Message boards :
Number crunching :
Different setting for different GPUs?
Message board moderation
Author | Message |
---|---|
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
Ok, what I want to know. Is it possible to run different number of work units on different GPUs within the same Computer? I have a mixed GPU with both NV 750 and 780. From what I see the 750 take 2 times as long as the 780. What I would like to do is run 2 work units on the 780 and 1 work unit on the 750. Can that be done? Just curious. Zalster |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
The easy way to do that is start 2 instances of Boinc on the same host, each one optimized for one of the GPU´s. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14649 Credit: 200,643,578 RAC: 874 |
The easy way to do that is start 2 instances of Boinc on the same host, each one optimized for one of the GPU´s. I don't think anyone except you calls that 'easy' :P - though I don't doubt it's effective, but with side-effects. |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
The easy way to do that is start 2 instances of Boinc on the same host, each one optimized for one of the GPU´s. It is pretty easy to setup, but BOINC can be rather temperamental in that situation. I had a few issues the last time I was doing that with BOINC 7.0.64. -Sometimes an instance would generate a new CPID -Tasks run on one instance would be reported as valid on the other running instance. -One instance would just stop running tasks with no log as to why. To get it running again all BOINC instances would have to be stopped and then started. I figured I might try it again with a newer version at some point. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
What about using the <coproc> inside of the app_config file. I was looking at that and trying to figure out if I could add that into the app_config and tell it if GPU 0 or GPU 1 then run this many, if GPU 2 or GPU 3 then run this.. I'm not sure how to fix it into the app_config. What I was thinking was along these lines. <app_config> <app> <name>astropulse_v6</name> <coproc> <type>GTX 780</type> <count>2</count> <device_nums>0 1</device_nums> <gpu_versions> <gpu_usage>0.5</gpu_usage> <cpu_usage>1</cpu_usage> </gpu_version> </coproc> <coproc> <type>GTX 750</type> <count>2</count> <device_nums>2 3</device_nums> <gpu_versions> <gpu_usage>1</gpu_usage> <cpu_usage>1</cpu_usage> </gpu_version> </coproc> </app> </app_config> Ok, I don't do code so I don't even know if what the above looks like is even remotely possible but that is why I am asking. Last thing I want to do is put this in the computer and have it go berzerk. Thoughts and Ideas? |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
What about using the <coproc> inside of the app_config file. I was looking at that and trying to figure out if I could add that into the app_config and tell it if GPU 0 or GPU 1 then run this many, if GPU 2 or GPU 3 then run this.. I'm not sure how to fix it into the app_config. What I was thinking was along these lines. You can add whatever you like to the app_config.xml. However since there is no code to read your made up values they will be ignored. With <coproc> declared twice I imagine the last one will take effect. In another post Raistmer was mentioning it would be nice if we had a per GPU config and I made this suggestion. Perhaps one of the Lunatics forwarded it or a better idea to the BOINC devs. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
I think that would be a great option if they could do it. Nice to know both of you were thinking already about that. Guess I'll wait and see if anything comes from that. Zalster |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
If you are on the BOINC Alpha list you could mention it also. I would think that the more times something is mentioned the more likely it is to happen. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
The easy way to do that is start 2 instances of Boinc on the same host, each one optimized for one of the GPU´s. I run multiple instances of Boinc in most of my hosts, realy works and allow you to mix any GPU family without be worry about that. That´s why i say "the easy way" not say it´s easy to do that. Something i need to agree, it´s for advanced users only. But i belive only advanced users worry about configurate separately each GPU. Most of the "Normal" users don´t need or worry about that just set and runs with defoult parameters. There are a very good "how to do" in Vypers site: http://vyper.kafit.se/wp/index.php/2011/02/04/running-different-nvidia-architectures-most-optimal-at-setihome/ |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Exactly. And with latest issues with crazy spam filters @different mail providers mentioning from different sides is especially actual. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
And to please "normal" users more in APv7 defaults will be scalable, different for different GPUs. So more powerful GPUs now will be more close to sweet spot of best performance than before w/o any additional tweaking. For entusiast tweaking still possible of course, even more params added to improve fine tuning. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.