Message boards : Number crunching : Not getting any work.
Author | Message |
---|---|
sculptor Send message Joined: 25 Mar 09 Posts: 6 Credit: 2,604,245 RAC: 0 |
Ive just installed boinc 10 43. and it only requests work for the GPU from Rosetta. after leaving it for 10 mins it still hasent got any work for Rosetta. It recieved work for Enstien@Home straight away so must be configured correctly. Any ideas guys. Tom. |
Jochen Send message Joined: 6 Jun 06 Posts: 133 Credit: 3,847,433 RAC: 0 |
Just be patient. It will start asking for CPU-WUs shortly. Winhin in the first 30 minutes after installation. Jochen |
sculptor Send message Joined: 25 Mar 09 Posts: 6 Credit: 2,604,245 RAC: 0 |
Still havent recieved any work. Ill leave it untill tomorrow, finish the remaining Einstien work then try a reinstall of the boinc client. I started another PC about an hour after having problems with this one and it started downloading Rosetta work after a couple of minutes. I can only think that it is a problem with boinc. Tom, |
Jochen Send message Joined: 6 Jun 06 Posts: 133 Credit: 3,847,433 RAC: 0 |
What are the messages in the log, when the client is trying to download new work? Does it at least ask for CPU-WUs by now? |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2126 Credit: 41,252,656 RAC: 7,882 |
Still haven't received any work. It almost certainly is a problem with BOINC, from my experience, but a reinstall isn't going to make it work any better unfortunately. I often get calls for GPU WUs only from Rosetta, even though they don't exist and I'm needing CPU WUs. If you're also running other projects and have tasks remaining to complete this will complicate matters further, but I've learnt that with a little patience it sorts itself out of its own accord. As those tasks work through there'll be more space for Rosetta in your forward schedule. Give it a day or two. There are stacks of WUs at Rosetta (a few million) just as soon as your BOINC manager calls for them. |
sculptor Send message Joined: 25 Mar 09 Posts: 6 Credit: 2,604,245 RAC: 0 |
You are right Sid Ive had this problem before as well, only solution then was to go back to using the older boinc client, unfortunatly the latest Einstien work requires this client so Im stuck with it. Ill give it a couple of days as you suggest, if it dosent sort itself out Ill just have to swap around which PCs are running each project. its a shame though as Rosetta and Einstien work well together on a quad core. Ill just have to suck-it-and-see. Tom. |
mikey Send message Joined: 5 Jan 06 Posts: 1895 Credit: 9,183,586 RAC: 3,322 |
You are right Sid Ive had this problem before as well, only solution then was to go back to using the older boinc client, unfortunatly the latest Einstien work requires this client so Im stuck with it. You can also check your Long Term Debts and your Duration Correction Factors if you'd like. They are both inside a file called client-state.xml and could be in a hidden folder on your pc. The lines in the file will look like this: <duration_correction_factor>100.000000</duration_correction_factor> <short_term_debt>0.000000</short_term_debt> <long_term_debt>0.000000</long_term_debt> Your numbers WILL be different! Check and see what your numbers are and maybe we can help you get them back towards normal so you can get Rosetta work again. Also do NOT make any changes unless you KNOW what you are changing, things can get pretty messed up if you are not careful. I only run one project on this machine so that affects the numbers greatly, how many projects do you run on the machine in question? |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2126 Credit: 41,252,656 RAC: 7,882 |
You are right Sid. I've had this problem before as well, only solution then was to go back to using the older boinc client, unfortunately the latest Einstein work requires this client so I'm stuck with it. I'm not suggesting you roll back to an older BOINC version. I understand they're worse rather than better. As long as you're not out of work it should balance out in the end (so I'm led to believe anyway). One of my many faults is messing round with settings. As long as you won't run out of work you might want to set Einstein to "No new tasks" for a period and see if that breaks the logjam. |
sculptor Send message Joined: 25 Mar 09 Posts: 6 Credit: 2,604,245 RAC: 0 |
Working now. Thanks for your help guys, started recieving work at 15:49GMT today. Im running both Rosetta and Einstien on this PC, I have done for over a year with very little trouble. Any problems I have had have been down to the boinc manager, I know my way around ajusting the settings, you do when you you run mutiple projects, however v.10 43 dosent seem to respond as well as previous versions, to me atleast. During last week I was having problems getting work for Einstien on the older boinc manager, on their forum they suggest having the latest boinc manager as it solves these problems. Thats the reason I changed to 10 43. hopefully it will go ok from now on. Thanks again, Tom. |
Aroundomaha Send message Joined: 11 Sep 08 Posts: 14 Credit: 55,623,619 RAC: 0 |
I've had problems getting and upload work from all my computers today. Doh... scratch that looks like the floodgates just opened as I was typing this. Wow that THAT'S service :) |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2126 Credit: 41,252,656 RAC: 7,882 |
I have to agree, the recent versions of BOINC are very 'lumpy' in grabbing WUs. Constant feast then famine. On my two machines, one runs two projects, the other just one. Both suffer from this 'lumpiness'. On the plus side, when one core has completed a WU and my 2.5 day buffer is empty, a new one comes down straight away. But only one. A day or two later I build back up to 20 or so WUs, but just one or two arrive every 4 minutes. Or, like today, 12 all at once. Unsurprisingly, I've now given up trying to second guess what will happen. I just hope it resolves itself before the project goes down at some inopportune time... :( |
mikey Send message Joined: 5 Jan 06 Posts: 1895 Credit: 9,183,586 RAC: 3,322 |
I've had problems getting and upload work from all my computers today. Some of your Windows machines are using the 6.10.43 version of Boinc but some are still using 6.10.18, this may make a difference in the way units are received but then again if it works..... One thing to know about the 6.10.30 and newer versions is there is now a setting to not crunch if more than x% of your cpu is otherwise busy, Boinv will STOP crunching if this is reached. Most people are setting that to zero so Boinc keeps crunching. I do not see this setting on the webpage but it is in the Boinc Manager under Advanced, Preferences, and then the processor usage tab. It says 'while processor usage is less than [] percent (0 means no restriction)'. If you make the change you have to click ok, but it then overwrites the webpage settings, so be careful!! |
FoldingSolutions Send message Joined: 2 Apr 06 Posts: 129 Credit: 3,506,690 RAC: 0 |
I am having exactly the same problem as the creator of this thread! It is requesting GPU tasks from a non-GPU project, but no CPU tasks. I've waited quite a while now for this to change but every time it says 'requesting new tasks for GPU'. This is quite a major flaw is there no way of fixing it? And is it affecting the throughput of Rosetta as a result? EDIT: I just got some! Still annoying though! it seems like BOINC knows i've had a moan on here so it got it's act together :D |
mikey Send message Joined: 5 Jan 06 Posts: 1895 Credit: 9,183,586 RAC: 3,322 |
I am having exactly the same problem as the creator of this thread! It is requesting GPU tasks from a non-GPU project, but no CPU tasks. I've waited quite a while now for this to change but every time it says 'requesting new tasks for GPU'. This is quite a major flaw is there no way of fixing it? And is it affecting the throughput of Rosetta as a result? One problem Boinc gets itself into is LTD, Long term Debt, issues. If you CAREFULLY go into the file client_state.xml, it is in a hidden folder, you will see the line: "<long_term_debt>0.000000</long_term_debt>". The fact that mine is zero is because I only have one cpu project running on this pc right now, it does not OWE time to another project. If you check your file you will probably find that yours is not zero, and if you stop Boinc and then bring the number down to the single digits, not zero but say 1.121212, then save the file and restart Boinc, then you will get a more balanced workload from each project. The problem is that Boinc is trying to comply with your percentage settings for each Boinc project and at the same time dealing with down times etc at each project. This is a hack to get Boinc back to more of what you think it should be doing NOW, if you just leave it alone it WILL do it on its own!!! It will just take some time, as you noticed it is trying because you got some work already. |
Paul D. Buck Send message Joined: 17 Sep 05 Posts: 815 Credit: 1,812,737 RAC: 0 |
Sadly this is another of those areas where to this point the complaints of the users posting to the Alpha list have been unheeded by UCB... we have noted this "lumpy" behavior in a number of different ways including the one where if you have just reset your debts you will see BOINC go and get a lot of work from just about all the projects to the point where you do have a good buffer fill ... then, a week or so later you will note that the buffer is starting to be a little "thin" in that it does not really have a full bag of work ... I usually keep the flag to reset the debts in my CC CONFIG file and set it to 1 and stop and restart BOINC every couple weeks ... |
mikey Send message Joined: 5 Jan 06 Posts: 1895 Credit: 9,183,586 RAC: 3,322 |
Sadly this is another of those areas where to this point the complaints of the users posting to the Alpha list have been unheeded by UCB... we have noted this "lumpy" behavior in a number of different ways including the one where if you have just reset your debts you will see BOINC go and get a lot of work from just about all the projects to the point where you do have a good buffer fill ... then, a week or so later you will note that the buffer is starting to be a little "thin" in that it does not really have a full bag of work ... I do it by running only one project of a type per pc, ie only one cpu project and one gpu project per pc. |
Message boards :
Number crunching :
Not getting any work.
©2024 University of Washington
https://www.bakerlab.org