[llvm] r352670 - [GlobalISel] Add IRTranslator support for @llvm.sqrt -> G_FSQRT

[Jessica Paquette via llvm-commits]

Author: paquette
Date: Wed Jan 30 12:58:14 2019
New Revision: 352670

URL: http://llvm.org/viewvc/llvm-project?rev=352670&view=rev
Log:
[GlobalISel] Add IRTranslator support for @llvm.sqrt -> G_FSQRT

Follow-up commit to https://reviews.llvm.org/D57359. (r352668)

This adds IRTranslator support for recognising a @llvm.sqrt intrinsic and
translating it into a G_FSQRT.

https://reviews.llvm.org/D57360

Modified:
    llvm/trunk/lib/CodeGen/GlobalISel/IRTranslator.cpp
    llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-irtranslator.ll

Modified: llvm/trunk/lib/CodeGen/GlobalISel/IRTranslator.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/GlobalISel/IRTranslator.cpp?rev=352670&r1=352669&r2=352670&view=diff
==============================================================================
--- llvm/trunk/lib/CodeGen/GlobalISel/IRTranslator.cpp (original)
+++ llvm/trunk/lib/CodeGen/GlobalISel/IRTranslator.cpp Wed Jan 30 12:58:14 2019
@@ -1086,6 +1086,11 @@ bool IRTranslator::translateKnownIntrins
         .addDef(getOrCreateVReg(CI))
         .addUse(getOrCreateVReg(*CI.getArgOperand(0)));
     return true;
+  case Intrinsic::sqrt:
+    MIRBuilder.buildInstr(TargetOpcode::G_FSQRT)
+        .addDef(getOrCreateVReg(CI))
+        .addUse(getOrCreateVReg(*CI.getArgOperand(0)));
+    return true;
   }
   return false;
 }

Modified: llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-irtranslator.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-irtranslator.ll?rev=352670&r1=352669&r2=352670&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-irtranslator.ll (original)
+++ llvm/trunk/test/CodeGen/AArch64/GlobalISel/arm64-irtranslator.ll Wed Jan 30 12:58:14 2019
@@ -2315,3 +2315,11 @@ define float @test_sin_f32(float %x) {
   %y = call float @llvm.sin.f32(float %x)
   ret float %y
 }
+
+declare float @llvm.sqrt.f32(float)
+define float @test_sqrt_f32(float %x) {
+  ; CHECK-LABEL: name:            test_sqrt_f32
+  ; CHECK: %{{[0-9]+}}:_(s32) = G_FSQRT %{{[0-9]+}}
+  %y = call float @llvm.sqrt.f32(float %x)
+  ret float %y
+}