Message boards :
Questions and problems :
Slots - usage
Message board moderation
Author | Message |
---|---|
Send message Joined: 29 Aug 05 Posts: 15480 |
Both single and multi-core jobs run from a single slot. One slot per job, doesn't matter how much resources the job uses. |
Send message Joined: 29 Aug 05 Posts: 15480 |
So he runs two or more 'tasks' at the same time. One task per VM, one VM per slot. |
Send message Joined: 29 Aug 05 Posts: 15480 |
I have emailed one of the developers, but due to time differences and him working elsewhere than on BOINC these days, it can take hours or days for him to answer. |
Send message Joined: 29 Aug 05 Posts: 15480 |
All right, according to the BOINC developers, this is an artifact of VMs failing. It's not a bug, and doesn't mean that a task actually uses 2 slots. |
Send message Joined: 3 Apr 14 Posts: 10 |
Both single and multi-core jobs run from a single slot. One slot per job, doesn't matter how much resources the job uses. So does that mean that when task x is suspended in favor of a higher priority task, task x has to wait until its original slot is available again in order to be resumed? |
Send message Joined: 29 Aug 05 Posts: 15480 |
So does that mean that when task x is suspended in favor of a higher priority task, task x has to wait until its original slot is available again in order to be resumed? A task stays in its slot until it has met a boinc_finish() and is unloaded. It's not that when a task suspends that it releases the slot it's in, because we only have 4 or 8 slots. There's no real limit to slot directories that can be made and used/be in use. So tasks stay in their own slot from their start to their finish, no matter how many times you suspend them in between. |
Send message Joined: 3 Apr 14 Posts: 10 |
It's not that when a task suspends that it releases the slot it's in, because we only have 4 or 8 slots. There's no real limit to slot directories that can be made and used/be in use. So tasks stay in their own slot from their start to their finish, no matter how many times you suspend them in between. Thanks for the enlightenment! |
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.