Message boards : Number crunching : No Rosetta Wu's being downloaded
Author | Message |
---|---|
Jim Send message Joined: 1 Dec 05 Posts: 3 Credit: 26,219 RAC: 0 |
I'm participating in Rosetta and Seti. The Seti WU's are downloading fine, but none for Rosetta. This started after getting an initial error for Rosetta that no WU's were available. I'm not getting and error when updating Rosetta, just no WU's. Both projects have the same settings and nothing is suspended. Any ideas? Thanks Jim |
Feet1st Send message Joined: 30 Dec 05 Posts: 1755 Credit: 4,690,520 RAC: 0 |
How long as this been the case? If you go to the projects tab in BOINC Manager, what is shown for the "status"? If it says "no new work", select Rosetta, and then click the button that appears at the left that says to allow new work. Is it possible you were crunching a little more Rosetta work to meet deadlines and now are simply making up that debt to SETI? 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/ |
Cvi Send message Joined: 5 May 06 Posts: 2 Credit: 574,936 RAC: 1 |
I have the same problem. Even enlarging the resource share in preference, and resetting project didn't help I have not been at home for 4 days, it happened during these last days. Cvi How long as this been the case? |
FluffyChicken Send message Joined: 1 Nov 05 Posts: 1260 Credit: 369,635 RAC: 0 |
Try hitting the update button for each project in BoincManager and also (probably first) the retry communications in the advanced menu. That should recent any time delay roll-backs/retries and put the projects to where they should be. But given you say it happened when there was no work, BOINC will have got as much work as it could from Seti@home and now i will complete these first. But in this case even altering your resource share wouldn't (shouldn't) do much since boinc will probably see that it can only complete the seti tasks. So either just leave it and it will sort itself out OR delete some seti tasks to make room. Team mauisun.org |
Jim Send message Joined: 1 Dec 05 Posts: 3 Credit: 26,219 RAC: 0 |
Thanks for the replies everyone. It seemed to resolve itself last night. It was 8 days since I got any WU's from Rosetta. It was strange, no work for 8 days and only the first day the message stated there was no work from the project. I tried updating the Rosetta project and it would say that (no work was requested) / Request succeeded. Well all is well and I appreciate everyone input! Thanks Jim |
Cvi Send message Joined: 5 May 06 Posts: 2 Credit: 574,936 RAC: 1 |
Everything went back to normal after a day or so. Cvi I have the same problem. |
River~~ Send message Joined: 15 Dec 05 Posts: 761 Credit: 285,578 RAC: 0 |
Jim: When you did get work you got 28 WU all at once - am I right that you have your cache set at 3.5 days? (assumption: you have the default 3hr target runtime) The rule of thumb is that to keep work coming steadily from N projects where the shortest deadline of any project is D is to set a connect interval of less than D / (N+2) This will allow a connect interval for each project, plus one connect interval overall for crunching, plus one more waiting to return it. In short, the above rule of thumb allows enough breathing space for the client to "feel happy" to download work steadily. Your settings don't. So, assuming that SETI has a deadline of 7 days then that would be 7/4 days, = 1.75. Take off a sliver for luck and I'd use 1.74 or 1.744 at the most. This will make these glitches less likely, but will never rule them out totally - for example cvi reports this problem too and I think he has a cache of ~ 1 day. ~~~~ Talking of settings... There is a second tip I'd like to offer. Each WU has a network and a database overhead and this project offers user selectable run lengths. You can find this setting under your account -> project specific prefs. The suggested length is the one that will give you 1 WU per day calendar time on average. So if you do an even split with SETI, then ask for 12hr WU. Likewise adjust up/down for unequal project shares and adjust again if the box is turned off regularly. This setting will not alter your credit, nor the scientific value of your crunching. It will will help conserve the project's network bandwidth, help the database server keep up with the work, reduce download time, and if your ISP / phoneline meters your downloads it will save your own costs. It also keeps the size of the pages listng your results manageable! This is a very mild request - many feel they prefer a different target run length and they are welcome - the ability to choose is a positive feature of this project. HTH River~~ |
River~~ Send message Joined: 15 Dec 05 Posts: 761 Credit: 285,578 RAC: 0 |
... PS I should have said, if you are going to increase the run length and you currently have a connect interval of more than 0.2 days, then you need to do so one step at a time in the prefs. If you increase several steps at once, that can cause problems just like the one you first mentioned... For more info on this, please see runtime preference on work units. R~~ |
Keck_Komputers Send message Joined: 17 Sep 05 Posts: 211 Credit: 4,246,150 RAC: 0 |
This formula works well for a small number of projects (about 4 or less) but the denonimator does not grow fast enough, D / (N*2) would work better as the number of projects grows. The one I like to use is 40% of the shortest deadline divided by the number of projects, that allows a fudge factor if the estimated run time is off. BOINC WIKI BOINCing since 2002/12/8 |
Jim Send message Joined: 1 Dec 05 Posts: 3 Credit: 26,219 RAC: 0 |
Where in Boinc am i able to adjust the Target RunTime? I have the number of days set to 1.5 for Rosetta at this time. I greatly appreciate everyone's input! Thank you Jim |
AlphaLaser Send message Joined: 19 Aug 06 Posts: 52 Credit: 3,327,939 RAC: 0 |
|
Message boards :
Number crunching :
No Rosetta Wu's being downloaded
©2024 University of Washington
https://www.bakerlab.org