[llvm] r366685 - AMDGPU/GlobalISel: Remove unnecessary code

[Matt Arsenault via llvm-commits]

Author: arsenm
Date: Mon Jul 22 06:05:25 2019
New Revision: 366685

URL: http://llvm.org/viewvc/llvm-project?rev=366685&view=rev
Log:
AMDGPU/GlobalISel: Remove unnecessary code

The minnum/maxnum case are dead, and the cvt is handled by the
default.

Modified:
    llvm/trunk/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp

Modified: llvm/trunk/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp?rev=366685&r1=366684&r2=366685&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp (original)
+++ llvm/trunk/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp Mon Jul 22 06:05:25 2019
@@ -567,10 +567,6 @@ bool AMDGPUInstructionSelector::selectG_
   MachineInstr &I, CodeGenCoverage &CoverageInfo) const {
   unsigned IntrinsicID =  I.getOperand(I.getNumExplicitDefs()).getIntrinsicID();
   switch (IntrinsicID) {
-  case Intrinsic::maxnum:
-  case Intrinsic::minnum:
-  case Intrinsic::amdgcn_cvt_pkrtz:
-    return selectImpl(I, CoverageInfo);
   case Intrinsic::amdgcn_if_break: {
     MachineBasicBlock *BB = I.getParent();
     MachineFunction *MF = BB->getParent();