Re: Graph coloring and JIT compilers.

"oliverhunt@gmail.com" <oliverhunt@gmail.com>
3 Apr 2006 01:33:12 -0400

          From comp.compilers
Related articles
Graph coloring and JIT compilers. julian_solo13@hotmail.com (Poseidon13) (2006-03-29)
Re: Graph coloring and JIT compilers. dnovillo@redhat.com (Diego Novillo) (2006-04-01)
Re: Graph coloring and JIT compilers. aghuloum@cs.indiana.edu (Abdulaziz Ghuloum) (2006-04-01)
Re: Graph coloring and JIT compilers. bonzini@gnu.org (Paolo Bonzini) (2006-04-01)
Re: Graph coloring and JIT compilers. torbenm@app-4.diku.dk (2006-04-03)
Re: Graph coloring and JIT compilers. richard@imagecraft.com (Richard) (2006-04-03)
Re: Graph coloring and JIT compilers. oliverhunt@gmail.com (oliverhunt@gmail.com) (2006-04-03)
Re: Graph coloring and JIT compilers. anton@mips.complang.tuwien.ac.at (2006-04-03)
Graph coloring and JIT compilers. inderaj@gmail.com (Inderaj Bains) (2006-04-03)
Re: Graph coloring and JIT compilers. anton@mips.complang.tuwien.ac.at (2006-04-08)
| List of all articles for this month |
From: "oliverhunt@gmail.com" <oliverhunt@gmail.com>
Newsgroups: comp.compilers
Date: 3 Apr 2006 01:33:12 -0400
Organization: http://groups.google.com
References: 06-03-097
Keywords: optimize, registers
Posted-Date: 03 Apr 2006 01:33:12 EDT

Well there's also linear scan register allocation, which is meant to be
fairly close (performance wise) to graph colouring (there's a paper at
http://www.research.ibm.com/jalapeno/papers/toplas99.pdf though google
should return a number of others).


I haven't personally implemented any form of register allocation as I'm
not writing a JIT copmiler, so can safely be lazy and target higher
level compilers like C/C#/etc that do all that hard work for me :)


Cheers,
    Oliver


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.