Do people have an idea about what fracton of work units that are getting stuck, and whether any proteins or run conditions are getting stuck more frequently than others?

In my view you won't get a reliable answer to this, David, as the variation between machines is enormous - probably luck of the draw rather than anything systematic.

I had 17 machines runnning Rosetta over the break. I could pick out spurious patterns on one and refute them on another. Four boxes got 90% downtime for a couple of days, one box got off unscathed, the others varying in between.

If I'd had just one of those boxes what story I'd give you would depend on which box it was.

I'd suggest getting an SQL wizard to coax the frequencies of errors etc and run-times out of the returned work pile. Even better because it is less dependent on cpu speed, frequecies of errors against credit claims for the various flavours of wu.

It won't have so much detail but at least it won't depend so much on the vagaries of what wu went to which observant/too-busy users.

River~~

plesae check this WU, look at the points wasted :(
(and yes, the last 95 points were mine :()

https://boinc.bakerlab.org/rosetta/workunit.php?wuid=3761771

Do you mean one like this?

---

Along with that, the computation errors, and bandwidth required for data downloads, it's not looking so good.

---

Job NO_SIM_ANNEAL_1hz6_228_5495_0 went stuck here. 12hrs on 2% on a PIII 733mhz

---

The signature is away on the moment, just leave a message after the beep

Job MORE_FRAGS_W_BARCODE_2reb_229_5862 stuck for 9hrs at 1%

MORE_FRAGS_W_BARCODE_2reb_229_5862

XP2100+ o/c 13x166 (Lister) stable prime95 / memtest86+ et al

---

Hi,...

MORE_FRAGS_W_BARCODE_2reb_229_6182_0 (Result ID 5827255) been running for 5:30 CPU time, with 1% progress and more than 12 hours to completion. Should I Abort? ... be Credited?

O&O

These "stuck at 1%" results _seem_ to restart and complete successfully if you exit BOINC and relaunch it. The project has not said that credit will be granted if you abort them, so I can't advise that - it would help them locate the problem, however, if you could get the last 20 lines from the stdout.txt file in the slots directory and paste them here before you relaunch BOINC.

---

These "stuck at 1%" results _seem_ to restart and complete successfully if you exit BOINC and relaunch it. The project has not said that credit will be granted if you abort them, so I can't advise that - it would help them locate the problem, however, if you could get the last 20 lines from the stdout.txt file in the slots directory and paste them here before you relaunch BOINC.

Thanks Bill.

I did abort the result before having the chance to read your advise... which resulted ...
07/01/2006 17:14:42|rosetta@home|Unrecoverable error for result MORE_FRAGS_W_BARCODE_2reb_229_6182_0 (aborted via GUI RPC)
07/01/2006 17:14:43|rosetta@home|Computation for result MORE_FRAGS_W_BARCODE_2reb_229_6182_0 finished

Reporting 2 good results for Rosetta or PrimeGrid...probably 5 hours of CPU crunching time be added to a Climateprediction unit if this result was not a...stuck...turns to be a waste of a normal BOINC operation.

The project did not say that by crunching its results you may experience a lower Average creditis/day neither, did it?

Until "they" locate the problems causing such "abnormalities", I'll be allowing 30 minutes to have a Rosetta result...progresses beyond the 1%...otherwise I'll abort it.

O&O (Systems management 101: Kill the monster while it is young)

Hi,
I didn't read much of this thread, so I'm just assuming you still need information. I had this WU which stayed at 1% for 2 hours. It was still in the ab initio phase, which I've never seen before. I restarted BOINC and it completed normally.

MrS

---

Scanning for our furry friends since Jan 2002

Wao, I got one :-)

WU-Name: INCREASE_CYCLES_10_1hz6_226_6922_0

The last 20 lines of stdout:

Size: 3 NUMBER OF FRAGS FOR POS: 53 200
Size: 3 NUMBER OF FRAGS FOR POS: 54 200
Size: 3 NUMBER OF FRAGS FOR POS: 55 200
Size: 3 NUMBER OF FRAGS FOR POS: 56 200
Size: 3 NUMBER OF FRAGS FOR POS: 57 200
Size: 3 NUMBER OF FRAGS FOR POS: 58 200
Size: 3 NUMBER OF FRAGS FOR POS: 59 200
[T/F OPT]New TRUE value for [-jitter_frag]
[REAL OPT]Default value for [-jitter_amount] 2
[STR OPT]New value for [-jitter_variation] gauss.
score0 done: (best, low) rms
2 0 10.9471054
---------------------------------------------------------
score1 done: (best, low) rms (best,low)
-4.41134071 -17.6313858 8.64116001 4.5229826
standard trials: 20000 accepts: 585 %: 2.925
-----------------------------------------------------
Alternate score2/score5...
kk score2 score5 low_score n_low_accept rms rms_min low_rms
0 -27.503 -12.841 -27.503 28 4.523 3.605 4.523


I have saved the whole boinc-directory, so, if you are interested, I can zip it for you and put it on one of my servers so that you can download it from me

The WU should normally take 6 hours on this machine; until now it has taken 6 hours and says 1%

---

Supporting BOINC, a great concept !

Wao, I got one :-)

WU-Name: INCREASE_CYCLES_10_1hz6_226_6922_0

The last 20 lines of stdout:

Size: 3 NUMBER OF FRAGS FOR POS: 53 200
Size: 3 NUMBER OF FRAGS FOR POS: 54 200
Size: 3 NUMBER OF FRAGS FOR POS: 55 200
Size: 3 NUMBER OF FRAGS FOR POS: 56 200
Size: 3 NUMBER OF FRAGS FOR POS: 57 200
Size: 3 NUMBER OF FRAGS FOR POS: 58 200
Size: 3 NUMBER OF FRAGS FOR POS: 59 200
[T/F OPT]New TRUE value for [-jitter_frag]
[REAL OPT]Default value for [-jitter_amount] 2
[STR OPT]New value for [-jitter_variation] gauss.
score0 done: (best, low) rms
2 0 10.9471054
---------------------------------------------------------
score1 done: (best, low) rms (best,low)
-4.41134071 -17.6313858 8.64116001 4.5229826
standard trials: 20000 accepts: 585 %: 2.925
-----------------------------------------------------
Alternate score2/score5...
kk score2 score5 low_score n_low_accept rms rms_min low_rms
0 -27.503 -12.841 -27.503 28 4.523 3.605 4.523


I have saved the whole boinc-directory, so, if you are interested, I can zip it for you and put it on one of my servers so that you can download it from me

The WU should normally take 6 hours on this machine; until now it has taken 6 hours and says 1%

thanks. I'll try to reproduce this locally.

---

Wao, I got one :-)

WU-Name: INCREASE_CYCLES_10_1hz6_226_6922_0

The last 20 lines of stdout:

Size: 3 NUMBER OF FRAGS FOR POS: 53 200
Size: 3 NUMBER OF FRAGS FOR POS: 54 200
Size: 3 NUMBER OF FRAGS FOR POS: 55 200
Size: 3 NUMBER OF FRAGS FOR POS: 56 200
Size: 3 NUMBER OF FRAGS FOR POS: 57 200
Size: 3 NUMBER OF FRAGS FOR POS: 58 200
Size: 3 NUMBER OF FRAGS FOR POS: 59 200
[T/F OPT]New TRUE value for [-jitter_frag]
[REAL OPT]Default value for [-jitter_amount] 2
[STR OPT]New value for [-jitter_variation] gauss.
score0 done: (best, low) rms
2 0 10.9471054
---------------------------------------------------------
score1 done: (best, low) rms (best,low)
-4.41134071 -17.6313858 8.64116001 4.5229826
standard trials: 20000 accepts: 585 %: 2.925
-----------------------------------------------------
Alternate score2/score5...
kk score2 score5 low_score n_low_accept rms rms_min low_rms
0 -27.503 -12.841 -27.503 28 4.523 3.605 4.523


I have saved the whole boinc-directory, so, if you are interested, I can zip it for you and put it on one of my servers so that you can download it from me

The WU should normally take 6 hours on this machine; until now it has taken 6 hours and says 1%

thanks. I'll try to reproduce this locally.

---

David,...

I have two INCREASE_CYCLES_10_1xxx_226_xxxx_0 (Same Batch I presume) in a "ready to run" status, should I abort?

Thanks,
O&O