Message boards : BOINC client : Sending Deadline Info to Core Client
Message board moderation
Author | Message |
---|---|
Send message Joined: 17 Aug 09 Posts: 19 |
Hi all, I'm relatively new to PRC, so please forgive any naive observations/assumptions! Also, I hope this is the correct place for the post? I was wondering how BOINC currently handles resuming a suspended task/wu, either through user input or via detection of keyboard or mouse I/O. I can envisage a scenario where a wu is sent to a client, and suspended at some point. Does the client, upon resuming processing of this wu, check that the deadline has not passed, either through a scheduler contact (server determines the wu should be cancelled), or analysis by the core client of the suspend time vs start-time & deadline? Enabling client side deadline awareness would be a great way of ensuring that clients are not wasting time on wu's that are "dead" i.e. having been suspended for a long enough period of time to be considered un-returnable. If someone who is aware of the code base where this may be implemented could let me know whether this functionality exists, so that I could potentially work on a branch to include it? Regards Chris |
Send message Joined: 8 Jan 06 Posts: 36 |
CCs are already aware of what the deadline is for all tasks onboard. That gets sent when a task is assigned to the hosts. The CC will post a message in the logs about tasks which are over the deadline, advising that you may not get credit for it and might want to abort them. However you have to do that manually by design. Projects already have a couple of methods of trying to reduce the potential amount of time hosts can waste on useless tasks. The most commonly seen one is the '221' Redundant Task abort. What happens here is when a scheduler request comes in from a host, a list of all the tasks a host has onboard for the project is included. This list is checked against the BOINC DB and any tasks the host has which are no longer needed by the project and have not yet been started by the host will be aborted. The catch is the project has to have that functionality enabled. Most do, but some don't. The other alternative is the project has the ability to send an unconditional abort to a host. However, AFAIK no project uses this feature except in extraordinary conditions. HTH, Alinator |
Send message Joined: 17 Aug 09 Posts: 19 |
The CC will post a message in the logs about tasks which are over the deadline, advising that you may not get credit for it and might want to abort them. However you have to do that manually by design How does it notify (a pop up box or a log print), and why is a user input required to cancel/abort jobs over the deadline? Surely if the deadline has passed, the job should be automatically cancelled as it cannot be used/processed by the server? The redundant task abort makes perfect sense, although this would not address tasks currently being processed that are over the deadline. Is there some scope for developing this feature here? |
Send message Joined: 8 Jan 06 Posts: 448 |
CPDN has a soft deadline because of its extemely long WU. It will accept results past the deadlines. As long as the client continues to send trickle on a regular basis. The manager will just continue to run the past due WU in priority mode to ensure it is finished as quickly as possible. |
Send message Joined: 17 Aug 09 Posts: 19 |
This strikes me as an attempt to achieve some balance in retaining efficiency; whether by retaining users by allowing them to receive credit for uploading complete but useless results at a cost of time lost in processing redundant tasks, vs cancelling overdue/redundant tasks to reduce said lost time but with a potentially detrimental effect on the size of the host pool. In which case, 6.10 favours the later. Thank you all for the comments, very helpful! |
Send message Joined: 19 Jan 07 Posts: 1179 |
Think 6.10 has the feature to client side automate the cancellations of unstarted / overdue tasks, message log recorded, without need to contact server. That's in 6.6, or possibly 6.4. |
Send message Joined: 1 Aug 09 Posts: 33 |
[...]why is a user input required to cancel/abort jobs over the deadline? Surely if the deadline has passed, the job should be automatically cancelled as it cannot be used/processed by the server? I totally agree with you. Why is a human intervation necessary to cancel/abort a task which is over the deadline? If I remember well, the task will be automatically redistributed to another computer, and it shows that the computer which has a task with an over deadline is completely unuseful for the processing of this task, isn't it? |
Send message Joined: 29 Aug 05 Posts: 15581 |
I totally agree with you. Why is a human intervation necessary to cancel/abort a task which is over the deadline? As already said earlier, the ClimatePredction.Net project will take models that ran over their deadline. In other projects as well, when you run longer tasks, you still get credit when you manage to upload & report the task before the computer with the resent task does so. I've gotten credit on Einstein tasks that were 5 days over their deadline. This was at a time I didn't have an internet connection, but continued running down my cache anyway. So it's not all so clear cut that the client should automatically abort/delete tasks over the deadline. What if you had 1%, or 3 minutes to go on a task but it was deleted due to it being over the deadline? Would you like that? Then people would demand a certain time point after the deadline before this came into effect, then demand that that time point would be increased as they'd always be at that 1% or 3 minutes before their task was deleted. Thus, it's better to leave it to human intervention. You decide what you do with the task in question. That's why BOINC says "Consider aborting the task". It doesn't tell you that you must. |
Send message Joined: 1 Aug 09 Posts: 33 |
I totally agree with you. Why is a human intervation necessary to cancel/abort a task which is over the deadline? I did not know it. Your help was precious. Thanks, I learnt something. See http://www.student.montefiore.ulg.ac.be/~merciadri/ for more pieces of info. |
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.