Restarting project?

Questions and Answers : Getting started : Restarting project?

To post messages, you must log in.

AuthorMessage
Kashii

Send message
Joined: 9 Apr 06
Posts: 2
Credit: 145
RAC: 0
Message 13995 - Posted: 18 Apr 2006, 1:08:36 UTC

I have successfully completed 3 projects, each withing a few hours. But my fourth one has been taking days! How long can these things take? I got it up to 10 hours CPU, and it still wasn't done. I had to turn off the computer, and when I restarted it later, it was back at 0 CPU. Did it completely restart that project? Why am I having trouble with this one, when I didn't have any with the first 3? I hadn't hit the "suspend" button on the first 3; are you supposed to any time you have to shut down your computer?
ID: 13995 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Feet1st
Avatar

Send message
Joined: 30 Dec 05
Posts: 1755
Credit: 4,690,520
RAC: 0
Message 14139 - Posted: 19 Apr 2006, 21:26:38 UTC

Kashii, welcome to Rosetta! You have completed three "Work Units" (called WUs for short), Rosetta is the "project". Yes, you've struck on one of the current hot issues with the project. There are a couple of possible issues here that have cropped up recently. Let me take them one at a time:

If the work unit has one of the four names [/url=https://boinc.bakerlab.org/rosetta/forum_thread.php?id=1383#13576]mentioned here[/url], you should probably just "abort" it. Those had some problems with them, and on some systems were not completing normally. I see one of your WUs is named like those in the list. That's probably why you are having this problem. Just go to the Work tab in BOINC Manager, select that WU, and click the ABORT button.

The graphics display will show the "model" and "step" that is in progress. Recently they sent out WUs which took siginificantly longer than normal to complete model 1. If you shutdown your computer before completing a model, then Rosetta has to begin that model again when you start up. It's nothing that you did wrong. That's the other issue being discussed presently is that if the project is going to have such long processing to complete model 1, then they need to save their work (i.e. "checkpoint") more frequently.

There are a number of threads discussing these issues. Most of them are in the "Crunching" board because people could crunch more work, if they didn't have to start over at the beginning every time.

One thing you might do to help would be to select the "leave in memory" option in your "General Preferences". This way, Rosetta can continue where it left off on model 1 when BOINC suspends it to crunch on another project for a while. But if you have to turn off your computer, then it will have to start the model that it is on from the beginning.

Rosetta must complete at least one model before it will mark the WU as completed. For some WUs this first model can be completed in 15 mintues, for others it takes many hours. This is because the specific proteins the WUs are studying vary in size. So, as you crunch other WUs you may notice this. Ideally, to lose the smallest amount of work possible, you would end BOINC and shutdown your computer just after Rosetta completes a model and resets the step.
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/
ID: 14139 · Rating: 1 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Feet1st
Avatar

Send message
Joined: 30 Dec 05
Posts: 1755
Credit: 4,690,520
RAC: 0
Message 14194 - Posted: 20 Apr 2006, 20:16:39 UTC

It looks like there are changes coming very soon to do the checkpoints more frequently. They are testing these changes presently over in the Ralph project, which means they should be here soon!

This means that even the processing of a partial model can be preserved if you power down your PC, or if BOINC switches to another project for a while. This will mean more productive crunch time for everyone. They are also testing some solutions to a few bugs they've been working hard to kill for some time. So this is all great news.

You will notice the application name will change from Rosetta 4.98 to a 5.xx number when these changes are proven on Ralph and rolled out to Rosetta.
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/
ID: 14194 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Questions and Answers : Getting started : Restarting project?



©2024 University of Washington
https://www.bakerlab.org