Transpose the calculation of spill weights such that we are calculating one

register at a time. This turns out to be slightly faster than iterating over
instructions, but more importantly, it allows us to compute spill weights for
new registers created after the spill weight pass has run.

Also compute the allocation hint at the same time as the spill weight. This
allows us to use the spill weight as a cost metric for copies, and choose the
most profitable hint if there is more than one possibility.

The new hints provide a very small (< 0.1%) but universal code size improvement.

llvm-svn: 110631

1 files changed, 0 insertions, 1 deletions

diff --git a/llvm/lib/CodeGen/SimpleRegisterCoalescing.h b/llvm/lib/CodeGen/SimpleRegisterCoalescing.h
index 321cef8a38d..855bdb98b36 100644
--- a/llvm/lib/CodeGen/SimpleRegisterCoalescing.h
+++ b/llvm/lib/CodeGen/SimpleRegisterCoalescing.h
@@ -47,7 +47,6 @@ namespace llvm {
     const MachineLoopInfo* loopInfo;
     AliasAnalysis *AA;
     
-    BitVector allocatableRegs_;
     DenseMap<const TargetRegisterClass*, BitVector> allocatableRCRegs_;
 
     /// JoinedCopies - Keep track of copies eliminated due to coalescing.