Author: Inmann Werner
Date: 13:50:24 03/19/99
Go up one level in this thread
On March 19, 1999 at 16:44:47, Dann Corbit wrote: >On March 19, 1999 at 16:06:10, Inmann Werner wrote: > >>Hello. >> >>Today I changed my hashtables from fixed one ta allocated one. >> >>Example >>Old >> >>long hindex[65536] >>signed char hcolor[65536] >>signed short int value[65536] >>..... >> >>New >> >>long *hindex >>signed char *hcolor >>signed short int *value >>.... >> >>hindex=malloc(4*65536) >Suggestion: sizeof(long)*65536 is better if you want to use other machines. > >>hcolor=malloc(65536) >>value=malloc(2*65536) >Suggestion: sizeof(short)*65536 is better if you want to use other machines. > >>.... >> >>Works fine but: >> >>On my P90, the NPS decreases about 30% !!!!! >No surprise. P90 does not have a 64 bit memory transfer. You are doing a lot >more memory read writes. This wimpy chip can't handle it very well. > >>On my Cyrix 233 everything is fine (same speed) >No surprise, Cyrix 233 has a 64 bit memory transfer. > >>Do I anything extremly wrong or what happens here. Has this something to do >>with Prozessor cache or what? >Memory bandwidth. A big cache will help but not as much as a wider transfer. > >>Any suggestions?? >Don't worry about it. Nobody sells P90 any more. > You mean, all new prozessors can handle my "new code" ? But I really do not understand why this happens. Why do I a lot more memory read writes? The code in the program is the same as before, only the mallocs are new. Werner
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.