[PATCH] D31790: [PowerPC] Assume intrinsics with 128bit integer operands use CTR

[Hal Finkel via Phabricator via llvm-commits]

hfinkel added a comment.

> A somewhat minified example IR which reproduces the assertion is attached to the linked bug report -- here I would also appreciate some guidance on in what form / where this can be integrated in the regression tests.

Just running llc on the resulting IR in the usual way should be fine.



================
Comment at: lib/Target/PowerPC/PPCCTRLoops.cpp:250
         if (F->getIntrinsicID() != Intrinsic::not_intrinsic) {
+          // Preemtively assume intrinsics operating on 128 integers use CTR
+          for (unsigned OID = 0; OID < CI->getNumArgOperands(); OID++) {
----------------
It looks like we can have a more-targeted fix than this. The problem seems to be restricted to multiply. See below.


================
Comment at: lib/Target/PowerPC/PPCCTRLoops.cpp:317
           case Intrinsic::maxnum:    Opcode = ISD::FMAXNUM;    break;
           }
         }
----------------
Here add:

case Intrinsic::umul_with_overflow: Opcode = ISD::UMULO; break;
case Intrinsic::smul_with_overflow: Opcode = ISD::SMULO; break;

and I think that should take care of it.



https://reviews.llvm.org/D31790