There will always be "longer" jobs in an experiment for the toughest compounds.

That you had to wait 1.5 days for a whomp in credit I don't see. from 15:50 yesterday to now, UTC is not 24 hours. If you had to wait, then either there was a second job sent out after yours returned for verification, or an initial distribution of 2 was pushed out indicating that your client is up for "board" review

[Added: Today's noon chart, not published, shows 9.5 hours as average run time, meaning that the target 7-8 hours was exceeded somewhat. We'll survive.]

Hey Astrolab,

If you forgot it, take it on time.
If you don't forgot it, double the recommended dose

The problem with these variable length jobs is that if WCG would raise the average to 3 times, one would do it in 9 hours instead of 3 and the other part time computer would take 72 hours instead of 24, getting these into serious deadline stress. With the return times in mind, the batches need to continue to complete and flow back to the scientists, it would lock out many volunteers with lesser resource and that would be counter to the objective of have as many volunteers able to contribute to FAAH, or any other WCG project.

That said, there are developments underway to be able to cut work to size in a more dynamic fashion, meaning that powerful PCs get heavier jobs and not so powerful less hefty version so all can do it like the RICE project in e.g. 8 hours.

Present deadlines for all projects is 12 days btw. HPF2 is the exception with 20 days.

cheers

A 72 hour processing time would be accounted for in the estimated time to completion (i.e., 72 hours from now). Therefore, the part-time computer's job download algorithm would only download the appropriate number of jobs for a 12 day window (3 jobs in this case). So the part-time computer would run 3 or 4 jobs in the 12 days, while the 24/7 system would download and run about 31 jobs in the same 12 days. In both cases, network traffic will be cut by 2/3. Thus, I do not see how your scenario results in any actual harm. I do see, however, a benefit - namely, the 2/3 reduction in network traffic.

Please note that I have consistently and specifically proposed that the job length be set at 12 to 24 hours (as opposed to 8 hours currently). The upper bound on job length is, rather obviously, imposed by the need to avoid deadline difficulties. Assuming a 24-hour job length and a 12-day deadline window, the client needs to crunch for an average of only 2 hours per 24-hour period.