forked from OSchip/llvm-project
8084dbaf8e
unrolling threshold to the optimize-for-size threshold. Basically, for loops containing calls, unrolling can still be profitable as long as the loop is REALLY small. llvm-svn: 113439 |
||
---|---|---|
.. | ||
Hello | ||
IPO | ||
InstCombine | ||
Instrumentation | ||
Scalar | ||
Utils | ||
Makefile |