Author: Robert Hyatt
Date: 15:02:33 01/30/99
Go up one level in this thread
On January 30, 1999 at 13:08:37, Enrique Irazoqui wrote: >On January 30, 1999 at 12:44:17, Antonio Palma wrote: > >>But did you reboot and defragment memory ? > >I did reboot and load Memturbo. Is there any additional step needed to >defragment memory? I assumed it was done automatically. > >>I tried on Fritz 5.32 and Nimzo 99 and it works. > >With Fritz 5.32 I got 289 long Fritzmarks instead of 290 and 254 short instead >of 257, all with 192MB hashtables on a PII-400 with 256MB RAM. Rebel 10 went >exactly at the same speed with 120MB hashtables. > >Enrique > I don't want to post a long post on computer architecture to explain why this is a problem, but I certainly copy something I just wrote on the linux-smp mailing list for anyone interested in the 'problem'. In Enrique's case, he overlooked the point here... when you load a program into memory, sometimes you get a good physical memory layout. Othertimes you don't. If the program runs under dos, this is a moot point pretty much because memory is allocated linearly... Under windows it is a different thing completely. And you sometimes get stuck into pages that map well into the cache, other times you get stuck into a set of pages that map poorly into cache and you get lots of overwrites... >>On January 30, 1999 at 12:33:05, Enrique Irazoqui wrote: >> >>>On January 30, 1999 at 11:18:23, Antonio Palma wrote: >>> >>>>I have received e-mail of people interested about this ram defragmenter. >>>>I am waiting in this newsgroup for the follow-up about fritzmark obtained and >>>>the cpu to evaluate if this tool is really useful. >>> >>>I just tried it with Rebel 10, Fritz 5.32 and Junior 5, and I saw no improvement >>>in speed at all. I already uninstalled it. >>> >>>Enrique >>> >>>>Regards >>>>Antonio Palma
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.