[X86] Teach how to custom lower double-to-half conversions under fast-math.

This patch teaches the backend how to expand a double-half conversion into
a double-float conversion immediately followed by a float-half conversion.
We do this only under fast-math, and if float-half conversions are legal
for the target.

Added test CodeGen/X86/fastmath-float-half-conversion.ll

Differential Revision: http://reviews.llvm.org/D7832

llvm-svn: 230276

2 files changed, 67 insertions, 0 deletions

diff --git a/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp b/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp
index ed337eb9648..a26f30739c7 100644
--- a/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp
@@ -3548,6 +3548,21 @@ void SelectionDAGLegalize::ExpandNode(SDNode *Node) {
     break;
   }
   case ISD::FP_TO_FP16: {
+    if (!TM.Options.UseSoftFloat && TM.Options.UnsafeFPMath) {
+      SDValue Op = Node->getOperand(0);
+      MVT SVT = Op.getSimpleValueType();
+      if ((SVT == MVT::f64 || SVT == MVT::f80) &&
+          TLI.isOperationLegalOrCustom(ISD::FP_TO_FP16, MVT::f32)) {
+        // Under fastmath, we can expand this node into a fround followed by
+        // a float-half conversion.
+        SDValue FloatVal = DAG.getNode(ISD::FP_ROUND, dl, MVT::f32, Op,
+                                       DAG.getIntPtrConstant(0));
+        Results.push_back(
+            DAG.getNode(ISD::FP_TO_FP16, dl, MVT::i16, FloatVal));
+        break;
+      }
+    }
+
     RTLIB::Libcall LC =
         RTLIB::getFPROUND(Node->getOperand(0).getValueType(), MVT::f16);
     assert(LC != RTLIB::UNKNOWN_LIBCALL && "Unable to expand fp_to_fp16");
diff --git a/llvm/test/CodeGen/X86/fastmath-float-half-conversion.ll b/llvm/test/CodeGen/X86/fastmath-float-half-conversion.ll
new file mode 100644
index 00000000000..29308735cca
--- /dev/null
+++ b/llvm/test/CodeGen/X86/fastmath-float-half-conversion.ll
@@ -0,0 +1,52 @@
+; RUN: llc -mtriple=x86_64-unknown-unknown -mattr=+f16c < %s | FileCheck %s --check-prefix=ALL --check-prefix=F16C
+; RUN: llc -mtriple=x86_64-unknown-unknown -mattr=+avx < %s | FileCheck %s --check-prefix=ALL --check-prefix=AVX
+
+define zeroext i16 @test1_fast(double %d) #0 {
+; ALL-LABEL: test1_fast:
+; F16C-NOT: callq {{_+}}truncdfhf2
+; F16C: vcvtsd2ss %xmm0, %xmm0, %xmm0
+; F16C-NEXT: vcvtps2ph $0, %xmm0, %xmm0
+; AVX: callq {{_+}}truncdfhf2
+; ALL: ret
+entry:
+  %0 = tail call i16 @llvm.convert.to.fp16.f64(double %d)
+  ret i16 %0
+}
+
+define zeroext i16 @test2_fast(x86_fp80 %d) #0 {
+; ALL-LABEL: test2_fast:
+; F16C-NOT: callq {{_+}}truncxfhf2
+; F16C: fldt
+; F16C-NEXT: fstps
+; F16C-NEXT: vmovss
+; F16C-NEXT: vcvtps2ph $0, %xmm0, %xmm0
+; AVX: callq {{_+}}truncxfhf2
+; ALL: ret
+entry:
+  %0 = tail call i16 @llvm.convert.to.fp16.f80(x86_fp80 %d)
+  ret i16 %0
+}
+
+define zeroext i16 @test1(double %d) #1 {
+; ALL-LABEL: test1:
+; ALL: callq  {{_+}}truncdfhf2
+; ALL: ret
+entry:
+  %0 = tail call i16 @llvm.convert.to.fp16.f64(double %d)
+  ret i16 %0
+}
+
+define zeroext i16 @test2(x86_fp80 %d) #1 {
+; ALL-LABEL: test2:
+; ALL: callq  {{_+}}truncxfhf2
+; ALL: ret
+entry:
+  %0 = tail call i16 @llvm.convert.to.fp16.f80(x86_fp80 %d)
+  ret i16 %0
+}
+
+declare i16 @llvm.convert.to.fp16.f64(double)
+declare i16 @llvm.convert.to.fp16.f80(x86_fp80)
+
+attributes #0 = { nounwind readnone uwtable "unsafe-fp-math"="true" "use-soft-float"="false" }
+attributes #1 = { nounwind readnone uwtable "unsafe-fp-math"="false" "use-soft-float"="false" }