Message boards :
Number crunching :
This computer has reached a limit on tasks in progress??
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 . . . 8 · Next
Author | Message |
---|---|
Fred J. Verster Send message Joined: 21 Apr 04 Posts: 3252 Credit: 31,903,643 RAC: 0 |
Had 2 CUDA/CAL hosts, turned off, cause of the heat today, 36C and still 30C. atm., it's crazy lazy weather. And got, after UPLoading a few hundred tasks, a bunch of new tasks, MB WU's, excellent for this host, a QX9650(@3.4G;1450MHz FSB;1x9800GTX+ and 1 GTS250. Running Windows XP64, BOINC 6.10.56 (x64) and Optimized CPU (SSSE3x) and GPU app. And use REScheduler, so I won't send back a REJected/killed, VLAR! Use the [i]automatic mode, depending of the amount of work, between 3 and 6 hours. No RESents of killed VLARS, since it's a waiste and why NOT crunch them, as well, as I was astonished, how much tis happens. Including myself, no problems, only when you have > 250 - 400 WU's, you should check, Connecting to Local Host, or password incorrect(?). Hope I get some ASTRO PULSE tasks. (Take 2.5 -6 hours) |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Hm, outgoing traffic drops considerably and we have only half week to fill caches and even before after small outage it took much more hours on maxed bandwidth to low work demand... that is, very likely that almost all fast enough hosts are underloaded now and will not refill caches due to remaining time before next outage... I have big doubts in usefulness of such work restriction now. |
-= Vyper =- Send message Joined: 5 Sep 99 Posts: 1652 Credit: 1,065,191,981 RAC: 2,537 |
|
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13727 Credit: 208,696,464 RAC: 304 |
I have big doubts in usefulness of such work restriction now. In the Tech News Matt posted about several software updates they did. Wouldn't surprise me if this present restriction is an unintended result of that. There have been times in the past were updates & patches have clobbered existing tweaks & settings. Grant Darwin NT |
JDWhale Send message Joined: 6 Apr 99 Posts: 921 Credit: 21,935,817 RAC: 3 |
This is brilliant! Normally after a significant outage I can't upload my results and therefore can't download any new tasks. However this time that's not the case. Uploads were no problem after the outage and a few new tasks (number is 20) were downloaded to the hosts that had run dry. I can hope that this is the intended behavior and more WUs will flow freely as the network/server traffic after the outage eases. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
This is brilliant! The problen that it eased already, look graph. But limit still here so currently outgoing bandwidth heavely underused (and we will have it only half of prev time). |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
Sadly, While my CPU queue remains chock full of Astropulse and multibeam VLAR goodness, the New NinjaB@5t!Red is stone cold idle. Oh well. "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
perryjay Send message Joined: 20 Aug 02 Posts: 3377 Credit: 20,676,751 RAC: 0 |
It may be just to give it some time for the servers to process all the uploads we sent in this morning. I notice though the pendings are dropping fairly quick I still have a lot that have been finished by both me and my wingman that are still waiting validation. I hope that once they get caught up they will open the floodgates again. PROUD MEMBER OF Team Starfire World BOINC |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13727 Credit: 208,696,464 RAC: 304 |
This is brilliant! I doubt the new limit has had any bearing on you being able to upload all your results & download new work as the servers were being hammered as hard as they ever have been even with the limitation in place. It is most likely a result of the upgrades they did while the system was down. Grant Darwin NT |
JDWhale Send message Joined: 6 Apr 99 Posts: 921 Credit: 21,935,817 RAC: 3 |
This is brilliant! But if the downloads/outstanding WUs are limited to 20 tasks per host after the outage, the download pipe won't stay clogged for 24+ hours, thus blocking uploads as in the past. Bottom line... my host has work shortly after the servers came back online. In the past it might take a full day before my results could be fully uploaded and requests for new work issued. Again... This is brilliant!!! |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13727 Credit: 208,696,464 RAC: 304 |
But if the downloads/outstanding WUs are limited to 20 tasks per host after the outage, the download pipe won't stay clogged for 24+ hours, thus blocking uploads as in the past. I agree, but what i'm saying is that all those uploads went through early on- when the download pipe was as full as it's ever been. Not only that, but the uploads went through at almost the highest rate ever- while downloads were still going flat out. Which would indicate that the reduced upload bottleneck wasn't as a result of the limit on tasks in progress, but more likely upgrades/ patches done while the system was down. Or the empty queues from the system being down. Grant Darwin NT |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 65737 Credit: 55,293,173 RAC: 49 |
This is brilliant! I have 4 gpus here and I have no work for them, Brilliant? Sounds stupid to Me, As in Your not wanted, Go away... 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: 13727 Credit: 208,696,464 RAC: 304 |
Hmm, just got 15 Work Units to crunch, however the To Completion times are way, way, way too long. This is the first time it's requested work since they implemented the changes of a couple of weeks ago- as that resulted in estimated To Completion times that were way, way, way to short & my 4 day caches ended up with 10 days worth. I take it those chages have been reversed, or the updates have undone them? EDIT Just edited my DCF to see what would happen. Prior to the previous server changes it was around .1 most of the time, found it was around .95. Changed it to .1 again & downloaded & whole bunch of work, several times, till i got the "This computer has reached a limit on tasks in progress" message again. I probably got about 30 or so Work Units before the message came up again. Enough to keep me going for 1.5, maybe 2 days, but not enough to fill my 4 day cache. Grant Darwin NT |
Cameron Send message Joined: 27 Nov 02 Posts: 110 Credit: 5,082,471 RAC: 17 |
Like I already mentioned in one of the other threads.. I believe the Milkyway Limit is 6 Tasks per Core Got 11 WUs 15 minutes ago to go with the remaining 50 of the 78 given before the outage. So I'm Fine for the Time Being. The SETI app Data says http://setiathome.berkeley.edu/host_app_versions.php?hostid=5178843 My Guess is this new limit Includes everything (In Progress/Valid/Pending/Errored) hopefully I'll be able to get a few more when all this gets written to the database. |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 65737 Credit: 55,293,173 RAC: 49 |
Has anyone here generated any -12 errors? I have, 6 in the last 7 days... Out of thousands of good results... The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's |
perryjay Send message Joined: 20 Aug 02 Posts: 3377 Credit: 20,676,751 RAC: 0 |
I only had one -12 error during the outage. Turned in this morning. PROUD MEMBER OF Team Starfire World BOINC |
SciManStev Send message Joined: 20 Jun 99 Posts: 6652 Credit: 121,090,076 RAC: 0 |
Has anyone here generated any -12 errors? I have, 6 in the last 7 days... Out of thousands of good results... I have 2 errors with an exit status of 1. One was from the CPU, and one was from the GPU. Some of my wingmen timed out, but that was bound to happen. Steve Warning, addicted to SETI crunching! Crunching as a member of GPU Users Group. GPUUG Website |
Todd Hebert Send message Joined: 16 Jun 00 Posts: 648 Credit: 228,292,957 RAC: 0 |
Unfortunatly some of us have hosts that will process more than 20 WU's at a time. Case in point is a dual Intel X5680 system that I have with 2 Fermi cards running two instances per card. That is a total of 28 tasks at a time that could be processed if this limitation was lifted. Also I am building a system with a Quad sockets and each cpu has 8 cores and hyperthreading - and two fermi cards to boot. That will be 68 tasks at a time processing. If this behavior is temporary and expected that is fine but if this is the way it is going to be I am seriously considering other options for my resources. This system is being built as a mega-cruncher and has a significant price tag associated with it. If it cant be used to its full potential I won't bother. I'm not saying I'm dumping S@H or anything like that but do want to make it known that there are individuals with systems that are more than able to process more than 20 WU's at a time and a higher limit should be implemented. Todd |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13727 Credit: 208,696,464 RAC: 304 |
Has anyone here generated any -12 errors? I have, 6 in the last 7 days... Out of thousands of good results... I've had no errors at all. Grant Darwin NT |
Rasputin Send message Joined: 13 Jun 02 Posts: 1764 Credit: 6,132,221 RAC: 0 |
Good, maybe they'll lift the limit so we can build a cache. |
©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.