Author: Albert Bertilsson
Date: 11:23:12 12/11/03
Go up one level in this thread
On December 11, 2003 at 14:18:00, Aaron Gordon wrote: >On December 11, 2003 at 13:50:57, Albert Bertilsson wrote: > >>On December 11, 2003 at 13:34:36, Aaron Gordon wrote: >> >>>On December 11, 2003 at 12:39:59, Albert Bertilsson wrote: >>> >>>>On December 11, 2003 at 12:34:33, Aaron Gordon wrote: >>>> >>>>>On December 11, 2003 at 07:08:30, Albert Bertilsson wrote: >>>>> >>>>>>Hi! >>>>>> >>>>>>If you have any wasted cpu-resources join my distributed perft project to >>>>>>calculate perft(11). Any contribution large or small is welcome. >>>>>> >>>>>>The project is currently under testing to see how many will join (I'll not do it >>>>>>myself) and to make sure that the server is working ok on the internet, I've >>>>>>just tested it on LAN before. >>>>>> >>>>>>You need .Net framework from Microsoft to participate. >>>>>> >>>>>>Home page: >>>>>>http://www.albert.nu/programs/dperft >>>>>> >>>>>>/Regards Albert >>>>> >>>>>Can I put multiple computers doing this and have it use the same user name (and >>>>>still have it do different work units)? Or will I need to make a different user >>>>>(MoD, MoD2, etc) for each computer? Thanks. >>>> >>>>There is no technical problem with having multiple computers doing multiple >>>>workloads using the same username and password. But for statistics I recomend >>>>you to have diffrent names (password could be the same). For example I use >>>>AlbertProg and AlbertOffice to separate my two computers (otherwise they would >>>>be visible as one very powerful computer). If you have a multiple CPU machine >>>>I'd recommend that you use the same name/pass for all the clients running on >>>>that machine (no multi-cpu support). >>> >>>Ah, ok. I ended up putting all of my available computers that aren't already >>>doing something on your project. Hopefully it helps a little bit, here is what I >>>have so far on the username MoD: >>> >>>Athlon XP 2.5GHz w/ 512mb hash >>>Athlon XP 2.5GHz w/ 128mb hash >>>Athlon XP 1.6GHz w/ 128mb hash >>>Athlon TB 1.2GHz w/ 256mb hash >>>Celeron-2 1.0GHz w/ 128mb hash >>> >>>Also I was wondering about something... Would you be able to add a benchmark >>>mode into your current client, that way we can do some speed testing between the >>>various cpus? Thanks. >> >>You just doubled the number of computers on the project (Assuming the other >>participants are using one computer each)! If you'd like to bench them for this >>project just start the engine and type "perfthash 8". This will be a pretty good >>benchmark for this problem (I guess you 2.5GHz 512MB Athlon) will do quite well. >>Post your results here, I'll collect them. >> >>A benchmark in the client GUI would be possible to implement, sending the result >>to the server and storing it there, but I think the manual way is good enough. >>If we go after perft(12) I'll update the client to include this feature. >> >>/Regards Albert > >Thanks :) Just curious, do you plan a linux version at any point? >Also, any idea how long the benchmark may take? I was hoping for something that >could be run in a 1-3 minute period. Thanks again... For a one minute benchmark you'll have to use perfthash 7, that should compute in less than a minute on your fast machines, and a couple of minutes on you slow machines. I'm hoping that Mono project will make it possible to run the client on linux, and porting Sharper will be easy I think. This is only an option if many join this project otherwise perft(12) will take far to long to complete. /Regards Albert
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.