diff options
| author | Craig Topper <craig.topper@intel.com> | 2018-02-20 07:28:14 +0000 |
|---|---|---|
| committer | Craig Topper <craig.topper@intel.com> | 2018-02-20 07:28:14 +0000 |
| commit | 9256ac1a580cd6e30db4ed310383a9f12b602091 (patch) | |
| tree | c016d615d3e8f29574d5ab48b66b9ce61ef09c69 /llvm/lib/IR/AutoUpgrade.cpp | |
| parent | 718d04c6534cc83f8d48b9ce94dcb3f46e02a7e7 (diff) | |
| download | bcm5719-llvm-9256ac1a580cd6e30db4ed310383a9f12b602091.tar.gz bcm5719-llvm-9256ac1a580cd6e30db4ed310383a9f12b602091.zip | |
[X86] Add 512-bit unmasked pmulhrsw/pmulhw/pmulhuw intrinsics. Remove and auto upgrade 128/256/512 bit masked pmulhrsw/pmulhw/pmulhuw intrinsics.
The 128 and 256 bit versions were already not used by clang. This adds an equivalent unmasked 512 bit version. Then autoupgrades all sizes to use unmasked intrinsics plus select.
llvm-svn: 325559
Diffstat (limited to 'llvm/lib/IR/AutoUpgrade.cpp')
| -rw-r--r-- | llvm/lib/IR/AutoUpgrade.cpp | 43 |
1 files changed, 43 insertions, 0 deletions
diff --git a/llvm/lib/IR/AutoUpgrade.cpp b/llvm/lib/IR/AutoUpgrade.cpp index 9f6590dd609..0cf28c9fd22 100644 --- a/llvm/lib/IR/AutoUpgrade.cpp +++ b/llvm/lib/IR/AutoUpgrade.cpp @@ -170,6 +170,9 @@ static bool ShouldUpgradeX86Intrinsic(Function *F, StringRef Name) { Name.startswith("avx512.mask.cvtudq2pd.") || // Added in 4.0 Name.startswith("avx512.mask.pmul.dq.") || // Added in 4.0 Name.startswith("avx512.mask.pmulu.dq.") || // Added in 4.0 + Name.startswith("avx512.mask.pmul.hr.sw.") || // Added in 7.0 + Name.startswith("avx512.mask.pmulh.w.") || // Added in 7.0 + Name.startswith("avx512.mask.pmulhu.w.") || // Added in 7.0 Name.startswith("avx512.mask.packsswb.") || // Added in 5.0 Name.startswith("avx512.mask.packssdw.") || // Added in 5.0 Name.startswith("avx512.mask.packuswb.") || // Added in 5.0 @@ -1963,6 +1966,46 @@ void llvm::UpgradeIntrinsicCall(CallInst *CI, Function *NewFn) { { CI->getArgOperand(0), CI->getArgOperand(1) }); Rep = EmitX86Select(Builder, CI->getArgOperand(3), Rep, CI->getArgOperand(2)); + } else if (IsX86 && Name.startswith("avx512.mask.pmul.hr.sw.")) { + VectorType *VecTy = cast<VectorType>(CI->getType()); + Intrinsic::ID IID; + if (VecTy->getPrimitiveSizeInBits() == 128) + IID = Intrinsic::x86_ssse3_pmul_hr_sw_128; + else if (VecTy->getPrimitiveSizeInBits() == 256) + IID = Intrinsic::x86_avx2_pmul_hr_sw; + else if (VecTy->getPrimitiveSizeInBits() == 512) + IID = Intrinsic::x86_avx512_pmul_hr_sw_512; + else + llvm_unreachable("Unexpected intrinsic"); + + Rep = Builder.CreateCall(Intrinsic::getDeclaration(F->getParent(), IID), + { CI->getArgOperand(0), CI->getArgOperand(1) }); + Rep = EmitX86Select(Builder, CI->getArgOperand(3), Rep, + CI->getArgOperand(2)); + } else if (IsX86 && (Name.startswith("avx512.mask.pmulh.w.") || + Name.startswith("avx512.mask.pmulhu.w."))) { + bool IsUnsigned = Name[17] == 'u'; + VectorType *VecTy = cast<VectorType>(CI->getType()); + Intrinsic::ID IID; + if (!IsUnsigned && VecTy->getPrimitiveSizeInBits() == 128) + IID = Intrinsic::x86_sse2_pmulh_w; + else if (!IsUnsigned && VecTy->getPrimitiveSizeInBits() == 256) + IID = Intrinsic::x86_avx2_pmulh_w; + else if (!IsUnsigned && VecTy->getPrimitiveSizeInBits() == 512) + IID = Intrinsic::x86_avx512_pmulh_w_512; + else if (IsUnsigned && VecTy->getPrimitiveSizeInBits() == 128) + IID = Intrinsic::x86_sse2_pmulhu_w; + else if (IsUnsigned && VecTy->getPrimitiveSizeInBits() == 256) + IID = Intrinsic::x86_avx2_pmulhu_w; + else if (IsUnsigned && VecTy->getPrimitiveSizeInBits() == 512) + IID = Intrinsic::x86_avx512_pmulhu_w_512; + else + llvm_unreachable("Unexpected intrinsic"); + + Rep = Builder.CreateCall(Intrinsic::getDeclaration(F->getParent(), IID), + { CI->getArgOperand(0), CI->getArgOperand(1) }); + Rep = EmitX86Select(Builder, CI->getArgOperand(3), Rep, + CI->getArgOperand(2)); } else if (IsX86 && Name.startswith("avx512.mask.pack")) { bool IsUnsigned = Name[16] == 'u'; bool IsDW = Name[18] == 'd'; |
