llvm-project

Go to file

Quentin Colombet 8aa7abe2ae Modify how the formulae are rated in Loop Strength Reduce.

Namely, check if the target allows to fold more that one register in the
addressing mode and if yes, adjust the cost accordingly.

Prior to this commit, reg1 + scale * reg2 accesses were artificially preferred
to reg1 + reg2 accesses. Indeed, the cost model wrongly assumed that reg1 + reg2
needs a temporary register for the computation, whereas it was correctly
estimated for reg1 + scale * reg2.

<rdar://problem/13973908>

llvm-svn: 183021

2013-05-31 17:20:29 +00:00

clang

Testcase for r183015.

2013-05-31 17:01:30 +00:00

clang-tools-extra

cpp11-migrate: Add option to detect and use macros that expand to 'override'

2013-05-31 13:20:24 +00:00

compiler-rt

[ASan] Do not buffer stdin in asan_symbolize.py

2013-05-31 14:55:06 +00:00

debuginfo-tests

Remove IR scenario tests.