Message boards :
Number crunching :
Panic Mode On (85) Server Problems?
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 . . . 23 · Next
Author | Message |
---|---|
Filipe Send message Joined: 12 Aug 00 Posts: 218 Credit: 21,281,677 RAC: 20 |
Where are you at with the new larger MB tasks? I remember to read about a 4x bundle. MB tasks 4x times larger. So we can reduce the load on the database. |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
Where are you at with the new larger MB tasks? I remember to read about a 4x bundle. We were testing the large size wu's on beta, but then I think all focus shifted to v7 apps. Which gives longer run times by doing more science. At the moment the focus looks to be the release of a stock android app. Which was just released, on beta, this week. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Mal McKellar Send message Joined: 6 May 03 Posts: 3 Credit: 1,983,416 RAC: 5 |
Hi Guys, A quick question please: How do I set SETI to highest priority? I also run Einstein, LHC, Cosmology & Milky Way. Thanks for your help. Mal |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Is not allways a good ideia to try to run too many projects on one slow machine, concentrate the resources on one or two project is more eficient, but if you want to try... You could change the resource share form 100 (the defoult) to a large numer in SETI (500 will give SETI 5x more time than the others, etc.), that will give your SETI work more share of the processing time. |
rob smith Send message Joined: 7 Mar 03 Posts: 22273 Credit: 416,307,556 RAC: 380 |
Set SETI to 1000, the others to a number below 100. And WAIT, it will take BOINC some time to sort its life out and balance the work correctly, and remember it is a long term average share, not an "instant" share. Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Mal McKellar Send message Joined: 6 May 03 Posts: 3 Credit: 1,983,416 RAC: 5 |
Thanks Juan & Rob, I will try that. I agree with your comments about too many tasks - I actually only run one of the others at any one time with Seti always on. Many thanks, Mal |
Cosmic_Ocean Send message Joined: 23 Dec 00 Posts: 3027 Credit: 13,516,867 RAC: 13 |
The MB splitting back-log is making pretty good progress. I would guess it shouldn't be but 2-4 days before we get some new tapes to work with for AP. My cache is still hanging on. I've been getting 1-4 re-sends/day. Currently have just under 3 days until idle, unless I get more before then. Still not complaining. Just mentioning that relief should be coming soon. Linux laptop: record uptime: 1511d 20h 19m (ended due to the power brick giving-up) |
Thomas Send message Joined: 9 Dec 11 Posts: 1499 Credit: 1,345,576 RAC: 0 |
1 AP (GPU) in my list of WU's ! That's already something... :) |
Lionel Send message Joined: 25 Mar 00 Posts: 680 Credit: 563,640,304 RAC: 597 |
Where are you at with the new larger MB tasks? I remember to read about a 4x bundle. and less credit ...lets not forget this other fact shall we.. :) |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Not a real panic, but anyone have any ideia way for about more than a week no AP WU are spliting? I see in the servers stst page 11 tapes are allready AP done but there are a lot more space than only 11 tapes on the servers. Did i miss something? |
rob smith Send message Joined: 7 Mar 03 Posts: 22273 Credit: 416,307,556 RAC: 380 |
APs will become available again when a new batch of tapes are loaded. APs are split much faster than MB, so get shot out more quickly and each tape contains fewer APs than MBs because APs are bigger, and have a much smaller overlap than MBs (and this has been explained on quite a few occasions.) Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
William Send message Joined: 14 Feb 13 Posts: 2037 Credit: 17,689,662 RAC: 0 |
I believe the question was why there are only 11 tapes mounted. You might be able to mount more, but that would create an even larger backlog of MB still to split. Maybe the solution is to split less AP at a time, so the tapes last longer - spread the work a little more. But in the long run, you still have to split all the MB work. Now I know - AP crunching on GPU is far too efficient. therefore the AP WUs don't last long enough. Maybe if we throttled the speed of the AP GPU apps? ;) A person who won't read has no advantage over one who can't read. (Mark Twain) |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
I believe the question was why there are only 11 tapes mounted. Yes, why not mount few (not necessary all the possible at the same time) more tapes to split AP WU? Maybe the solution is to split less AP at a time, so the tapes last longer - spread the work a little more... Maybe it helps, spread the work sure helps, it´s far more eficient crunch 1AP+1MB on the GPU than 2AP or 2 MB on my host´s at least, by more eficient i means more work done (more WU crunched) in less time. As i say before several times, a balance must be achieved, specialy after the mess of creditnew with the credit given to the MB WU a lot of top hosts are hungry for new AP to crunch. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
I believe the question was why there are only 11 tapes mounted. Here's an idea. Since the only GPU tasks presently working on a Mac is the ATI AstroPulse task, just send all GPU AstroPulse tasks to Macs. That would work for me, and help keep APs from disappearing for weeks at a time... nods head... |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Yes the start to split one tape for AP. Anyone know how to automaticaly make the AP WU start to crunch as soon they DL and not need to wait to clear the entire MB cache? But by leaving the MB on the caches ready to start as soon no AP WU were avaiable? |
Cosmic_Ocean Send message Joined: 23 Dec 00 Posts: 3027 Credit: 13,516,867 RAC: 13 |
Oh sweet beautiful new APs! I still had ~2.5 days left of my re-send cache, and got topped-off to a full 10-day cache.. 2013-08-05 17:38:25 SETI@home Scheduler request completed: got 4 new tasks 2013-08-05 17:43:32 SETI@home Scheduler request completed: got 1 new tasks 2013-08-05 17:48:40 SETI@home Scheduler request completed: got 2 new tasks 2013-08-05 17:53:46 SETI@home Scheduler request completed: got 1 new tasks 2013-08-05 18:04:00 SETI@home Scheduler request completed: got 2 new tasks 2013-08-05 18:09:07 SETI@home Scheduler request completed: got 2 new tasks 2013-08-05 18:24:30 SETI@home Scheduler request completed: got 1 new tasks 2013-08-05 18:29:37 SETI@home Scheduler request completed: got 2 new tasks 2013-08-05 18:34:43 SETI@home Scheduler request completed: got 2 new tasks 2013-08-05 18:39:50 SETI@home Scheduler request completed: got 2 new tasks 2013-08-05 18:44:57 SETI@home Scheduler request completed: got 5 new tasks 2013-08-05 18:50:04 SETI@home Scheduler request completed: got 1 new tasks 2013-08-05 18:55:11 SETI@home Scheduler request completed: got 3 new tasks 2013-08-05 19:00:19 SETI@home Scheduler request completed: got 3 new tasks 2013-08-05 19:05:26 SETI@home Scheduler request completed: got 6 new tasks 2013-08-05 19:10:33 SETI@home Scheduler request completed: got 1 new tasks 2013-08-05 19:20:45 SETI@home Scheduler request completed: got 1 new tasks 2013-08-05 19:46:22 SETI@home Scheduler request completed: got 5 new tasks 2013-08-05 19:51:29 SETI@home Scheduler request completed: got 6 new tasks Linux laptop: record uptime: 1511d 20h 19m (ended due to the power brick giving-up) |
Lionel Send message Joined: 25 Mar 00 Posts: 680 Credit: 563,640,304 RAC: 597 |
I believe the question was why there are only 11 tapes mounted. Sten, you evil person :) |
Lionel Send message Joined: 25 Mar 00 Posts: 680 Credit: 563,640,304 RAC: 597 |
Yes the start to split one tape for AP. Juan, maybe someone with more knowledge can give you the answer your after however, in my case I just wait to see when AP is up, then I switch off v7 in preferences and abort all v7 WUs across all machines. From that point it takes a few cycles for all machines to fill up on AP. I know some people won't be happy about that but the way I see it, there is no point in crunching a WU (v7) that gives you 65% of the run rate of a different type of WU (AP). I move all machines back to v7 only when there are no AP WUs available. In short I suppose you can say that I now use v7 WUs as the filler between APs. |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Lionel That´s exactly why i ask, i want to do the same but without the need to abort the V7 already on the cache. I start a new thread about that and receive some sugestions, until now no one apears with a working solution, but i belive if i could do what i want a lot of others will do the same. I agree for now crunch MB "paid" a lot less credit, but i don´t want to simply stop crunching MB, just create a balance with both WU. |
Lionel Send message Joined: 25 Mar 00 Posts: 680 Credit: 563,640,304 RAC: 597 |
Juan, understand exactly where you are coming from. The sad thing is if the credit for v7 WUs was properly benched marked to AP, there would be no need for what you are after or for any of us to keep switching. Sad days I feel. cheers |
©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.