llvm-svn: 38443

llvm-svn: 38442

often.  This improves Anton's testcase from 36s to 32s.

llvm-svn: 38441

llvm-svn: 38440

in the set.  For the moment, this results in a slight performance decrease, but
it lays the groundwork for future improvements.

llvm-svn: 38439

llvm-svn: 38438

llvm-svn: 38437

absolute path to the llvm.css file.

llvm-svn: 38436

llvm-svn: 38435

llvm-svn: 38434

llvm-svn: 38433

llvm-svn: 38432

llvm-svn: 38431

llvm-svn: 38430

documentation is working.

llvm-svn: 38429

of llvm from that level.

llvm-svn: 38423

llvm-svn: 38422

the same value number.  This fixes an infinite loop on 444.namd.

llvm-svn: 37967

llvm-svn: 37966

llvm-svn: 37965

llvm-svn: 37964

llvm-svn: 37963

llvm-svn: 37962

llvm-svn: 37961

llvm-svn: 37960

llvm-svn: 37959

alias analysis, adding caching and lazy computation of queries.  This will
be used in planned improvements to memory access optimizations.

llvm-svn: 37958

llvm-svn: 37957

llvm-svn: 37956

llvm-svn: 37955

llvm-svn: 37954

llvm-svn: 37953

llvm-svn: 37952

for the GVN part of GVNPRE to apply.

llvm-svn: 37951

llvm-svn: 37950

critical edges be split before we begin.

llvm-svn: 37949

llvm-svn: 37948

so must be lowered to a value, not nothing at all.
Subtle point: I made eh_selector return 0 and
eh_typeid_for return 1.  This means that only
cleanups (destructors) will be run as the exception
unwinds [if eh_typeid_for returned 0 then it would
be as if the first catch always matched, and the
corresponding handler would be run], which is
probably want you want in the CBE.

llvm-svn: 37947

llvm-svn: 37946

files.
bitcode files are the only LLVM format left.

llvm-svn: 37945