|
1)
Message boards :
Questions and problems :
one thread of project by CPU
(Message 37370)
Posted 2 Apr 2011 by Einstein-Rosen-Podolsky Post: Hello, is there a solution to have compute a thread of project on ONE core CPU ? i have two project: cosmology and ClimatePrediction.net i would two units by project for engaged my 4 CPU cores, for any project be computed. Actually Projects send us 2 or more wu's. I don't know how limited send to one wu to compute. my example: cosmology: 4wu's (3 computed and one wait) - climate: 4wu's (one computed and 3 wait). ------------------------------ Bonjour, y at-il une solution pour avoir calculer un fil de projet sur un core CPU? J'ai deux projets: la cosmologie et ClimatePrediction.net Je voudrais deux unités par projet retenu mon CPU 4 cœurs, pour tout projet sera calculé. En fait, les projets nous envoyer 2 ou plus de wu. Je ne sais pas comment limitée envoyer à un wu à calculer. mon exemple: la cosmologie: 4wu de (3 calculés et une attente) - Climat: les 4wu (celui calculé et 3 d'attente). |
|
2)
Message boards :
Questions and problems :
end of wu's
(Message 37368)
Posted 2 Apr 2011 by Einstein-Rosen-Podolsky Post: bonjour, dans mes projets BOINC, il y a plusieurs unités de travail qui finissent après 1h de travail, souvent 1h et quelques mns. Sera-t'il possible, dans les options, d'ajouter l'option "laisser finir la wu avant de calculer une autre unité de travail". je sais, je pourrais aussi augmenter la durée dans "switch between" automatic english translate: hello, in my BOINC projects, there are several units of work that end after 1 hour of work, often a few mins and 1 hr. Will it be possible in the options, adding the option "let him finish before calculating the wu another work unit. I know, I could also increase the time to "switch between " |
|
3)
Message boards :
Questions and problems :
end of wu's
(Message 37363)
Posted 1 Apr 2011 by Einstein-Rosen-Podolsky Post: Hello, in options of BOINC, is it possible to define end of work unit after this wu's is finish of compute ? i compute on home server, many wu's finished computed after more 1 hour (1h05mns by example) |
|
4)
Message boards :
GPUs :
tow gpu _ boinc and folding
(Message 37238)
Posted 20 Mar 2011 by Einstein-Rosen-Podolsky Post: google translate french to english: http://forum.boinc-af.org/index.php/topic,4132.0.html Hello, 1_ I would like to run my GPU with BOINC and folding one after another on my HD4770 (in a few months to go for the HD5770 core v7 folding), but in the preferences of BOINC manager there is just "use gpu when the computer is used. No possibility of computed between 6am and 22h for example. 2_ Will there see this option in the next version of boinc (6.12.18). 3_ In addition, currently in single gpu, I wonder if he can just use a GPU that he has indicated, and then leave the other. |
|
5)
Message boards :
BOINC Manager :
My Wish List - part 3.
(Message 36114)
Posted 17 Dec 2010 by Einstein-Rosen-Podolsky Post: Hello, i'm with BOINC 6.10.58, i wich BOINC to can to change task of project to calculate when this task in running is finish, without switch with hours (switch task: 60mns). In french: Je souhaite que BOINC ne change de projet à calculé QUE lorsque la tache du projet en cours de calcul est terminé. Sans utilisé les heures (switch task: 60mns). ----------------------------------------------------------------------------------- switch task: 60mns milkyway: 5mns climate.net: 90days to calculate severales tasks of milkyway; but to calculate all the one task of climate.net wich very long. Running on server. i have a quad cores CPU |
|
6)
Message boards :
Questions and problems :
Can You Dedicate One Processor On A Quad To Handle A Lg WU?
(Message 19395)
Posted 9 Aug 2008 by Einstein-Rosen-Podolsky Post: CPU affinity will be integrated to BOINC 6.x or 7.x.x official ? For the quad or octo CPU as Nehalem, for to do to work a wu or a project on one core, not on 4 or 8 cores. The projects are not optimised for work on 8 cores. ps: i'm french |
|
7)
Message boards :
BOINC client :
(solved)migrating to 6.2.14 Benchmark problem
(Message 19394)
Posted 9 Aug 2008 by Einstein-Rosen-Podolsky Post: i've an error: 09/08/2008 20:38:56||Running CPU benchmarks 09/08/2008 20:38:56||Suspending computation - running CPU benchmarks 09/08/2008 20:39:27||[error] FP benchmark ran only 2.906250 sec; ignoring 09/08/2008 20:39:27||[error] CPU benchmarks error when folding_cpu run |
|
8)
Message boards :
BOINC client :
BOINC 6 & GPU CUDA
(Message 17339)
Posted 14 May 2008 by Einstein-Rosen-Podolsky Post: OK, but why to write GPU and CUDA ?! CUDA is a GPU specificly created for paralleling computing, and "NVIDIA CUDA™ Technology is the world’s only C language environment that provides access to processing power of NVIDIA GPUs. It enables developers to utilize NVIDIA GPUs to solve the most complex computation-intensive challenges such as oil and gas exploration, financial risk management, product design, medical imaging, and scientific research" http://www.nvidia.com/object/cuda_home.html i don't understand all :-( |
|
9)
Message boards :
BOINC client :
BOINC 6 & GPU CUDA
(Message 17329)
Posted 14 May 2008 by Einstein-Rosen-Podolsky Post: Hello, i saw that in a log: "client: for CUDA, don't confuse coproc.name ("CUDA") from coproc.prop.name (e.g., "Device Emulation (CPU)" on my PC, which doesn't have a recent-enough gpu)" What's mean ? if i have a good GPU Nvidia not very old, like 7xxx or 8xxx or quadro or CUDA, i can use power of my GPU for computing in BOINC in parallel of my CPU ? |
Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License,
Version 1.2 or any later version published by the Free Software Foundation.