Author: Rafael Andrist
Date: 11:17:17 04/16/02
Go up one level in this thread
On April 16, 2002 at 13:54:36, Slater Wold wrote: >On April 16, 2002 at 10:37:23, Rafael Andrist wrote: > >>On April 16, 2002 at 08:39:32, Gian-Carlo Pascutto wrote: >> >>>1) Would there be any advantage to my hash entries being 16 bytes instead of 14 >>>bytes? I'm thinking of possible alignment issues. The fact that I can have more >>>if they're only 14 bytes is irrelevant for me. >> >>I think you will have to reorganize your datastructure so that several >>14-bits-segments fit into a 8*n-segment (or 16*n, 32*n, 64*n - compiler >>dependent). >> >>>2) (For the Microsoft/Windows buffs) Is it possible to allocate more than 256M >>>in a contiguous chunck under older Windows versions? >> >>Under Win2000 with 1 GB RAM, I was able to allocate (malloc) around 900 MB, I >>wasn't able to allocate significantly more on a system with 2 GB. I haven't >>tried with other MS OS. Because this was for Nalimov-TB-Cache, I don't know >>wheter this is an OS dependent limit or caused by Nalimovs code. >> >>Rafael B. Andrist > > >An application can request up to 2GB of with malloc(). And it works in Windows >95/98 too. Or were you talking about 3.1? No, as said above, about Windows 2000. I know that this should theoretically possible, but as I tried it for the EGTB cache, I didn't work. I did not have the time to check why this happens. regards Rafael B. Andrist
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.