From 58dcd200b7f0ea01160b6159e0363cc96b1b83d9 Mon Sep 17 00:00:00 2001 From: Owen Anderson Date: Thu, 9 May 2013 22:27:13 +0000 Subject: [PATCH] Teach SelectionDAG to constant fold all-constant FMA nodes the same way that it constant folds FADD, FMUL, etc. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181555 91177308-0d34-0410-b5e6-96231b3b80d8 --- lib/CodeGen/SelectionDAG/SelectionDAG.cpp | 15 +++++++++++++++ test/CodeGen/X86/fma.ll | 8 ++++++++ 2 files changed, 23 insertions(+) diff --git a/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/lib/CodeGen/SelectionDAG/SelectionDAG.cpp index 15235c8ac307..f2c512dec7a5 100644 --- a/lib/CodeGen/SelectionDAG/SelectionDAG.cpp +++ b/lib/CodeGen/SelectionDAG/SelectionDAG.cpp @@ -3261,6 +3261,21 @@ SDValue SelectionDAG::getNode(unsigned Opcode, DebugLoc DL, EVT VT, // Perform various simplifications. ConstantSDNode *N1C = dyn_cast(N1.getNode()); switch (Opcode) { + case ISD::FMA: { + ConstantFPSDNode *N1CFP = dyn_cast(N1); + ConstantFPSDNode *N2CFP = dyn_cast(N2); + ConstantFPSDNode *N3CFP = dyn_cast(N3); + if (N1CFP && N2CFP && N3CFP) { + APFloat V1 = N1CFP->getValueAPF(); + const APFloat &V2 = N2CFP->getValueAPF(); + const APFloat &V3 = N3CFP->getValueAPF(); + APFloat::opStatus s = + V1.fusedMultiplyAdd(V2, V3, APFloat::rmNearestTiesToEven); + if (s != APFloat::opInvalidOp) + return getConstantFP(V1, VT); + } + break; + } case ISD::CONCAT_VECTORS: // A CONCAT_VECTOR with all operands BUILD_VECTOR can be simplified to // one big BUILD_VECTOR. diff --git a/test/CodeGen/X86/fma.ll b/test/CodeGen/X86/fma.ll index bd3514cc3f73..917eac0ca32d 100644 --- a/test/CodeGen/X86/fma.ll +++ b/test/CodeGen/X86/fma.ll @@ -34,6 +34,14 @@ entry: ret x86_fp80 %call } +; CHECK: test_f32_cst +; CHECK-NOT: fma +define float @test_f32_cst() nounwind readnone ssp { +entry: + %call = tail call float @llvm.fma.f32(float 3.0, float 3.0, float 3.0) nounwind readnone + ret float %call +} + declare float @llvm.fma.f32(float, float, float) nounwind readnone declare double @llvm.fma.f64(double, double, double) nounwind readnone declare x86_fp80 @llvm.fma.f80(x86_fp80, x86_fp80, x86_fp80) nounwind readnone