News for hackers. Stuff that matters.
Re:Greedy Register Allocation in LLVM 3.0 by Rusky on Monday September 19, 2011 @08:13AM (#3013208)
I love reading about LLVM internals. It's exciting to see LLVM still making major algorithmic improvements to something as important as register allocation- it's a really nicely designed library both to use and to work on, and bringing the generated code's performance closer to compilers like GCC makes it even better.
Re:Greedy Register Allocation in LLVM 3.0 by danssig on Monday September 19, 2011 @10:22AM (#3013642)
I thought it had already largely surpassed GCC?
Re:Greedy Register Allocation in LLVM 3.0 by zmanji on Monday September 19, 2011 @10:39AM (#3013696)
It may have largely surpassed GCC in terms of how easy it is to integrate and compilation speed, but in quality of generated code it is lagging.