Message boards :
Number crunching :
Havent seen this before. (Maybe I wasnt looking heh)
Message board moderation
Author | Message |
---|---|
Gonad the Destroyer®©™ Send message Joined: 6 Aug 99 Posts: 204 Credit: 12,463,705 RAC: 0 |
|
perryjay Send message Joined: 20 Aug 02 Posts: 3377 Credit: 20,676,751 RAC: 0 |
They had a bad run on some of the tapes. It's been discussed here on a couple of threads already. At least they are dieing a quick death. :) PROUD MEMBER OF Team Starfire World BOINC |
Gonad the Destroyer®©™ Send message Joined: 6 Aug 99 Posts: 204 Credit: 12,463,705 RAC: 0 |
|
Clyde C. Phillips, III Send message Joined: 2 Aug 00 Posts: 1851 Credit: 5,955,047 RAC: 0 |
At least they cause negligible delay. |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
At least they cause negligible delay. Except for special circumstances.... With my dial-up connection at its best, an AP_v5 WU takes about 25 minutes to download. The one I got April 28 was like that, I think. I had some previously downloaded MB work running, so the AP_v5 went at the end of the queue. Around 4 am the next morning its turn came and went in 7 seconds. The phone line was of course configured as a phone line at that time... So when I got up I found that host idle, after going online I saw that the project was in difficulty so put off trying for more work until afternoon when downloads started flowing again. The error reported, the Scheduler assigned another AP_v5. With the S@H download pipe still maxed, that one took a total of about an hour of download time, with several backoffs from HTTP errors and timeouts that was about 9 hours of wall time. Murphy was there, that one also had the problem and crunched in 6 seconds. After that I got some MB work, so it was only about 18 hours of idle time. The day was not the kind I like, but interesting nevertheless. I'm not complaining, it's rather amusing in retrospect. There's another aspect which is more generally applicable. Each one of those WUs will go to at least 7 hosts. Since an 8 MB file is 64 MBits, 3 downloads occupy the full download bandwidth for 2 seconds. The ones I've seen reported so far happened on tapes from 12mr through 20mr and counting up the 'tapes' in that range suggests there are at least 10000 of the bad WUs. They'll be spread out over time so won't tie up all download bandwidth for 1 hour and 51 minutes, but are still wasting part of a somewhat limited resource. Joe |
perryjay Send message Joined: 20 Aug 02 Posts: 3377 Credit: 20,676,751 RAC: 0 |
Josef, You might want to refigure. It's not quite that bad. It appears it is only one track on the tapes from those dates. The rest of the tracks are ok. PROUD MEMBER OF Team Starfire World BOINC |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
Josef, You might want to refigure. It's not quite that bad. It appears it is only one track on the tapes from those dates. The rest of the tracks are ok. Yes, I know, but one channel of a 50.20 GB 'tape' yields about 400 AP_v5 WUs and there are 25 such (plus a few more shorter 'tapes'). OTOH there's no absolute certainty that all 25 are affected, it might be some kind of intermittent which comes and goes. In any case, the other 130000 WUs from those dates ought to be fine. Joe |
Rob.B Send message Joined: 23 Jul 99 Posts: 157 Credit: 1,439,682 RAC: 0 |
Yup I had one of those too. Pity they are still being rescheduled out to even more unfortunates. |
©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.