Join GitHub today
GitHub is home to over 31 million developers working together to host and review code, manage projects, and build software together.
Sign upuse -O0 when building release LLVM on sparc #61
Conversation
This comment has been minimized.
This comment has been minimized.
|
Weird! Is the resulting LLVM usable? I figured that an LLVM at O0 would be unbearably slow... |
This comment has been minimized.
This comment has been minimized.
|
It's usable with -O0, just really slow (but not quite as slow as a normal debug build). But considering that it doesn't work at all otherwise, there's not much choice. Investigation is being done with one of the gcc maintainers. It's possible that some other combination of options might work here, or that I could potentially limit the application of -O0 to specific files. If you'd prefer to wait until I can refine this further to minimize the performance impact, that's perfectly fine with me. While I'm waiting on a response from the gcc maintainer, I'm investigating other combinations of options to see if they can workaround the issue. But the change I proposed here is the only one that works reliably so far. |
This comment has been minimized.
This comment has been minimized.
|
Ok sounds reasonable to me! |
binarycrusader commentedFeb 21, 2017
The bits produced by gcc above -O0 when building rust's private copy of llvm on
sparc result in either versions of llvm-tblgen that consistently segfault or
produce otherwise unuseable code. This does not affect cross-compiles for some
reason. Investigation with gcc upstream is in progress.