Message boards :
Number crunching :
Using a faster computer for a result
Message board moderation
Author | Message |
---|---|
Chiana Send message Joined: 2 Feb 05 Posts: 19 Credit: 165,472 RAC: 0 |
Is it possible to move a result that might not do the deadline into a faster computer that "takes off" and hour or so of computation and then move it back to have it finished in the original computer? |
Compare Watches Send message Joined: 24 Apr 04 Posts: 81 Credit: 2,862,596 RAC: 0 |
Ahem, I'm sure it's not the reason you're asking, but if that were possible, people could get a work unit to 99% completed then have several machines submit the unit to artificially inflate their credit without doing much work. This happened to SETI Classic (it's alleged) and they put a stop to it. Tomcat's Taekwondo Pages |
Chiana Send message Joined: 2 Feb 05 Posts: 19 Credit: 165,472 RAC: 0 |
No it's not, the machine had som accidental down-time due to a bad fuse (Similar to seti???) and have gotten somewhat "behind shedule"... I don't like the prospect of missing the dead-line after 3 days of cpu-time... > Ahem, I'm sure it's not the reason you're asking, but if that were possible, > people could get a work unit to 99% completed then have several machines > submit the unit to artificially inflate their credit without doing much work. > This happened to SETI Classic (it's alleged) and they put a stop to it. > |
Jaaku Send message Joined: 29 Oct 02 Posts: 494 Credit: 346,224 RAC: 0 |
You probably could, but if you are trying to gain credit easily ... ha, You will only be aloud to submit it once as seti knows what it has sent you and what to recieve back, you send the samw WU back 100 times it will only take one unless you hack the server to say it sent you it 100 times. Not saying this is your intension, just saying it, if it is! :D Edit: ( Just saw your post above :D ) |
Chiana Send message Joined: 2 Feb 05 Posts: 19 Credit: 165,472 RAC: 0 |
Yeah, I realize that, this is my slowest machine that I use for seti, it's on a P-III/333, my problem right now is that the deadline is about 1.5 hours away and there is still 2 hours of computing (according to boincView) left... |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
> Yeah, I realize that, this is my slowest machine that I use for seti, it's on > a P-III/333, my problem right now is that the deadline is about 1.5 hours away > and there is still 2 hours of computing (according to boincView) left... You'll probably still get credit if it's only 1/2 hour late. |
Compare Watches Send message Joined: 24 Apr 04 Posts: 81 Credit: 2,862,596 RAC: 0 |
I lost a bunch of WU's during the 48 hour server outage last week - my computers had done the work, but by the time the splitters were back online they were down as too late to be submitted. A bit frustrating, but that's how the cookie crumbles! Tomcat's Taekwondo Pages |
Chiana Send message Joined: 2 Feb 05 Posts: 19 Credit: 165,472 RAC: 0 |
Well, let's see what happens, the wu is actually in the einstein project, perhaps they have other preferences since they put out wu's that takes time to run, this particular one has been running 3 days 16 hours by now and for the moment i've adjusted my prefs to run only on einstein... |
Jaaku Send message Joined: 29 Oct 02 Posts: 494 Credit: 346,224 RAC: 0 |
> Well, let's see what happens, the wu is actually in the einstein project, > perhaps they have other preferences since they put out wu's that takes time to > run, this particular one has been running 3 days 16 hours by now and for the > moment i've adjusted my prefs to run only on einstein... > What i used to do on seti classic, when i was OBSESED (still am, building a 3.6GHZ too add) i would close all programs and endtask most of the hidden ones, i would also switch off the screen .. only gave me a few extra minutes but it works, try it if you want to risk it! :D! |
Chiana Send message Joined: 2 Feb 05 Posts: 19 Credit: 165,472 RAC: 0 |
It's a Linux box, and I've raised it's priority up to "realtime" already, can't even get a command from the keyboard in without having to wait about 10 secs for each keystroke... |
Alex Send message Joined: 26 Sep 01 Posts: 260 Credit: 2,327 RAC: 0 |
>Is it possible to move a result that might not do the deadline into a faster computer that "takes off" and hour or so of computation and then move it back to have it finished in the original computer? Actually, that can be done if you run Boinc from a removable disk, ie.. have boinc installed on a USB key or portable storage device (Mp3 player, usb hard drive etc) Exit the app from one machine. Unmount the drive, plug into another machine, restart the app. |
Chiana Send message Joined: 2 Feb 05 Posts: 19 Credit: 165,472 RAC: 0 |
well, what do you know! It just reported, with about 8 seconds to spare... Now I only have to get into contact with that machine again to restore it to normal... |
RB Send message Joined: 7 Mar 00 Posts: 103 Credit: 1,084,436 RAC: 0 |
> >Is it possible to move a result that might not do the deadline into a > faster computer that "takes off" and hour or so of computation and then move > it back to have it finished in the original computer? > > Actually, that can be done if you run Boinc from a removable disk, ie.. have > boinc installed on a USB key or portable storage device (Mp3 player, usb hard > drive etc) > Exit the app from one machine. Unmount the drive, plug into another machine, > restart the app. Very interesting !! Would it decrease the WU cruncing time? Starting a new thread with that topic of Boinc on USB Key. LOL. |
Speedy67 & Friends Send message Joined: 14 Jul 99 Posts: 335 Credit: 1,178,138 RAC: 0 |
> > Yeah, I realize that, this is my slowest machine that I use for seti, > it's on > > a P-III/333, my problem right now is that the deadline is about 1.5 hours > away > > and there is still 2 hours of computing (according to boincView) left... > > You'll probably still get credit if it's only 1/2 hour late. I think you're right.. i even got credit just now for a WU that was 8 days past the deadline... http://setiweb.ssl.berkeley.edu/result.php?resultid=35477962 Greetings, Speedy67 |
Ingleside Send message Joined: 4 Feb 03 Posts: 1546 Credit: 15,832,022 RAC: 13 |
> > I think you're right.. i even got credit just now for a WU that was 8 days > past the deadline... > > http://setiweb.ssl.berkeley.edu/result.php?resultid=35477962 This is since some of the backend-services have been backlogged or turned off. When everything is running as it should, the other result-files for the wu would have been deleted immediately after you reached your deadline, meaning no credit for you. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13746 Credit: 208,696,464 RAC: 304 |
> When everything is running as it should, the other result-files for the wu > would have been deleted immediately after you reached your deadline, meaning > no credit for you. Not so. Several times i've received credit for workunits that have been 3-5 days past the deadline. It just depoends on whether or not the other systems with the late result are also late themselves. Grant Darwin NT |
RandyC Send message Joined: 20 Oct 99 Posts: 714 Credit: 1,704,345 RAC: 0 |
> > When everything is running as it should, the other result-files for the > wu > > would have been deleted immediately after you reached your deadline, > meaning > > no credit for you. > > Not so. > Several times i've received credit for workunits that have been 3-5 days past > the deadline. It just depoends on whether or not the other systems with the > late result are also late themselves. > My understanding is that if you get your (valid) result in before the validator finds 3 on-time valid results, you'll still get credit...I could be wrong of course. |
Ingleside Send message Joined: 4 Feb 03 Posts: 1546 Credit: 15,832,022 RAC: 13 |
> > Not so. > Several times i've received credit for workunits that have been 3-5 days past > the deadline. It just depoends on whether or not the other systems with the > late result are also late themselves. > Well, based my comments on the wu linked to, and for this wu the only reason he got credit was due to backlogged services... Earlier when a seti-wu was initially sent out to 3 different users, missing your deadline wasn't normally a problem since for most had to re-issue to someone else and this would take atleast 1 day. But, now then a seti-wu is initially sent out to 4 different users, wouldn't gamble on there being another user also missing his deadline... Also, the systems isn't currently running "as it should", partially due to all the download-errors from v4.13-clients, but also since Validator have been backlogged and file_deleter reportedly also is backlogged. BTW, as long as the Assimilator is turned off, no newly validated or errored-out wu is being deleted, since the file_deleter must wait on the assimilator... |
Benher Send message Joined: 25 Jul 99 Posts: 517 Credit: 465,152 RAC: 0 |
Yes it is possible. I've done it before. You can't get more credit or anything, nor is cheating possible in this way. As far as any project is concerned one returned validatable result for each one passed out, to a computer with specific number. What happened in my case, was one machine had a glitch of some kind and requested 70 LHC WUs (during 2 days). Well since it was a Pentium II Celeron 500, it could never have finished them anywhere near deadline. So I copied the entire directory of BOINC over to another machine, not running boinc then, and started up the boinc_gui application with the WUs already in their proper place, and all XML files copied as well (whole directory remember). As far as LHC server was concerned it was the same as if I had installed the original Hard Drive into a new motherboard/new CPU setup. The temporary machine was a Pentium 4 HT 2.8Ghz and crunched through the results very nicely before deadline. I made sure it didn't download any new WUs by making its "connect every" number very small. |
©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.