Author: Steven Edwards
Date: 18:33:40 07/06/05
Go up one level in this thread
On July 06, 2005 at 13:53:12, Lance Perkins wrote: >If your single-cpu is enough for all your testing as you claim, then good luck >to you. I don't rely on good luck as my 36 years of programming experience is sufficient for my needs. I said that a single CPU is sufficient given proper mutex usage. Proof by example: Symbolic's toolkit was designed for multithreading from the onset. It has only one nontrivial global variable: the thread descriptor/task status table and it has a proper mutex guard. When running, the program has from three to twelve threads and has never had a thread fault or any other kind of fault that showed up on a dual that didn't show up on a single. Indeed, the thread related code would be scored the highest using a functionality/fault metric ratio. >I'm done here. We write more complex code than 10 lines of LISP. I'd rather write ten new lines of Lisp with three bugs than to re-write 100K lines of code that's been done a hundred times before.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.