Bill Wendling 984f0ce06b Fix for PR929. The PHI nodes were being gone through for each instruction ...

in a successor block for every block...resulting in some O(N^k) algorithm
which wasn't very good for performance. Calculating this information up
front and keeping it in a map made it much faster.

llvm-svn: 30697

2006-10-03 07:20:20 +00:00

18 KiB

Raw Blame History

View Raw