Long running work unit

[Beyond]

Instead of implementing additional M XL, etc. applications (which then might also be a bit confusing for the users to overlook)

Not confusing. Something like:
S (0-2 hour runtime)
M (2-4 hour runtime)
L (4-8 hour runtime)
XL (8-16 hour runtime)
XXL (> 16 hour runtime)

it might be an idea to dynamically adjust what is fed into S or XXL applications under special circumstances such that e.g. in case of availability of very few short WUs, the shortest ones of the XXL application would then be re-directed into S.

This would work too. Which would be easier to implement?

Regards/Beyond

[cein]

Could you please extend the following two tasks?
13306650
13306651

Thank you in advance

[yoyo]

done

[steinrar]

An I7 2600K has been chewing on WU 5286370 for around 146 hrs now, 100% CPU time. It timed out this morning.
Meaning no credit ?, waste of time ?. Should I just abort it ?. I was not aware this was a problem until now, and that
you could extend time. In school I learned 100% is finish, but this WU reached 100% days ago but still use one cpu core.

[Norman]

please let it run. if you have a Intron_qpI then it is normaly...
please post the task so that we can extend the deadline.
thx.

[yoyo]

Welcome to or forum.
Let the wu running. I extendet the deadline. Your wingman for this wu finished in 516h.
yoyo

[Michael H.W. Weber]

Instead of implementing additional M XL, etc. applications (which then might also be a bit confusing for the users to overlook)

Not confusing. Something like:
S (0-2 hour runtime)
M (2-4 hour runtime)
L (4-8 hour runtime)
XL (8-16 hour runtime)
XXL (> 16 hour runtime)

Unfortunately, such a clear definition of runtimes is not possible due to the inaccuracies the runtime forecast function (as implemented by the original developers of INFERNAL) currently suffers from.

Michael.

[Beyond]

Instead of implementing additional M XL, etc. applications (which then might also be a bit confusing for the users to overlook)

Not confusing. Something like:

S (0-2 hour runtime)
M (2-4 hour runtime)
L (4-8 hour runtime)
XL (8-16 hour runtime)
XXL (> 16 hour runtime)

Unfortunately, such a clear definition of runtimes is not possible due to the inaccuracies the runtime forecast function (as implemented by the original developers of INFERNAL) currently suffers from. Michael.

How close predictions can you get? I doubt if a certain amount of variation would upset anyone. Any level of expanded choice would be nice and I think would help the project by getting more moderately long WUs completed. Most people just don't have machines that can be counted on to stay up for hundreds of hours without either restarting or having BOINC itself cause a problem with the WU.

[yoyo]

Hello,
the current available S and XXL is enough. I will not deploy more instances. This much more effort in maintaining them and it will bring much more problems for the shared memory of the feeder who has slots foreach app, operating system and HR class.
yoyo

[Beyond]

Hello,
the current available S and XXL is enough. I will not deploy more instances. This much more effort in maintaining them and it will bring much more problems for the shared memory of the feeder who has slots foreach app, operating system and HR class. yoyo

I get it, hardware limitations and added admin effort (thought maybe it could be done with scripts). Thanks yoyo for the explanation!

[ConflictingEmotions]

I have two 'intron' WUs that need extensions as these should complete very close to or just after their respective deadlines:
5285203
5590990

Fortunately these are completed by wingmen but I still need about 14 days to finish.

Really you need to figure out why this intron WUs are so badly underestimated.

[Michael H.W. Weber]

Really you need to figure out why this intron WUs are so badly underestimated.

Well, we know that but, unfortunately, it cannot be fixed due to stochastic elements in the code.

Michael.