What do these "options" mean exactly?
____________
ID: 63694 | Rating: 0 | rate:
/
Yifan Song Forum moderator Project administrator Project developer Project scientist Joined: May 26 09 Posts: 62 ID: 318024 Credit: 7,322 RAC: 0
As David described in his journal (http://boinc.bakerlab.org/rosetta/forum_thread.php?id=1177&nowrap=true#63383), we are working on improving energy functions. Currently, a lot of parameters are either defined in the code or in minirosetta_database. The new options allow us to test different energy functions from command line.
One example of what we plan to test in the coming weeks is a hydrogen bond potential with sharper distance and angular dependence. As many of you know, hydrogen bond plays an important role in drug design, as well as in protein/DNA interface design for gene therapy. However the exact form and magnitude of hydrogen bond is still underdetermined. Now from command line we can change the shape of hydrogen bond potential, and test whether they agree with the experimental data we've collected.
What do these "options" mean exactly?
ID: 63695 | Rating: 0 | rate:
/
Mod.Sense Forum moderator Project administrator Joined: Aug 22 06 Posts: 2976 ID: 106194 Credit: 0 RAC: 0
The "command line" is what drives each task to get it started on your machine at home. Just like Linux/Unix or DOS commands, BOINC applications are started with a number of options that direct them on how to run. So it is easily changed, without changing the program. And also easy to create several tasks with the same protein and starting point, and see if the command line adjustments make a demonstrable difference in the outcome.
____________ Rosetta Moderator: Mod.Sense
ID: 63700 | Rating: 0 | rate:
/
Michael G.R. Joined: Nov 11 05 Posts: 263 ID: 11128 Credit: 6,252,138 RAC: 293
I'm getting some errors with lr8_score12_run03_rlbd WUs. They exit after a few seconds with the message:
ERROR: Illegal attempt to score with non-identical atom set between pose and etable
ERROR:: Exit from: src/core/scoring/etable/EtableEnergy.cc line: 72
BOINC:: Error reading and gzipping output datafile: default.out
Fri 16 Oct 2009 15:22:49 EST|rosetta@home|Output file lr8_score12_run03_rlbd_1py9_IGNORE_THE_REST_DECOY_14712_976_1_0 for task absent
ERROR: Illegal attempt to score with non-identical atom set between pose and etable
ERROR:: Exit from: src/core/scoring/etable/EtableEnergy.cc line: 72
BOINC:: Error reading and gzipping output datafile: default.out
called boinc_finish
same as with AMD_is_logical:
http://boinc.bakerlab.org/rosetta/result.php?resultid=288316563:
ERROR: Illegal attempt to score with non-identical atom set between pose and etable
ERROR:: Exit from: src/core/scoring/etable/EtableEnergy.cc line: 72
Task 288375137 (symm_lr8_seq_score12_ss_1.7_rlbd_1h75_IGNORE_THE_REST_DECOY_14923_2689_0) failed on Mac OS X 10.6
ERROR: !core::conformation::symmetry::is_symmetric( pose )
ERROR:: Exit from: src/core/optimization/AtomTreeMinimizer.cc line: 55
BOINC:: Error reading and gzipping output datafile: default.out
called boinc_finish
</stderr_txt>
]]>
____________
ID: 63716 | Rating: 0 | rate:
/
Yifan Song Forum moderator Project administrator Project developer Project scientist Joined: May 26 09 Posts: 62 ID: 318024 Credit: 7,322 RAC: 0
Hmm, I can't figure out right away where these errors could be from. They are rather old work units which has been running ok. The same type of jobs also passed the alpha test on ralph for v1.98. I'll do some more tests over the weekend.
I seem to be getting bad WUs back to back. Only on my Pentium D 3.2GHz machine, though.
____________
ID: 63721 | Rating: 0 | rate:
/
Jerry Goggin Joined: Jun 7 06 Posts: 4 ID: 92225 Credit: 226,010 RAC: 0
This task appears to have hung
10/16/2009 10:02:03 PM rosetta@home task denovo_design_rossmann2x3_flxbb_SAVE_ALL_OUT_r2x3_001_rbd_h15_001_0001.82_0001_15195_60_0 suspended by user
so I am going to abort it. Task properties indicate the State "Waiting for memory" whatever that means. There is plenty of memory available. I can resume it, but it simply does not run.
10/16/2009 10:08:35 PM rosetta@home task denovo_design_rossmann2x3_flxbb_SAVE_ALL_OUT_r2x3_001_rbd_h15_001_0001.82_0001_15195_60_0 resumed by user
____________
ID: 63724 | Rating: 0 | rate:
/
Mod.Sense Forum moderator Project administrator Joined: Aug 22 06 Posts: 2976 ID: 106194 Credit: 0 RAC: 0
Jerry, BOINC allows you to control the amount of memory used by BOINC. Check your settings to see what % of memory you allow when computer is in use, and when idle. Often more memory is allowed when idle, and so BOINC is waiting either for another task to complete, or for the machine to go idle again to resume work on the task.
So, what you are seeing is normal, not "hung".
____________ Rosetta Moderator: Mod.Sense
ID: 63726 | Rating: 0 | rate:
/
Jerry Goggin Joined: Jun 7 06 Posts: 4 ID: 92225 Credit: 226,010 RAC: 0
I understand what you are saying, but think something else was going on. Based on my settings and running processes, task manager indicated over 500MB free physical memory available and considerably more swap space. Settings allow 50% when computer in use and 100% when computer idle, and Rosetta has had no problems during the years I've been running it. Thing I noticed was that Rosetta had disappeared from the process list -- it wasn't waiting for anything, it was gone.
____________
ID: 63727 | Rating: 0 | rate:
/
Mod.Sense Forum moderator Project administrator Joined: Aug 22 06 Posts: 2976 ID: 106194 Credit: 0 RAC: 0
The amount free is not relevant. The total size times 50% when in use, and then how many CPUs do you have?
It sounds like what you are seeing is that Rosetta often takes more memory to run then it used to, and they increased the recommended minimum system memory in the past year as well to reflect this.
I see you have 1 Windows machine, with 1 CPU and 768MB of memory. So 50% of that is less then the 512MB recommended. But when your machine goes idle and you allow 100%, the task is probably running. As soon as you use the machine to try and see it, computer is in use and it must be suspended.
With such settings, be sure you check the box to keep tasks in memory while suspended to avoid losing work everytime you sit down to use your computer and more memory is required.
If BOINC just got started, and knows that the task was using 200MB when it was last running, and does not presently have 200MB to devote to it, then it may not start the task on the task list until sufficient memory is available.
Also, your original post said the task was waiting for memory, but the message you copied said that you (the user) had suspended it. Perhaps that was as you were preparing to abort it?
Anyway, next time, note the task name and check back in a day and see if it is still "hung". It should work it's way through normally without any intervention.
____________ Rosetta Moderator: Mod.Sense
Jerry Goggin Joined: Jun 7 06 Posts: 4 ID: 92225 Credit: 226,010 RAC: 0
Really appreciate the prompt feedback. The box to "keep tasks in memory while suspended" was not checked, so that got changed. This is a very old computer which needs replacing soon. Going from Win2K to Windows 7 will be interesting, I'm sure.
____________
Don't know quite where to post this but since the change to 1.98 all workunits result in Computation error. 5.98 is running just fine as are all other projects in boinc.
- On Windows XP, units fail after 5 minutes, calculation error.
- On Windows 2000, units run for 9 hours (my patience was over, normally 3), the percentage completed is still at 0%, and the time remaining does not move from the initial value. It does not let other projects with any time to run. Restarting Boinc manager resets the clock to zero, and it starts all over again with the same behaviour. I had to suspend the project until further fix.
____________
Don't know quite where to post this but since the change to 1.98 all workunits result in Computation error. 5.98 is running just fine as are all other projects in boinc.
Just seconding this, I'm having the same behavior. Please advise if there is anything that I can report to help diagnose.
____________
ID: 63760 | Rating: 0 | rate:
/
borg Joined: Dec 4 07 Posts: 3 ID: 224173 Credit: 142,556 RAC: 0
21.10.2009 9:14:49 rosetta@home Task lr8_A_seq_score12_shake_ss1.7_rlbd_1c8c_IGNORE_THE_REST_DECOY_14949_3992_0 exited with zero status but no 'finished' file
21.10.2009 9:14:49 rosetta@home If this happens repeatedly you may need to reset the project.
This was happening repeatedly. Finally I aborted the task.
http://boinc.bakerlab.org/rosetta/result.php?resultid=289275155 Gave a compute error after running for roughly 1hr 06min 3812.641 seconds. This is a lr8 task task ID 289275155
____________
Have a crunching good day!! Live in NZ y not join Smile City?
If this isn't the venue to report such glitches, though it seems to be, please let me know :) Incidentally, what % of WU's are found to be corrupted/abortive/unfinishable, and can they be rehabilitated for purposes of this project? :)
The inquiring layman's mind wishes to know... :)
____________
The lovely lady you see isn't I, but Hayley Westenra, a classical crossover singer from Christchurch, NZ. There is no known voice as hers. Check her out- she's seraphic.
Repeated lines in the stderr txt point to this error
Exception:
failure to read decoy F_00018_0004416_0_0001 from silent-file lr8_shake_1c8c.out
[2009-10-17 20:10:35:] :: BOINC:: Initializing ... ok.
Hi
Similar error with this:
http://boinc.bakerlab.org/rosetta/workunit.php?wuid=263361396
Exception:
failure to read decoy F_00023_0001585_0 from silent-file lr8_1lou.out
CPU time 15.23438
And a different one with these:
http://boinc.bakerlab.org/rosetta/workunit.php?wuid=263361627
Unhandled Exception Detected...
- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x00436BD8 read attempt to address 0x018BD000
Sun 25 Oct 2009 15:25:52 EST|rosetta@home|Output file lr8_score12_run03_rlbd_1ugh_IGNORE_THE_REST_DECOY_14712_835_2_0 for task absent
<message>
process exited with code 1 (0x1, -255)
</message>
ERROR: Illegal attempt to score with non-identical atom set between pose and etable
ERROR:: Exit from: src/core/scoring/etable/EtableEnergy.cc line: 72
BOINC:: Error reading and gzipping output datafile: default.out
called boinc_finish
____________
Many, many more errors of a similar type before it closes down after 3 seconds
Edit: wingman errored out after 11 seconds but with very different errors reported. Best to take a look yourself. Well beyond me.
____________
ID: 63849 | Rating: 0 | rate:
/
Mike Joined: Jun 28 09 Posts: 1 ID: 324073 Credit: 6,427 RAC: 0
I found the following error when I returned to my computer this afternoon:
Unhandled exception at 0x7c9101b3 in minirosetta_1.98_windows_intelx86.exe: 0xC0000005: Access violation writing location 0x00450eb
Actually I got a dialog offering to run the Visual Studio debugger, (BOINC is running on a development machine) and this is what the debugger identified as the error.
This has happened twice before, possibly in the same work unit.
Addendum - this happens whenever the work unit runs - aborting it
Work Unit: symm_lr13_seq_score12_A_rlbd_1bkr_IGNORE_THE_REST_DECOY_15334_130_1
Environment:
OS: Windows XP Professional Service Pack 3
Processor: Intel Core2 CPU 6420@ 2.13GHz (2 CPUs)
Memory: 3584MB RAM
Page File: 2359MB used, 6421MB available
1. Do you want more debugging info if it happens again?
2. Should I abort the work unit?
I would like to know if there is a way to avoid units that run under 1.98.
Not only am I loosing houres of crunching time for R@H but all other projects are being affected. At present I am manually aborting all 1.98 units but this is a hit and miss operation. The alternative is to suspend R@H until the problem is fixed.
Not only am I losing houres of crunching time for R@H but all other projects are being affected.
Happened to me too a while back... it's why I walked away from Rosetta and went to POEM. Their WU's don't foul up. I gave Rosetta time to fix their issues, and while my experience hasn't been as bad as the last poster, I have had 2-3 comp error aborts. Also, the graphics for 1.98 often won't show when prompted, a symptom of troubles generally in the code.
The alternative is to suspend R@H until the problem is fixed.
Managers of R@H, take note of that last line. The original poster, joined by myself, just said a mouthful... we value the use of our flops!
____________
The lovely lady you see isn't I, but Hayley Westenra, a classical crossover singer from Christchurch, NZ. There is no known voice as hers. Check her out- she's seraphic.
My computer has stopped DLing any new work from the rosetta project since
27 Oct 2009 3:20:11 UTC. I double-checked all my settings. Nothing has changed,
I'm just not receiving any new work. So what's happened now, any ideas?
____________
ID: 63865 | Rating: 0 | rate:
/
Mod.Sense Forum moderator Project administrator Joined: Aug 22 06 Posts: 2976 ID: 106194 Credit: 0 RAC: 0
My computer has stopped DLing any new work from the rosetta project since
27 Oct 2009 3:20:11 UTC. I double-checked all my settings. Nothing has changed,
I'm just not receiving any new work. So what's happened now, any ideas?
Whatcha getting for messages?
____________ Rosetta Moderator: Mod.Sense
My computer has stopped DLing any new work from the rosetta project since
27 Oct 2009 3:20:11 UTC. I double-checked all my settings. Nothing has changed,
I'm just not receiving any new work. So what's happened now, any ideas?
I see you're getting tasks through from Einstein ok. Could it be something to do with debt on one project compared to another? Or the split of work between the two? Or just Boinc messing up scheduling again.
Additional info on what Boinc is reporting under the messages tab would help to pin the reason down.
____________
ID: 63869 | Rating: 0 | rate:
/
Yifan Song Forum moderator Project administrator Project developer Project scientist Joined: May 26 09 Posts: 62 ID: 318024 Credit: 7,322 RAC: 0
Mike has figured out where the bugs are and is currently working on them.
there seems to be a conflict between the symmetry code and disulphide, which is why a lot of symm runs are failing.
also, there is an api bug for zip, which causes some of the i/o problems.
hopefully we'll be able to update at the beginning of next week.
My computer has stopped DLing any new work from the rosetta project since
27 Oct 2009 3:20:11 UTC. I double-checked all my settings. Nothing has changed,
I'm just not receiving any new work. So what's happened now, any ideas?
I see you're getting tasks through from Einstein ok. Could it be something to do with debt on one project compared to another? Or the split of work between the two? Or just Boinc messing up scheduling again.
Additional info on what Boinc is reporting under the messages tab would help to pin the reason down.
Problem solved, I notice. And in a big way.
New problem - too many WUs! ;)
____________
ID: 63892 | Rating: 0 | rate:
/
Rob Lilley Joined: Jan 11 06 Posts: 11 ID: 49465 Credit: 53,921 RAC: 84
Error on this WU after a couple of wasted hours of crunching, as follows:
Unhandled Exception Detected...
- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x00436BD8 read attempt to address 0x016AE000
I was using this machine, running XP SP3 and usinb BOINC version 6.6.41
Will suspend work fetch and await developments...
____________
Mod.Sense Forum moderator Project administrator Joined: Aug 22 06 Posts: 2976 ID: 106194 Credit: 0 RAC: 0
Let's leave the sticky on this thread until the existing 1.98 WUs have had 10 days to reach their expiration.
____________ Rosetta Moderator: Mod.Sense
This may be ancient history with the release of 2.0 but I had several tasks with names like threading_bongs_pipeline_hb* hang under Windows System 7 at random percentage completion values. Bringing up the graphics window simply resulted in a blank window: I had to abort the tasks.
This may be ancient history with the release of 2.0 but I had several tasks with names like threading_bongs_pipeline_hb* hang under Windows System 7 at random percentage completion values. Bringing up the graphics window simply resulted in a blank window: I had to abort the tasks.
Mod.Sense Forum moderator Project administrator Joined: Aug 22 06 Posts: 2976 ID: 106194 Credit: 0 RAC: 0
svincent, did you get a look at the task manager? Was the processing task getting CPU? It is possible that the graphic had a problem but the processing was continuing.
____________ Rosetta Moderator: Mod.Sense
svincent, did you get a look at the task manager? Was the processing task getting CPU? It is possible that the graphic had a problem but the processing was continuing.
Graphics worked OK for other workunits. I didn't take a look at the task manager (will do next time) but was going on the combination of the Progress and Elapsed Time fields in the Boinc Manager: the former was stuck and the latter kept going. I had one such task that went on over 25 hours before I aborted it: unfortunately my results page doesn't go back far enough to find it. On the other hand I did have some workunits named threading_bongs_* complete successfully.
