Message boards :
Number crunching :
14 Day Deadline to be introduced at Einstein
Message board moderation
Author | Message |
---|---|
Gary Roberts Send message Joined: 31 Oct 99 Posts: 95 Credit: 2,301,228 RAC: 0 |
With all the current difficulties with Seti, there have been quite a few suggestions about shifting resources to other projects to try to help the Seti Devs to get the various parts of their server systems operating efficiently. Einstein@Home is one of those projects but the 7 day deadline there has been a significant turnoff for people wanting a viable alternative. There is a current thread in Cafe Einstein where Prof Bruce Allen has recently announced that due to his faith in the EAH servers handling the load, a 14 day deadline will be introduced for their next set of work units from the S4 science run. Maybe EAH will now be seen as a more attractive alternative project for the embattled Seti users wondering about just how large their validation queues will get before things start to improve. I just thought people over here might be interested to know about this as I haven't seen it announced formally yet. Cheers, Gary. |
STE\/E Send message Joined: 29 Mar 03 Posts: 1137 Credit: 5,334,063 RAC: 0 |
IMO all the Projects should have the same Deadline, the way it is now you have some with a 1 Day, some with a 4 day, some with a 7 day & some with a 14 Day Deadline and it messes everything up if you have a favorite Project but still want to run a little of the other Projects. If one of your favorite Projects has a 14 day Deadline and you try to run Projects with a shorter Deadline the shorter Deadline Projects pretty much just take over and thats all you end up running. |
Pooh Bear 27 Send message Joined: 14 Jul 03 Posts: 3224 Credit: 4,603,826 RAC: 0 |
CPDN is impossible to have a 2 week deadline. It takes months to run. I do not think all projects needs to be the same length. When different sized units come, it will be too large for a 2 week deadline. I have no problem with the 1 week on Einstein. I think Protien Predictor could actually be a week, cause those are small enough. My movie https://vimeo.com/manage/videos/502242 |
Keck_Komputers Send message Joined: 4 Jul 99 Posts: 1575 Credit: 4,152,111 RAC: 1 |
IMO all the Projects should have the same Deadline, the way it is now you have some with a 1 Day, some with a 4 day, some with a 7 day & some with a 14 Day Deadline and it messes everything up if you have a favorite Project but still want to run a little of the other Projects. I have to disagree with you there. Allthough a set deadline per CS might be a viable idea. ie. if the average workunit at project A is worth 1 CS and the deadline is 1 day, then if the average workunit at project B is worth 10 CS the deadline should be 10 days. Still in general different projects will have different needs, and therefore different deadlines. The CPU scheduler in 4.45 and later handles this reasonably well and is improving. What will be interesting is if/when folding@home develops a real BOINC application. To fit thier needs we are liable to see workunits that take more time than E@H workunits and have shorter deadlines. They will almost certainly have the tightest deadlines of any project. BOINC WIKI BOINCing since 2002/12/8 |
Raithmir Send message Joined: 3 Apr 99 Posts: 89 Credit: 385,065 RAC: 0 |
With all the current difficulties with Seti, there have been quite a few suggestions about shifting resources to other projects to try to help the Seti Devs to get the various parts of their server systems operating efficiently. Einstein@Home is one of those projects but the 7 day deadline there has been a significant turnoff for people wanting a viable alternative. This is good news, I do run einstein but currently have it disabled on this machine for exactly that reason. Raithmir's SPARC64/UltraSPARC Linux Builds http://www.kulthea.net/boinc/ |
Bronco Send message Joined: 22 Jun 05 Posts: 123 Credit: 19,340 RAC: 0 |
I've test that point. The only problem I've got with BOINC scheduler (4.45) is that it seems to handle only the currently running projects, and doesn't take into account the full queues. Recently, due to a planned change in my ADSL access, involving some days off, I've change my settings to connect every 10 days. I've downloaded 6 E@H, no setis (was not working at that time), and CPDN was already OK. I've been obliged to handle manually the E@H queue as the scheduler didn't see that crunching all the E@H was only possible by affecting 100% of the ressources to each. But with short queues, 0.5 at the moment, it works perfectly. To get back in topic, even if the scheduler does it well, having longer time to crunch E@H will be a good thing. Having the same dead-line for all projects doesn't make sens, but having a single ratio computation time / Time to do the job for all projects will be better "In a world without walls and fences, who needs windows and gates ?" for the team |
Bronco Send message Joined: 22 Jun 05 Posts: 123 Credit: 19,340 RAC: 0 |
And it's working, I just get 2 WUs from E@H with 2 weeks to crunch them. Fine. "In a world without walls and fences, who needs windows and gates ?" for the team |
BarryAZ Send message Joined: 1 Apr 01 Posts: 2580 Credit: 16,982,517 RAC: 0 |
Indeed -- thanks for posting that -- I'm one of those who has shifted CPU cycles over to Einstein, but have had some problems with the 7 day cycle. For that matter, I've had problems in some setups getting the 4.45 client to work (I mention that because that client is said by a number of Einstein mavens to work better at shared resource handling for multiple projects). With all the current difficulties with Seti, there have been quite a few suggestions about shifting resources to other projects to try to help the Seti Devs to get the various parts of their server systems operating efficiently. Einstein@Home is one of those projects but the 7 day deadline there has been a significant turnoff for people wanting a viable alternative. |
BarryAZ Send message Joined: 1 Apr 01 Posts: 2580 Credit: 16,982,517 RAC: 0 |
|
©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.