undef, align 4 -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <2 x double> undef, <2 x double>* undef, align 4 +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <2 x double> undef, <2 x double>* undef, align 4 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <2 x i64> undef, <2 x i64>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <4 x i32> undef, <4 x i32>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <8 x i16> undef, <8 x i16>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <4 x float> undef, <4 x float>* undef, align 1 -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: store <2 x double> undef, <2 x double>* undef, align 1 +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: store <2 x double> undef, <2 x double>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; CHECK-NEON-LABEL: 'stores' @@ -158,14 +158,14 @@ ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %13 = load <8 x i16>, <8 x i16>* undef, align 2 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %14 = load <16 x i8>, <16 x i8>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %15 = load <4 x float>, <4 x float>* undef, align 4 -; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %16 = load <4 x double>, <4 x double>* undef, align 4 +; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %16 = load <4 x double>, <4 x double>* undef, align 4 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %17 = load <2 x float>, <2 x float>* undef, align 4 -; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %18 = load <2 x double>, <2 x double>* undef, align 4 +; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %18 = load <2 x double>, <2 x double>* undef, align 4 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %19 = load <2 x i64>, <2 x i64>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %20 = load <4 x i32>, <4 x i32>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %21 = load <8 x i16>, <8 x i16>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %22 = load <4 x float>, <4 x float>* undef, align 1 -; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %23 = load <2 x double>, <2 x double>* undef, align 1 +; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %23 = load <2 x double>, <2 x double>* undef, align 1 ; CHECK-NOVEC-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; CHECK-FP-LABEL: 'loads' @@ -184,14 +184,14 @@ ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %13 = load <8 x i16>, <8 x i16>* undef, align 2 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %14 = load <16 x i8>, <16 x i8>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %15 = load <4 x float>, <4 x float>* undef, align 4 -; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %16 = load <4 x double>, <4 x double>* undef, align 4 +; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %16 = load <4 x double>, <4 x double>* undef, align 4 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %17 = load <2 x float>, <2 x float>* undef, align 4 -; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %18 = load <2 x double>, <2 x double>* undef, align 4 +; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %18 = load <2 x double>, <2 x double>* undef, align 4 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %19 = load <2 x i64>, <2 x i64>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %20 = load <4 x i32>, <4 x i32>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %21 = load <8 x i16>, <8 x i16>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %22 = load <4 x float>, <4 x float>* undef, align 1 -; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %23 = load <2 x double>, <2 x double>* undef, align 1 +; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %23 = load <2 x double>, <2 x double>* undef, align 1 ; CHECK-FP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; CHECK-MVE-LABEL: 'loads' @@ -210,14 +210,14 @@ ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %13 = load <8 x i16>, <8 x i16>* undef, align 2 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %14 = load <16 x i8>, <16 x i8>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %15 = load <4 x float>, <4 x float>* undef, align 4 -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %16 = load <4 x double>, <4 x double>* undef, align 4 +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %16 = load <4 x double>, <4 x double>* undef, align 4 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %17 = load <2 x float>, <2 x float>* undef, align 4 -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %18 = load <2 x double>, <2 x double>* undef, align 4 +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %18 = load <2 x double>, <2 x double>* undef, align 4 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %19 = load <2 x i64>, <2 x i64>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %20 = load <4 x i32>, <4 x i32>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %21 = load <8 x i16>, <8 x i16>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %22 = load <4 x float>, <4 x float>* undef, align 1 -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %23 = load <2 x double>, <2 x double>* undef, align 1 +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %23 = load <2 x double>, <2 x double>* undef, align 1 ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; CHECK-NEON-LABEL: 'loads' Index: llvm/test/Analysis/CostModel/ARM/select.ll =================================================================== --- llvm/test/Analysis/CostModel/ARM/select.ll +++ llvm/test/Analysis/CostModel/ARM/select.ll @@ -10,26 +10,26 @@ ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v4 = select i1 undef, i64 undef, i64 undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v5 = select i1 undef, float undef, float undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v6 = select i1 undef, double undef, double undef -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %v7 = select <2 x i1> undef, <2 x i8> undef, <2 x i8> undef +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v7 = select <2 x i1> undef, <2 x i8> undef, <2 x i8> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v8 = select <4 x i1> undef, <4 x i8> undef, <4 x i8> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v9 = select <8 x i1> undef, <8 x i8> undef, <8 x i8> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v10 = select <16 x i1> undef, <16 x i8> undef, <16 x i8> undef -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %v11 = select <2 x i1> undef, <2 x i16> undef, <2 x i16> undef +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v11 = select <2 x i1> undef, <2 x i16> undef, <2 x i16> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v12 = select <4 x i1> undef, <4 x i16> undef, <4 x i16> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v13 = select <8 x i1> undef, <8 x i16> undef, <8 x i16> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v13b = select <16 x i1> undef, <16 x i16> undef, <16 x i16> undef -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %v14 = select <2 x i1> undef, <2 x i32> undef, <2 x i32> undef +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v14 = select <2 x i1> undef, <2 x i32> undef, <2 x i32> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v15 = select <4 x i1> undef, <4 x i32> undef, <4 x i32> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v15b = select <8 x i1> undef, <8 x i32> undef, <8 x i32> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v15c = select <16 x i1> undef, <16 x i32> undef, <16 x i32> undef -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %v16 = select <2 x i1> undef, <2 x i64> undef, <2 x i64> undef -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %v16a = select <4 x i1> undef, <4 x i64> undef, <4 x i64> undef -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %v16b = select <8 x i1> undef, <8 x i64> undef, <8 x i64> undef -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 80 for instruction: %v16c = select <16 x i1> undef, <16 x i64> undef, <16 x i64> undef +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %v16 = select <2 x i1> undef, <2 x i64> undef, <2 x i64> undef +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v16a = select <4 x i1> undef, <4 x i64> undef, <4 x i64> undef +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v16b = select <8 x i1> undef, <8 x i64> undef, <8 x i64> undef +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %v16c = select <16 x i1> undef, <16 x i64> undef, <16 x i64> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v17 = select <2 x i1> undef, <2 x float> undef, <2 x float> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v18 = select <4 x i1> undef, <4 x float> undef, <4 x float> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v19 = select <2 x i1> undef, <2 x double> undef, <2 x double> undef -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v20 = select <1 x i1> undef, <1 x i32> undef, <1 x i32> undef +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v20 = select <1 x i1> undef, <1 x i32> undef, <1 x i32> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v21 = select <3 x i1> undef, <3 x float> undef, <3 x float> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %v22 = select <5 x i1> undef, <5 x double> undef, <5 x double> undef ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void Index: llvm/test/Analysis/CostModel/ARM/shuffle.ll =================================================================== --- llvm/test/Analysis/CostModel/ARM/shuffle.ll +++ llvm/test/Analysis/CostModel/ARM/shuffle.ll @@ -4,17 +4,17 @@ define void @broadcast() { ; CHECK-MVE-LABEL: 'broadcast' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v7 = shufflevector <2 x i8> undef, <2 x i8> undef, <2 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v8 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v9 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v10 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v11 = shufflevector <2 x i16> undef, <2 x i16> undef, <2 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v12 = shufflevector <4 x i16> undef, <4 x i16> undef, <4 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v13 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v14 = shufflevector <2 x i32> undef, <2 x i32> undef, <2 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v15 = shufflevector <4 x i32> undef, <4 x i32> undef, <4 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v16 = shufflevector <2 x float> undef, <2 x float> undef, <2 x i32> zeroinitializer -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v17 = shufflevector <4 x float> undef, <4 x float> undef, <4 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v7 = shufflevector <2 x i8> undef, <2 x i8> undef, <2 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %v8 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %v9 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %v10 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v11 = shufflevector <2 x i16> undef, <2 x i16> undef, <2 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %v12 = shufflevector <4 x i16> undef, <4 x i16> undef, <4 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %v13 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v14 = shufflevector <2 x i32> undef, <2 x i32> undef, <2 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %v15 = shufflevector <4 x i32> undef, <4 x i32> undef, <4 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %v16 = shufflevector <2 x float> undef, <2 x float> undef, <2 x i32> zeroinitializer +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %v17 = shufflevector <4 x float> undef, <4 x float> undef, <4 x i32> zeroinitializer ; CHECK-MVE-NEXT: Cost Model: Unknown cost for instruction: %v18 = shufflevector <8 x half> undef, <8 x half> undef, <4 x i32> zeroinitializer ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; @@ -55,18 +55,18 @@ ;; Reverse shuffles should be lowered to vrev and possibly a vext (for quadwords, on neon) define void @reverse() { ; CHECK-MVE-LABEL: 'reverse' -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v7 = shufflevector <2 x i8> undef, <2 x i8> undef, <2 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v8 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v9 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v10 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v11 = shufflevector <2 x i16> undef, <2 x i16> undef, <2 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v12 = shufflevector <4 x i16> undef, <4 x i16> undef, <4 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v13 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v14 = shufflevector <2 x i32> undef, <2 x i32> undef, <2 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v15 = shufflevector <4 x i32> undef, <4 x i32> undef, <4 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v16 = shufflevector <2 x float> undef, <2 x float> undef, <2 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %v17 = shufflevector <4 x float> undef, <4 x float> undef, <4 x i32> -; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v18 = shufflevector <8 x half> undef, <8 x half> undef, <8 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v7 = shufflevector <2 x i8> undef, <2 x i8> undef, <2 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %v8 = shufflevector <4 x i8> undef, <4 x i8> undef, <4 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v9 = shufflevector <8 x i8> undef, <8 x i8> undef, <8 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v10 = shufflevector <16 x i8> undef, <16 x i8> undef, <16 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v11 = shufflevector <2 x i16> undef, <2 x i16> undef, <2 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %v12 = shufflevector <4 x i16> undef, <4 x i16> undef, <4 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v13 = shufflevector <8 x i16> undef, <8 x i16> undef, <8 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v14 = shufflevector <2 x i32> undef, <2 x i32> undef, <2 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %v15 = shufflevector <4 x i32> undef, <4 x i32> undef, <4 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %v16 = shufflevector <2 x float> undef, <2 x float> undef, <2 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %v17 = shufflevector <4 x float> undef, <4 x float> undef, <4 x i32> +; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v18 = shufflevector <8 x half> undef, <8 x half> undef, <8 x i32> ; CHECK-MVE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void ; ; CHECK-NEON-LABEL: 'reverse'