Message boards :
Number crunching :
Progress doesn't grow evenly
Message board moderation
Author | Message |
---|---|
mahdia Send message Joined: 3 Apr 99 Posts: 12 Credit: 1,776,582 RAC: 0 |
Hello ! Watching the calculation of a WU I wonder why at the first 4 minutes my BOINC client says, that "time to completion" is about 5 minutes, but reaching a progress of ca. 55 %, the completion-time continuesly rises (up to 3.5 hours, which is my average WU calculation time) and the progress slow down very much ? So what's the reason for this confusing behavior ? Regards |
huns0004 Send message Joined: 14 Jun 01 Posts: 46 Credit: 3,208,956 RAC: 1 |
> Hello ! > > Watching the calculation of a WU I wonder why at the first 4 minutes my BOINC > client says, that "time to completion" is about 5 minutes, but reaching a > progress of ca. 55 %, the completion-time continuesly rises (up to 3.5 hours, > which is my average WU calculation time) and the progress slow down very much > ? > So what's the reason for this confusing behavior ? > Regards > > The algorithm for estimating the time to completion just doesn't work well for SETI units. Ignore the estimate and the progress indicator, it's not linear. |
Jaaku Send message Joined: 29 Oct 02 Posts: 494 Credit: 346,224 RAC: 0 |
|
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
|
PyroFox Send message Joined: 5 Apr 03 Posts: 155 Credit: 213,891 RAC: 0 |
Where do we see all the milestone criteria anyways? i've been searching for it to no avail.. [/url] |
Christopher Hauber Send message Joined: 10 Feb 01 Posts: 196 Credit: 71,611 RAC: 0 |
It's been a while since anyone has mentioned this problem. Maybe a good solution would just to make the time calculation logarithmic instead. Sort of "fudge" it. Still wouldn't be perfect, but would probably be closer. Especially if you tweaked the equation a little. > The algorithm for estimating the time to completion just doesn't work well for > SETI units. Ignore the estimate and the progress indicator, it's not linear. > > |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
> It's been a while since anyone has mentioned this problem. Maybe a good > solution would just to make the time calculation logarithmic instead. Sort of > "fudge" it. Still wouldn't be perfect, but would probably be closer. > Especially if you tweaked the equation a little. > > > The algorithm for estimating the time to completion just doesn't work > well for > > SETI units. Ignore the estimate and the progress indicator, it's not > linear. > > > > Maybe something like ((1-n)*original estimate + current CPU time) * (1-n) where n is the fraction done (0 to 1) or (% done / 100). This would tend to stabilize the calculation somewhat in the earlier stages, and it would converge on the same result in the end. |
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
> Where do we see all the milestone criteria anyways? i've been searching for it > to no avail.. Off of the "Your Account" page there is a spot for Pending results. This is where you would see the list of work accomplished and what you have pending. Tracking down the links you would be able to go and look at the actual results below. The issue here is that SETI@Home has such a wide audience that the amount of data sloshing aroud is HUGE. So, the core table has high activitiy. With this high activity came problems. So, the pages were turned off to allow the system to "run" without performance issues. We all hope this will end soon. I do have a description page for pending results that you may be able to get the flavor of it if you are real curious. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.