Message boards :
Number crunching :
Why did they limit dl's at all?
Message board moderation
Author | Message |
---|---|
PhonAcq Send message Joined: 14 Apr 01 Posts: 1656 Credit: 30,658,217 RAC: 1 |
I can't find out why they had to limit downloads at all, when they switched to the 3d off line schedule. Any morosoph knows not to make multiple changes at once, when the consequences are essentially completely unknown. So they must have had a reason. I hope so because so many of us are runnin' dry. I love that Norah Jones clip, "what am I to you". Good post. Apropos. |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Makes ME cry........ I thought the heartfelt thought was appropriate. So again I ask him What Am I to You? "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
I can't find out why they had to limit downloads at all, when they switched to the 3d off line schedule. So that everybody could get some work, not just some people. Grant Darwin NT |
JohnDK Send message Joined: 28 May 00 Posts: 1222 Credit: 451,243,443 RAC: 1,127 |
I can't find out why they had to limit downloads at all, when they switched to the 3d off line schedule. That is/was a good idea, but when "we" can't get GPUs then it's not working optimal. Don't have a clue if it's possible to fix that and still have the limit in some fashion. Joe talked about only 1% gets affected so maybe it's not a top priority on the fix list. |
Bill Walker Send message Joined: 4 Sep 99 Posts: 3868 Credit: 2,697,267 RAC: 0 |
Joe talked about only 1% gets affected so maybe it's not a top priority on the fix list. Yes, but what % of annoyed posters is affected? ;) |
PhonAcq Send message Joined: 14 Apr 01 Posts: 1656 Credit: 30,658,217 RAC: 1 |
All participants get work? Please. That sounds like progressivist chant. With that line of reasoning, we'll be asking, say, NASA to focus on third world out-reach programs pretty soon. Oops! Barry has already done that! Oh, well, square peg round hole... look who is in charge of the asylum. I thought that the point of our part of the project (when I'm not being cynical) is to produce processed wu's. So I don't see the reason for throttling D/L's from that perspective, yet. Put another way, if somebody like me, without any GPU's, gets angry and walks to another project, that would not be such a loss. But chase a GPU'er away, and the damage is much larger. |
Bill Walker Send message Joined: 4 Sep 99 Posts: 3868 Credit: 2,697,267 RAC: 0 |
But chase a GPU'er away, and the damage is much larger. People have been saying that for years, and moaning as the total number of users goes steadily down. But, if I read the graphs at BOINCstats correctly, the rate of total daily credit has been going steadily up for the last year, and even seems to be increasing the rate of increase. If the guys at Berkeley see this, and see people regularly asking for more WUs, where are they going to spend there plentiful spare time and money - hand-holding a few users to keep them from drifting, or getting on with the science? |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
I can't find out why they had to limit downloads at all, when they switched to the 3d off line schedule. 1. Jeff expressed some concern about running with the download bandwidth pegged for a long time. Although last Monday's run was impressive, running any system at peak for a long time has risks. {Anybody had to back off an overclock slightly to adapt to higher ambient temperatures lately?} 2. The BOINC defaults are something like "connect interval" .01 day, extra 0.25 days. IMO a very high majority of users have simply set up BOINC and attached to S@H as asked on the home page. All who have done so should be out of work at the end of each 3 day outage, getting those hosts back doing productive work is the civil thing to do. The oft-quoted idea the project doesn't guarantee 24/7 work is of course true, but the project would like to deliver work to all hosts whenever practical. 3. Statistically, hosts with large cache settings ought to still be busy at the end of the outage. Delivering large amounts of work to them simply to rebuild their caches should be delayed until other hosts are back doing productive work. ....... OK, none of that is a help for those who have gone beyond what's expected. Running optimized applications with an app_info.xml with or without <flops> entries is risky, changes unrelated to the 3 day outage make it hard to judge which way is the least risk. Hosts doing both GPU and CPU crunching are particularly at risk, and using Reschedule 1.9 magnifies the probability of sufficient imbalance to make the limits counterproductive. Transients from the failed attempt to use server-side averages to provide per application DCFs for anonymous platform hosts probably haven't faded completely either. With 6.5 hours project uptime, the servers seem to be doing well. The result received rate is down close to normal, the assimilator queue hasn't grown to an alarming size, etc. My guess is Jeff may increase the limits before his quitting time today, or sometime later this evening. But he'll have a much more detailed view of how the servers are doing, and almost certainly a different perspective on what's desirable for the weekend. Joe |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
Network traffic has just started tapering off. Took about 2-2.5 hrs longer than the drop ofter the intial surge last week. Will be interesting to see how long the priods of low traffic last this time round, and how long the following surges last for. Grant Darwin NT |
soft^spirit Send message Joined: 18 May 99 Posts: 6497 Credit: 34,134,168 RAC: 0 |
Josef, seems like Jeff is on the right track. Still some big crunchers hurting because they have over 140 in their CPU's..(from previous cache I presume).. so they are not getting any GPU's.. the 5/40 portion looks like it did great. I would suggest easing that up through the weekend.. to see how it flies. Thumb and 1/2 up. Good progress. Janice |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 65745 Credit: 55,293,173 RAC: 49 |
Not Me Joe, I only backed off My overclock of 3.40GHz to 2.98GHz cause I have two GTX295 cards operating off of a Corsair TX 850w psu, Plus now I'm running at a stock 1.35v cpu voltage vs 1.45v before. The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
Network traffic has just started tapering off. Looks promising- Network traffic has just kicked off again, about 2.5hrs since it dropped off. Grant Darwin NT |
James Nelson Send message Joined: 23 Mar 02 Posts: 381 Credit: 4,806,382 RAC: 0 |
Network traffic has just started tapering off. no that just means everyone chewed threw there limit and are requesting more |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
Network traffic has just started tapering off. Which is a good sign as last week the periods of activity were much shorter & periods of inactivty much longer. Grant Darwin NT |
James Nelson Send message Joined: 23 Mar 02 Posts: 381 Credit: 4,806,382 RAC: 0 |
Network traffic has just started tapering off. thats just because of the increase quota for GPU, starting to feel second class just because I'M a CPU cruncher... |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
Network traffic has just started tapering off. Network traffic dropping off again. Unfortunately the Server Stats aren't updating, so i can't see how things are going there. EDIT- spoke too soon, traffic drop was just a glitch. Still going strong. Grant Darwin NT |
JDWhale Send message Joined: 6 Apr 99 Posts: 921 Credit: 21,935,817 RAC: 3 |
I said it last week and I'll repeat it again now... Throttling back downloads after the outage is BRILLIANT !!! Uploads went quickly and without a hitch and downloads are available as needed... BRILLIANT !!! |
soft^spirit Send message Joined: 18 May 99 Posts: 6497 Credit: 34,134,168 RAC: 0 |
I said it last week and I'll repeat it again now... Not everyone is getting downloads as needed.. and people have not yet been able to stock up for next one... so ALMOST brilliant. Steps on the right direction. But not quite there yet. Janice |
JDWhale Send message Joined: 6 Apr 99 Posts: 921 Credit: 21,935,817 RAC: 3 |
Not everyone is getting downloads as needed.. and people have not yet been able to stock up for next one... so ALMOST brilliant. Plenty of time to 'stock up' as seen last Monday... Apriori is being able to upload and report results after the outage. In the past this might take 24+ hours, now it's almost instantaneous with new work distributed to the hosts that need it most. Without upload/reporting new downloads were impossible... Not anymore. BRILLIANT !!! (Agreed, some fine tuning is desirable... but it's still much, much better than after prior outages.) Did I say BRILLIANT ? |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
Network traffic has just started tapering off. Looks like we've finally reached the end of the 2nd surge, rather abruptly. Grant Darwin NT |
©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.