Message boards : Number crunching : Why am I not getting Rosetta work units ??
Author | Message |
---|---|
Alan Payne Send message Joined: 13 Nov 05 Posts: 12 Credit: 21,584 RAC: 0 |
There seems to be something not quite right with the way work units are given within BOINC. I am supporting Seti, Climate Prediction and Rosetta. Several weeks ago I gave Rosetta HALF of the usage and 25% to the other two because I was keen to compete in the CASP7 competition. But Seti and Climate Prediction seems to be hogging all of my user time. In late May I wanted to compete in the CASP7 competition. I even upgraded my RAM from 512 to 1.5 gigs. Can anyone tell me why I am not getting sufficient work from Rosetta ?? regards Alan Payne |
Astro Send message Joined: 2 Oct 05 Posts: 987 Credit: 500,253 RAC: 0 |
Well, you're the only one in a position to answer that, but I can help you figure it out. First read the Work Fetch Policy in the Boinc Wiki. If you want to find out your LTD (long term debt), you can either manually look through the "clientstate.xml" file (located in C:programfilesboinc) or you can install Skips small script and tell it to extract the data from the clientstate.xml file. The program is boinc debt viewer (BoincDV). I think you'll find time is owed to those two projects and it is being caught up. I'm SURE you'll have more questions, but you can give it a read, and someone will be around. tony |
Feet1st Send message Joined: 30 Dec 05 Posts: 1755 Credit: 4,690,520 RAC: 0 |
It would be helpful if you could let us know how long this has been going on? And what your General Preference for time to switch between tasks is? Here's a brief QA item on the subject as well with related links to learn more. And another less directly related QA item. Add this signature to your EMail: Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might! https://boinc.bakerlab.org/rosetta/ |
tralala Send message Joined: 8 Apr 06 Posts: 376 Credit: 581,806 RAC: 0 |
There seems to be something not quite right with the way work units are given within BOINC. I would suspend climateprediction and set Seti to no new work till August 4th, when CASP ends. After that you can catch up with climateprediction and Seti. You can delete all your long-term and short-term-credit with the boinc debt viewer (BoincDV) mentioned by mmciastro. |
Astro Send message Joined: 2 Oct 05 Posts: 987 Credit: 500,253 RAC: 0 |
You can delete all your long-term and short-term-credit with the boinc debt viewer (BoincDV) mentioned by mmciastro. BoincDV can reset all your Long Term Debt Values, so you're essentially starting from scratch as it pertains to resource share. It doesn't reset any credit. Suspending a Project stops LTD changes for a Project. Setting NNT/NNW (no new tasks/work) stops LTD accumulation as long as the project has NO work, but if it does, then LTD still accumulates. tony |
Alan Payne Send message Joined: 13 Nov 05 Posts: 12 Credit: 21,584 RAC: 0 |
It would be helpful if you could let us know how long this has been going on? And what your General Preference for time to switch between tasks is? Hi I have set my preferences to switch between applications every 60 minutes |
Astro Send message Joined: 2 Oct 05 Posts: 987 Credit: 500,253 RAC: 0 |
Hi That preference setting means the "work scheduler" will examine all available work and deadlines, and make a decision every 60 minutes as to what should be run next. if your ltd is very negative for one project it will continue to run that project. tony |
Astro Send message Joined: 2 Oct 05 Posts: 987 Credit: 500,253 RAC: 0 |
Perhaps a quick read of the work scheduler in the wiki might be in order. The wiki can answer pretty much any question. tony |
Alan Payne Send message Joined: 13 Nov 05 Posts: 12 Credit: 21,584 RAC: 0 |
Well, you're the only one in a position to answer that, but I can help you figure it out. First read the Work Fetch Policy in the Boinc Wiki. If you want to find out your LTD (long term debt), you can either manually look through the "clientstate.xml" file (located in C:programfilesboinc) or you can install Skips small script and tell it to extract the data from the clientstate.xml file. The program is boinc debt viewer (BoincDV). OK Tony I downloaded the boinc dept viewer, unzipped it and ran it and this is what I got:- PRJ: CLIMATEPRE STD: 37561.218750 LTD: 643346.921559 RSRC: 75 ------------------------------- PRJ: ROSETTA@HO STD: 0.000000 LTD: -14923.125000 RSRC: 150 ------------------------------- PRJ: SETI@HOME STD: -37561.218750 LTD: -628423.796559 RSRC: 75 I CLOSED my BOINC manager and clicked on "clear debts" and it came up with exit BOINC NOW. I'm too lazy to add code to check if it's running. Any suggestions ?? regards Alan |
Alan Payne Send message Joined: 13 Nov 05 Posts: 12 Credit: 21,584 RAC: 0 |
Well, you're the only one in a position to answer that, but I can help you figure it out. First read the Work Fetch Policy in the Boinc Wiki. If you want to find out your LTD (long term debt), you can either manually look through the "clientstate.xml" file (located in C:programfilesboinc) or you can install Skips small script and tell it to extract the data from the clientstate.xml file. The program is boinc debt viewer (BoincDV). |
Astro Send message Joined: 2 Oct 05 Posts: 987 Credit: 500,253 RAC: 0 |
OK Tony See the Negative LTD for Seti? Seti needs to run 628,423 seconds (7.27 days) to get back to zero. It appears you might have manually stopped seti from running. You don't need to code anything to clear debts with boincdv. The "exit boinc now" prompt is just there to make sure you shut down boinc before it edits your clientstate.xml file. If you don't shut down boinc when editting that file Boinc/projects will crash badly. Just make sure you've shut it down and click on the box. Note: clearing debts resets the debts and all projects will begin following your resource share again from this point. I.E Seti will not get those 7.27 days of work it's owed to met the resource share you've requested. As long as you know. tony |
Alan Payne Send message Joined: 13 Nov 05 Posts: 12 Credit: 21,584 RAC: 0 |
OK Tony |
Alan Payne Send message Joined: 13 Nov 05 Posts: 12 Credit: 21,584 RAC: 0 |
OK Tony Hi again Tony, Well Rosetta still wont give me any work because they now say I do not have enough resources available. Here are the latest messages received from them. 14/07/2006 9:12:11 p.m.|rosetta@home|Message from server: No work sent 14/07/2006 9:12:11 p.m.|rosetta@home|Message from server: (there was work but you don't have enough disk space allocated) 14/07/2006 9:12:11 p.m.|rosetta@home|Message from server: No disk space (YOU must free 276.2 MB before BOINC gets space). Review preferences for minimum disk free space allowed. 14/07/2006 9:12:11 p.m.|rosetta@home|No work from project I have 4 gigs of disk space allocated for this project. Crazy, Crazy, Crazy !! Any ideas of what is going on Tony ?? regards Alan |
Astro Send message Joined: 2 Oct 05 Posts: 987 Credit: 500,253 RAC: 0 |
Any ideas of what is going on Tony ?? go back to the main page, click on "your account" then, General Preferences "view or edit", then copy and paste the "Disk and memory usage" section to this thread. Here's what mine looks like: (what size is your hard drive) Disk and memory usage Use no more than 400 GB disk space Leave at least (Values smaller than 0.001 are ignored) .25 GB disk space free Use no more than 85% of total disk space Write to disk at most every 600 seconds Use no more than 100% of total virtual memory |
Alan Payne Send message Joined: 13 Nov 05 Posts: 12 Credit: 21,584 RAC: 0 |
Any ideas of what is going on Tony ?? Hi Tony Thanks once again for very useful info. I looked again at all my preferences in all three of my projects and found that I had screwed up in one of them. I have to admit, until I read the Rosetta FAQ's (which was prompted by the info you had given me) I hadn't realised that the first "Use no more than" was for all three of my projects. And the second "Use no more than" was the maximum amount that could be used by the individual project. I have never seen this explained in SETI (where I started some 7 years ago). Anyway all is well now and thanks again Tony for "nursing" me through all of these technical points. It is all coming in handy for me to understand how the BOINC works. For example SETI has sent out some more work units with a completion date just over a week hence (which HAVE to be completed before any other project can get a look in). I had a look at SETI's LTD and it's now clocking up 'negative' seconds because of this. So now I understand that no project really gets away with getting more work done than they are entitled to. I have had a lot of work units from SETI with this early completion date which tells me that they dont't really have enough volunteers number crunching for them ?? Gotta go, I've waffled on enough. Once again thanks a bundle for all the time you have given to bring me up to date on understanding the crucial points of BOINC. By the way SETI, quite a while ago, was hoping to get some telescope time in the Southern Hemisphere (where I live) to look for ET. But this was delayed because they didn't have the equipment to do it. Any buzz about this in your circle ? All the best Alan Payne |
Astro Send message Joined: 2 Oct 05 Posts: 987 Credit: 500,253 RAC: 0 |
The "work scheduler" determines which mode to run in. the two different ones are RR (round robin), and EDF (earliest deadline first) (note: in the newest Alpha versions of boinc this mode is going away, to be replaced by another definition). If a project like seti gets close to being possible to miss a deadline, it will switch from RR mode to EDF mode and crunch those WU in danger of missing deadlines. Note: the larger your "connect to setting" the more often it will run in EDF. If it runs one project in EDF, it will still continue to count the cpu seconds used and will repay the other projects over the long term. The work scheduler won't necessarily appear to follow the resource share in the short term, but will long term. Without watching all the data on a frequent basis, it's nearly impossible to determine if it's working, and a user needs to pretty much have "faith" that it is working. As for "southern hemisphere" data collected at Parkes, I've heard nothing other than it's being investigated. As you know Seti has been undertaking alot of new stuff this year (Seti Enhanced, new data recorder at arecibo, getting data from Alfa-multibeam receiver, and starting Astropulse). tony |
Message boards :
Number crunching :
Why am I not getting Rosetta work units ??
©2024 University of Washington
https://www.bakerlab.org