Message boards : BOINC client : Cpu preferences
Message board moderation
Author | Message |
---|---|
Send message Joined: 22 Jun 09 Posts: 29 |
Is there a way to tell the client to not crunch CPU units? Basically I have one PC crunching Milkyway@Home and Collatz Conjecture on a ATI GPU and another machine crunching Seti on a CPU, Now the machine that does Seti is regularly sat there doing nothing waiting for there servers to come back up so I want to put ether Milkyway or Collatz on there as well for when its idle, however I cant figure out how to set it so that machine can crunch it on the CPU but stop the other machine from downloading and doing CPU units at the same time? |
Send message Joined: 23 Apr 07 Posts: 1112 |
You're got a total of 4 venues, move which ever PC you want to have different preferences to an unused venue at each project, then go and set for that venue computing preferences and project preferences to suit. Claggy |
Send message Joined: 29 Aug 05 Posts: 15575 |
In the project preferences at Collatz and Milkyway (needs to be done at both separately), set to use CPU = No, Nvidia GPU = No and ATI GPU = Yes. You don't have to set this at Seti, since they don't have an ATI application and thus no option to disable anything running on the ATI GPU. |
Send message Joined: 22 Jun 09 Posts: 29 |
Cheers, how do I set each PC to a location?, i.e have the PC with the ATI card as home and the PC crunching on the CPU to work? |
Send message Joined: 23 Apr 07 Posts: 1112 |
Cheers, how do I set each PC to a location?, i.e have the PC with the ATI card as home and the PC crunching on the CPU to work? Go to your computer's Details page at each project, right at the bottom is a location selection, Claggy |
Send message Joined: 22 Jun 09 Posts: 29 |
Yea saw it on there, how do I tell the local clients which profile to use? |
Send message Joined: 23 Apr 07 Posts: 1112 |
Yea saw it on there, how do I tell the local clients which profile to use? Select whichever location you want for each computer on the Computer Details page, then on the client, select the project, and click update, if it worked, then in the messages tab there will be a message about the new location, Claggy |
Send message Joined: 22 Jun 09 Posts: 29 |
Cheers, I might be being thick here but how then does the computer know which profile its to use? |
Send message Joined: 23 Apr 07 Posts: 1112 |
Go to your 'Your computer' page, then on the left, underneath where it says Computer ID, there's links to each computer's Detail's page, and tasks page's, click where it says 'Details' At the bottom of the Details page change the Location selection to the location desired, then click update, Claggy |
Send message Joined: 22 Jun 09 Posts: 29 |
Yea got the bit on the website, but the local client, how do you configure that to run the selected location? |
Send message Joined: 22 Jun 09 Posts: 29 |
Ahh ok I see how its done for Seti, but doenst seem to be the same on Milkyway@home or Collantz? |
Send message Joined: 22 Jun 09 Posts: 29 |
Ahh found it, thanks guys |
Send message Joined: 23 Apr 07 Posts: 1112 |
Yea got the bit on the website, but the local client, how do you configure that to run the selected location? On the client, click update for each project, it'll see it's location has changed, and use those preferences, Just make sure you've set up preferences for each project for that location, ie, SETI@home preferences, Collatz Conjecture preferences Claggy |
Send message Joined: 22 Jun 09 Posts: 29 |
Ok looks like I have it sorted, one last thing though, the latest project I add, first Milky way then Collantz's work units show as high priority and get done first even though other units were mid run? |
Send message Joined: 23 Apr 07 Posts: 1112 |
That's the nature of the Boinc client, it will only do GPU tasks in FIFO (First In, First Out) order, unless there's possible deadline problems, it'll do tasks that aren't in obvious deadline trouble, but might be in another few days/weeks/etc, before it'll switch over to tasks that are getting close to deadline, the way round it is to run a smaller cache, or just run one GPU project at a time, Claggy |
Copyright © 2025 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.