For a long time now I have my 'Target CPU run time' parameter set to 10 hours. This past week I have noticed several WUs on my system which completed in times that were NOT near 10 hours. I'm wondering how come ??

One WU ran 6 nstructs from 6 attempts and took 8.8 hours total. Was that close enough to 10 hours for it to not make another attempt? Another WU ran 1 nstruct from 1 attempt and took 5.8 hours total. At that rate I _would_ have expected it to try another attempt. A similar one ran 2 nstructs from 2 attempts in 7.4 hours total.

On the other hand, there was one WU that ran 2 nstructs from 4 attempts in 13.3 hours total. I guess it stopped when it realized it was over my 10 hour target.

Are such deviations from my 10 hour target normal ?
.

...Are such deviations from my 10 hour target normal ?

In a word, Yes.

The time setting has some variation due to the run length of a single model for a particular Work Unit. In some cases other activities on the computer can slow the processing, and that is not taken into account in the calculation. So if the time setting predicts that the work unit could run three models in the 10 hours, and something slows the system down, it will still run the 3 models but it will run over the time. In most cases if the up front calculations show it will run over the 10 hours if it tries to run the third model, it will only run 2 and it will finish early.

In any case the variation is normal.

---

Moderator9
ROSETTA@home FAQ
Moderator Contact

One WU ran 6 nstructs from 6 attempts and took 8.8 hours total.

So each one took about 88 minutes or 1.47hrs. A 7th model would have exceeded your 10 hr preference.

Another WU ran 1 nstruct from 1 attempt and took 5.8 hours total.

So a second model would be estimated to take 5.8hrs more, and would have exceeded your preference.

A similar one ran 2 nstructs from 2 attempts in 7.4 hours total.

That's 3.7hrs each, the next one would have exceeded your preference.

On the other hand, there was one WU that ran 2 nstructs from 4 attempts in 13.3 hours total.

It happens, perhaps the estimated time of the next run was thrown off by other activity on the PC as suggested below.

More often than not, the WUs finish before your preference. The golden exception to that being when it takes longer than your preference to complete model 1.

---

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/

More often than not, the WUs finish before your preference. The golden exception to that being when it takes longer than your preference to complete model 1.

Not counting the ones I asked about, of the last 14 WUs (that had finished recently) 4 took longer than my preference. [Their nstructs/attempts were 32/34, 52/54, 7/8, 4/5.]

That's a sizeable minority that finished longer than my preference.
.

I don't generally look at my WUs in that much detail to see the decoys vs nstructs. But I ran through many of my recently reported WUs and find that all of those I looked at had decoys = nstructs.

Perhaps you've uncovered the root cause of the longer running WUs.

Are these longer running WUs causing you a problem somehow?

---

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/