Linked by Thom Holwerda on Tue 11th Nov 2008 11:44 UTC, submitted by tyrione
LLVM 2.4 has been released. "LLVM 2.4 includes many bug fixes, much faster compile times at -O0, substantially better code generation in various cases, a new PIC16 target, new IR features, and numerous other improvements and features (see the release notes for details)." You can get it at the project's download page.
Thread beginning with comment 337019
To view parent comment, click here.
To read all comments associated with this story, please click here.
To view parent comment, click here.
To read all comments associated with this story, please click here.
Member since:
2006-01-24
adkilla wrote:
-"Anybody had any better luck? There has been claims that llvm-gcc produces code faster than vanilla gcc."
AFAIK llvm-gcc just uses gcc as a front end, the code optimization is all done by llvm. And yes, I've seen claims of better optimized code generation but I've yet to confirm it for myself (actually if anyone has some benchmarks or some such it would be very interesting). IIRC one of the authors of llvm claimed an overall 5%-10% speed increase in the generated code when compiling using llvm (which I personally think sounds almost too good) as compared to gcc. Of course both compilers are in heavy development and any such statements can be obsolete from day to day, but given that llvm is a relatively new project it's likely alot easier to implement new modern optimization techniques into it's codebase than into the much older and larger gcc toolchain which also has to make sure it doesn't break since basically the entire oss world depends on it.
I can't vouch for it's overall useability as I've yet to try it out, but given that it consists of many components I suppose some are more mature than others. For instance I believe the jit compiler is being used in OSX's opengl stack.