David E K Forum moderator Project administrator Project developer Project scientist Joined: Jul 1 05 Posts: 660 ID: 14 Credit: 838,217 RAC: 28
Please post bugs and issues regarding minirosetta version 1.45.
This update includes fixes to long runtimes for 'relax' jobs, validation errors, check point recovery issues, and numerical instability in hydrogen-bond scoring.
We think we might have fixed the preemption problem so please keep an eye out for this. The "can't acquire lockfile" issue might also be related. If you are having lockfile problems, please make sure there are no other boinc applications running in the same slot. If necessary, turn off the client and make sure all boinc apps are not running, and then restart the client.
Please, please, please post new versions in Rosetta Application Version Release Log. That's how we get notified by email so that we can adjust our firewalls--it's very important.
ID: 57625 | Rating: 0 | rate:
/
David E K Forum moderator Project administrator Project developer Project scientist Joined: Jul 1 05 Posts: 660 ID: 14 Credit: 838,217 RAC: 28
Please, please, please post new versions in Rosetta Application Version Release Log. That's how we get notified by email so that we can adjust our firewalls--it's very important.
This problem is getting very frustratung. I run Eosetta as well as nine other BOINC tasks so I try to balance approximate compute times across all but the Climate Prediction task which taks 2000 CPU minutes per task but allows two years to complete each task.
So when a Rosetta task runs away for 18-22 hours of CPU time I end up aborting it since it says a little more than 9 minutes left out of about 6 hours it was estimated to run.
All other BOINC based projects are fairly accurate and none have come close to 3-4 times the initial estimate as these are. Right now I have two Rosetta tasks running and both have hung at a little more than nine min for one and ten for the other.
So until this problem is resolved I have no choice but to suspend all further Rosetta tasks. I feel bad having to abort those 5-6 tasks previously since that is over 100 hours of CPU time wasted with two still suspended. That comes to over 160 hours of CPU time wasted when that amount of time could have completed dozens of tasks for other projects. Event SETI hasn't had a task that took more the 40 hours while the LHCAT tasks only take about two hours each so I could have processed 80 of them with the wasted CPU time from Rosette 1.40 tasks.
So I do hope they can fix what is wrong so I can get back to processing Rosette tasks again.
____________
ID: 57639 | Rating: 0 | rate:
/
David E K Forum moderator Project administrator Project developer Project scientist Joined: Jul 1 05 Posts: 660 ID: 14 Credit: 838,217 RAC: 28
This problem is getting very frustratung. I run Eosetta as well as nine other BOINC tasks so I try to balance approximate compute times across all but the Climate Prediction task which taks 2000 CPU minutes per task but allows two years to complete each task.
So when a Rosetta task runs away for 18-22 hours of CPU time I end up aborting it since it says a little more than 9 minutes left out of about 6 hours it was estimated to run.
All other BOINC based projects are fairly accurate and none have come close to 3-4 times the initial estimate as these are. Right now I have two Rosetta tasks running and both have hung at a little more than nine min for one and ten for the other.
So until this problem is resolved I have no choice but to suspend all further Rosetta tasks. I feel bad having to abort those 5-6 tasks previously since that is over 100 hours of CPU time wasted with two still suspended. That comes to over 160 hours of CPU time wasted when that amount of time could have completed dozens of tasks for other projects. Event SETI hasn't had a task that took more the 40 hours while the LHCAT tasks only take about two hours each so I could have processed 80 of them with the wasted CPU time from Rosette 1.40 tasks.
So I do hope they can fix what is wrong so I can get back to processing Rosette tasks again.
do you still have the names of those problem tasks? can you try our recently updated version and see if you have the same problems?
ID: 57640 | Rating: 0 | rate:
/
Mod.Sense Forum moderator Project administrator Joined: Aug 22 06 Posts: 2397 ID: 106194 Credit: 0 RAC: 0
ChiTownDale have you seen problems like this with v1.45?? It includes changes that should eliminate the long running models, and unpredictable completion times.
____________ Rosetta Moderator: Mod.Sense
Task 212334548, workunit 193576676 failed on my iMac2 10.4.11. after about half an hour. It seems to have been completed successfully by someone on an XP system.
<core_client_version>6.2.18</core_client_version>
<
