> Unfortunately the credit per WU has dropped by about the same percentage as
> well... but I guess that's only fair.
>
>
> - Marco
> Team
> Canada
>
Thats an logical effect of the droipping time, because credit is calculated over your crunched time.
This might be a reasonable place to post this comment and question.
I am thinking in the next several weeks to increase the length of the average workunit by about a factor of five to ten. So the run time would go from 30-60 minutes to five to ten hours. The main purpose is because this will cut down the strain on our server if there are fewer workunits to keep track of.
It doesn't make sense to do this quite yet. Until our code is stable, shorter workunits provide quicker feedback to everyone about what works and does not work.
But I'd like to know what our users think of this. Are half-day or even one-day long workunits reasonable?
> But I'd like to know what our users think of this. Are half-day or even
> one-day long workunits reasonable?
>
There's no reason I can see where long running work units would have a negative effect for the users. The only ones who would care are those of us stuck with dial-up. It would be better for me to upload a few times a day rather than every 30 min. The only concern I would have is if the result were several MB's in size.
> > But I'd like to know what our users think of this. Are half-day or even
> > one-day long workunits reasonable?
> >
>
> There's no reason I can see where long running work units would have a
> negative effect for the users. The only ones who would care are those of us
> stuck with dial-up. It would be better for me to upload a few times a day
> rather than every 30 min. The only concern I would have is if the result were
> several MB's in size.
>
But maybe it would be possible to split the result before?!
> I am thinking in the next several weeks to increase the length of the average
workunit by about a factor of five to ten. So the run time would go from
30-60 minutes to five to ten hours. The main purpose is because this will cut down the strain on our server if there are fewer workunits to keep track of.
>
> But I'd like to know what our users think of this. Are half-day or even
> one-day long workunits reasonable?
>
> Cheers,
> Bruce
>
Yes....please, please, please do it.
It would be soooo appreciated.
Especially for us dinosaurs who still have dialup.
WOW the new einstein app.4.46 is 100 % faster
)
Thats good to hear, I'm awaiting the new version :).
Greetings from Germany
Basti
Join Ad Astra
So first wu has been
)
So first wu has been calculated and time has been sunken from 40min to 30min :).
Greetings from Germany
Basti
Join Ad Astra
Unfortunately the credit per
)
Unfortunately the credit per WU has dropped by about the same percentage as well... but I guess that's only fair.
- Marco
Team Canada
- Marco
Team Canada
> Unfortunately the credit
)
> Unfortunately the credit per WU has dropped by about the same percentage as
> well... but I guess that's only fair.
>
>
> - Marco
> Team
> Canada
>
Thats an logical effect of the droipping time, because credit is calculated over your crunched time.
Greetings from Germany
Basti
Join Ad Astra
This might be a reasonable
)
This might be a reasonable place to post this comment and question.
I am thinking in the next several weeks to increase the length of the average workunit by about a factor of five to ten. So the run time would go from 30-60 minutes to five to ten hours. The main purpose is because this will cut down the strain on our server if there are fewer workunits to keep track of.
It doesn't make sense to do this quite yet. Until our code is stable, shorter workunits provide quicker feedback to everyone about what works and does not work.
But I'd like to know what our users think of this. Are half-day or even one-day long workunits reasonable?
Cheers,
Bruce
Director, Einstein@Home
I would have no problem with
)
I would have no problem with longer wu's. (Have a look on CPDN lol) But I think the deadline has also to be changed then :).
Greetings from Germany
Basti
Join Ad Astra
> But I'd like to know what
)
> But I'd like to know what our users think of this. Are half-day or even
> one-day long workunits reasonable?
>
There's no reason I can see where long running work units would have a negative effect for the users. The only ones who would care are those of us stuck with dial-up. It would be better for me to upload a few times a day rather than every 30 min. The only concern I would have is if the result were several MB's in size.
> > But I'd like to know what
)
> > But I'd like to know what our users think of this. Are half-day or even
> > one-day long workunits reasonable?
> >
>
> There's no reason I can see where long running work units would have a
> negative effect for the users. The only ones who would care are those of us
> stuck with dial-up. It would be better for me to upload a few times a day
> rather than every 30 min. The only concern I would have is if the result were
> several MB's in size.
>
But maybe it would be possible to split the result before?!
Greetings from Germany
Basti
Join Ad Astra
> I am thinking in the next
)
> I am thinking in the next several weeks to increase the length of the average
workunit by about a factor of five to ten. So the run time would go from
30-60 minutes to five to ten hours. The main purpose is because this will cut down the strain on our server if there are fewer workunits to keep track of.
>
> But I'd like to know what our users think of this. Are half-day or even
> one-day long workunits reasonable?
>
> Cheers,
> Bruce
>
Yes....please, please, please do it.
It would be soooo appreciated.
Especially for us dinosaurs who still have dialup.
Cheers.
5 or 10 hours are OK for me
)
5 or 10 hours are OK for me
http://www.guidonews.de