Messages in this thread | | | Date | Tue, 17 Nov 1998 01:09:43 -0800 | From | "David S. Miller" <> | Subject | Re: The Kommunity vs. Dick Johnson |
| |
Date: Tue, 17 Nov 1998 03:16:25 GMT From: "Theodore Y. Ts'o" <tytso@mit.edu>
The MD5 algorithm is just one example which is proof-positive that gcc can't find a register allocation scheme which allows all of the MD5 variables to be left in registers, and a human-coded MD5 algorithm handily proves that it can be found --- it's just that GCC can't find it.
Those of us who work on GCC in it's current state realize that register allocation technology is the one place where the compiler is still 10 years or so out of date. A graph coloring register allocation scheme plus even a half-assed forward-substitution pass added to reload would improve things significiantly. And this is what I spend my spare (gasp) time working on these days...
A few compilers designed on and for the original IBM 370 series machines have a more effective register allocater than anything GCC has ever posessed.
Later, David S. Miller davem@dm.cobaltmicro.com
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |