Computer Chess Club Archives


Search

Terms

Messages

Subject: Timings for Bit Scans

Author: Matt Taylor

Date: 03:48:03 12/19/02


Athlon data:
http://65.186.75.165/~mtaylor/bitscan_athlon.html

I'll try to post source a little later on today. I have a 3-hr. drive ahead of
me, and I should arrive at my destination around 10-10:30 AM my time (EST).
Hopefully 11 AM-12 PM EST when it gets posted, but it could easily be 3 PM EST
if I have to clean up more code than I thought. I'll stick a link in the HTML
file.

Obviously I have a few timing issues left unresolved. One such issue involves
interrupts in the middle of a timed routine. (You will notice the occasional
test skyrocketing by a few thousand clocks.) For most of the other timings, the
average (computed over 128 trials) is equal to the min. or max., usually the
former. That indicates that the first time in, the function has some extra
latency. I'm warming the cache, so I'm not sure why this is happening.

If anyone wants to make suggestions for the code, I'd love to take them. Keep in
mind that this code is a hacked/adapted version of code I was using to test IPC
and processor efficiency.

-Matt



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.