[X86][SSE] Auto upgrade PADDUS/PSUBUS intrinsics to UADD_SAT/USUB_SAT generic intrinsics (clang)
Sibling patch to D55855, this emits UADD_SAT/USUB_SAT generic intrinsics for the SSE saturated math intrinsics instead of expanding to a IR code sequence that could be difficult to reassemble.
Differential Revision: https://reviews.llvm.org/D55879
llvm-svn: 349631
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 0f8daa6..486e7d8 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -9487,31 +9487,14 @@
return CGF.Builder.CreateSExt(Mask, DstTy, "vpmovm2");
}
-// Emit addition or subtraction with saturation.
-// Handles both signed and unsigned intrinsics.
-static Value *EmitX86AddSubSatExpr(CodeGenFunction &CGF, const CallExpr *E,
+// Emit addition or subtraction with unsigned saturation.
+// TODO: Handle signed intrinsics.
+static Value *EmitX86AddSubSatExpr(CodeGenFunction &CGF,
SmallVectorImpl<Value *> &Ops,
bool IsAddition) {
-
- // Collect vector elements and type data.
- llvm::Type *ResultType = CGF.ConvertType(E->getType());
-
- Value *Res;
- if (IsAddition) {
- // ADDUS: a > (a+b) ? ~0 : (a+b)
- // If Ops[0] > Add, overflow occurred.
- Value *Add = CGF.Builder.CreateAdd(Ops[0], Ops[1]);
- Value *ICmp = CGF.Builder.CreateICmp(ICmpInst::ICMP_UGT, Ops[0], Add);
- Value *Max = llvm::Constant::getAllOnesValue(ResultType);
- Res = CGF.Builder.CreateSelect(ICmp, Max, Add);
- } else {
- // SUBUS: max(a, b) - b
- Value *ICmp = CGF.Builder.CreateICmp(ICmpInst::ICMP_UGT, Ops[0], Ops[1]);
- Value *Select = CGF.Builder.CreateSelect(ICmp, Ops[0], Ops[1]);
- Res = CGF.Builder.CreateSub(Select, Ops[1]);
- }
-
- return Res;
+ Intrinsic::ID IID = IsAddition ? Intrinsic::uadd_sat : Intrinsic::usub_sat;
+ llvm::Function *F = CGF.CGM.getIntrinsic(IID, Ops[0]->getType());
+ return CGF.Builder.CreateCall(F, {Ops[0], Ops[1]});
}
Value *CodeGenFunction::EmitX86CpuIs(const CallExpr *E) {
@@ -11382,14 +11365,14 @@
case X86::BI__builtin_ia32_paddusw256:
case X86::BI__builtin_ia32_paddusb128:
case X86::BI__builtin_ia32_paddusw128:
- return EmitX86AddSubSatExpr(*this, E, Ops, true /* IsAddition */);
+ return EmitX86AddSubSatExpr(*this, Ops, true /* IsAddition */);
case X86::BI__builtin_ia32_psubusb512:
case X86::BI__builtin_ia32_psubusw512:
case X86::BI__builtin_ia32_psubusb256:
case X86::BI__builtin_ia32_psubusw256:
case X86::BI__builtin_ia32_psubusb128:
case X86::BI__builtin_ia32_psubusw128:
- return EmitX86AddSubSatExpr(*this, E, Ops, false /* IsAddition */);
+ return EmitX86AddSubSatExpr(*this, Ops, false /* IsAddition */);
}
}