Author: Robert Hyatt
Date: 12:38:07 09/04/02
Go up one level in this thread
On September 04, 2002 at 14:35:53, Roy Eassa wrote: >On September 03, 2002 at 11:13:02, Robert Hyatt wrote: > >>On September 02, 2002 at 02:11:54, Jouni Uski wrote: >> >>>. >> >> >>It is something written by Jason Deines a few years ago. It is basically >>an estimate of SMP efficiency, although I am not sure it is really very >>meaningful. >> >>It is computed as follows: >> >>(640.0 / (total_time_used / 100)) >> >>which means bigger numbers are better. 640 probably comes from some >>machine Jason used while modifying bench.c, and I would suspect it >>represents some "factor" of overall speedup compared to that machine, >>whatever it was... >> >>But as total-time-used shrinks, the above computation will produce larger >>and larger values... I will probably change the computation to produce >>something more meaningful, such as "average speedup" or something similar... > > >Pardon my math, but isn't that just 64000/total_time_used? yep. As I said, I don't know what the significance of the 640/64000 is. But that simply came right out of the bench.c source cod...
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.