[PowerPC] Use the two-constant NR algorithm for refining estimates

The single-constant algorithm produces infinities on a lot of denormal values.
The precision of the two-constant algorithm is actually sufficient across the
range of denormals. We will switch to that algorithm for now to avoid the
infinities on denormals. In the future, we will re-evaluate the algorithm to
find the optimal one for PowerPC.

Differential revision: https://reviews.llvm.org/D60037

llvm-svn: 360144

1 files changed, 3 insertions, 1 deletions

diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
index 3b61f4da351..9ff817e2f50 100644
--- a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
+++ b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
@@ -11145,7 +11145,9 @@ SDValue PPCTargetLowering::getSqrtEstimate(SDValue Operand, SelectionDAG &DAG,
     if (RefinementSteps == ReciprocalEstimate::Unspecified)
       RefinementSteps = getEstimateRefinementSteps(VT, Subtarget);
 
-    UseOneConstNR = true;
+    // The Newton-Raphson computation with a single constant does not provide
+    // enough accuracy on some CPUs.
+    UseOneConstNR = !Subtarget.needsTwoConstNR();
     return DAG.getNode(PPCISD::FRSQRTE, SDLoc(Operand), VT, Operand);
   }
   return SDValue();