llvm-6502/lib/Transforms/InstCombine
Chris Lattner 979ed44feb Fix a serious performance regression introduced by r108687 on linux:
turning (fptrunc (sqrt (fpext x))) -> (sqrtf x)  is great, but we have
to delete the original sqrt as well.  Not doing so causes us to do 
two sqrt's when building with -fmath-errno (the default on linux).



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113260 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-07 20:01:38 +00:00
..
CMakeLists.txt
InstCombine.h
InstCombineAddSub.cpp
InstCombineAndOrXor.cpp
InstCombineCalls.cpp
InstCombineCasts.cpp Fix a serious performance regression introduced by r108687 on linux: 2010-09-07 20:01:38 +00:00
InstCombineCompares.cpp
InstCombineLoadStoreAlloca.cpp
InstCombineMulDivRem.cpp
InstCombinePHI.cpp
InstCombineSelect.cpp
InstCombineShifts.cpp
InstCombineSimplifyDemanded.cpp
InstCombineVectorOps.cpp
InstCombineWorklist.h
InstructionCombining.cpp
Makefile