Message boards :
Number crunching :
6.08, Cuda, and yay!
Message board moderation
Author | Message |
---|---|
Astlor.ca.SETI@home Send message Joined: 29 Jun 99 Posts: 12 Credit: 2,180,506 RAC: 0 |
Well, I've been running 6.08 for a while now, ever since it came out in main. It runs good, no need to abort any "stuck" units, and FINALLY that 8600 on my HTPC/File Server is used for SOMETHING. My system, based on workunits completed per day seems to have only one thing to say about this new Cuda 6.08: OM NOM NOM NOM Have a great week! |
Borgholio Send message Joined: 2 Aug 99 Posts: 654 Credit: 18,623,738 RAC: 45 |
Every now and then I get a stuck workunit with 6.08. I have a feeling it's a bug in the Boinc Manager however, as restarting the client seems to restart any stuck workunits. You will be assimilated...bunghole! |
Spear Send message Joined: 15 Nov 01 Posts: 49 Credit: 6,365,604 RAC: 0 |
It gave me lockups pretty consistently and got removed just as quickly as the previous release was. |
GM3EIY Send message Joined: 1 Nov 08 Posts: 3 Credit: 1,538,756 RAC: 0 |
Can anyone throw any light in this please, the referenc to "The number of results detected exceeds the storage space allocated." is the point in question. is this on my nachine? or the server? OF DUE TO THE 30 SPIKE COUNT? or what and if its me how do I fix it please, seen a few of these now. Mike GM3EIY/VP8DIA <core_client_version>6.4.5</core_client_version> <![CDATA[ <stderr_txt> setiathome_enhanced 6.02 DevC++/MinGW libboinc: 6.3.6 Work Unit Info: ............... WU true angle range is : 2.606176 Optimal function choices: ----------------------------------------------------- name ----------------------------------------------------- v_BaseLineSmooth (no other) v_vGetPowerSpectrumUnrolled 0.00013 0.00000 sse3_ChirpData_ak 0.01360 0.00000 v_vTranspose4 0.00602 0.00000 AK SSE folding 0.00097 0.00000 SETI@Home Informational message -9 result_overflow NOTE: The number of results detected exceeds the storage space allocated. Flopcounter: 9429565434.151686 Spike count: 30 Pulse count: 0 Triplet count: 0 Gaussian count: 0 called boinc_finish </stderr_txt> |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14653 Credit: 200,643,578 RAC: 874 |
Can anyone throw any light in this please, the referenc to Yes, it's the 30 spike count that it's referring to, though it might have been pulses, triplets or gaussians, or a combination of all four - the limit is 30 in total across all categories. Assuming that your 'wingmate' (the other BOINC user working on exactly the same data pack, to ensure consistency in the results) gets the same answer, there's nothing to worry about at your end at all. You'll be able to check tomorrow whether the task is marked 'valid' and whether you've been granted credit - no point in looking now, because the project validators have been playing up all weekend and there's a massive backlog. If you find any cases where your computer has reached the 30 pulse limit, but your wingmate hasn't, then by all means post again - but it very rarely happens. When it does, it can sometimes be a symptom of computer problems like overheating. These 'overflow' results happen because we're searching for incredibly faint radio signals amongst the background noise recorded at the Arecibo radio telescope. Sometimes the radio reception is contaminated by earthly sources - there's a military radar nearby which is a consistent culprit. When there's local contamination of the reception, the SETI program is designed to give up and move quickly on to the next job - that's all you're seeing. |
MarkJ Send message Joined: 17 Feb 08 Posts: 1139 Credit: 80,854,192 RAC: 5 |
Well, I've been running 6.08 for a while now, ever since it came out in main. It runs good, no need to abort any "stuck" units, and FINALLY that 8600 on my HTPC/File Server is used for SOMETHING. I gave it a whirl using the 6.08 client and it processed quite happily under BOINC 6.5.0 without any tweaks being required. It would appear to be working now. My two gripes are: 1. The time they take to run, almost the same as AKv8 2. You can't run cpu and gpu ones at the same time (unless you use the Team setup that Raistmer has posted). Supposedly the BOINC dev guys are going to address point 2 in a later release, once they sort out the current version of BOINC with the new work-fetch logic. BOINC blog |
GM3EIY Send message Joined: 1 Nov 08 Posts: 3 Credit: 1,538,756 RAC: 0 |
Thanks Richard, all is clear now Regards Mike GM3EIY/VP8DIA |
koubi Send message Joined: 12 Jan 09 Posts: 2 Credit: 68,316 RAC: 0 |
hello everybody, sorry to my english who is bad (lol) i have a problem with seti ehanced+cuda my computer: amd athlon64 X2 4800+@3GHz os: linux debian 64bit (testing) graphic card: twintech gtx260 sp216 55nm (709/1518/1096) drivers: nvidia 180.22 boinc version: 4.5 cuda:seti@home v6.08 x86_64 all work very well (no error,etc...) but i've a lot of work unit in pending since 4 days.... i would say if it was normal that wu leaves in pending without give credit?? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14653 Credit: 200,643,578 RAC: 874 |
.... Not normal, but not your problem. You cannot earn credit here by yourself. Three things have to happen: 1) You compute the result 2) Another user's computer computes the same result 3) The project (Berkeley) receives both results, and compares them. With the CUDA application (and your very fast graphics card), it is likely that you will finish and return your results before the second computer. While you are waiting for that second result, your results will show as 'pending'. Also, unusually, this weekend the computers at Berkeley which do the comparisons between each pair of results seem to have failed - they stopped working about midnight Saturday UTC, about 38 hours ago. So no (or very few) comparisons since then, and no credit either - everything will be showing as 'pending' since that time, and there's nothing you can do about it. I expect the coputers which make the comparisons - the 'validators' - will start working again soon after the staff open the laboratory for the new working week, in about three hours time. Your backlog of 'pending' work should start to decrease then. Welcome to the boards, and don't worry about your English - your question was perfectly clear. |
perryjay Send message Joined: 20 Aug 02 Posts: 3377 Credit: 20,676,751 RAC: 0 |
Yes, it is a very normal thing for Work Units to stay in pending. It could be a few things, your wingmen are slow turning in work so you have to wait for them to finish or the results sent in don't quite match so they have to send the WU out to another person to try again. Many times it could be on Berkeley's side too. They have problems keeping up with us sometimes. :) PROUD MEMBER OF Team Starfire World BOINC |
Cappy [Team Musketeers] Send message Joined: 19 Feb 03 Posts: 18 Credit: 1,180,143 RAC: 0 |
|
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14653 Credit: 200,643,578 RAC: 874 |
have a bunch piling up trying to upload now :( if this continues ill have to As previously discussed in Panic Mode On - the communications link is saturated. |
perryjay Send message Joined: 20 Aug 02 Posts: 3377 Credit: 20,676,751 RAC: 0 |
It's all part of the game. I've got one stuck waiting to upload, one waiting to report and eleven just reported a few minutes ago. They will work out, it just takes time. PROUD MEMBER OF Team Starfire World BOINC |
koubi Send message Joined: 12 Jan 09 Posts: 2 Credit: 68,316 RAC: 0 |
thanks for yours answers; i've the same problem (wu who don't upload)..i will go to the "panic mode" topic... i've taken 10 days of cache,4 days ago,but my graphic card has finished it ^^ see u later,thanks |
Morten Ross Send message Joined: 30 Apr 01 Posts: 183 Credit: 385,664,915 RAC: 0 |
Every now and then I get a stuck workunit with 6.08. I have a feeling it's a bug in the Boinc Manager however, as restarting the client seems to restart any stuck workunits. Hi, I'd like to know if we're experiencing the same issue: During the "stuck" situation - does the cuda-app (setiathome_6.08_windows_intelx86__cuda.exe) use one cpu 100% (25% on a quad core)? Morten Morten Ross |
Borgholio Send message Joined: 2 Aug 99 Posts: 654 Credit: 18,623,738 RAC: 45 |
Every now and then I get a stuck workunit with 6.08. I have a feeling it's a bug in the Boinc Manager however, as restarting the client seems to restart any stuck workunits. No, my GPU goes down from 100% to 0% and simple idles. My CPU tasks don't change. You will be assimilated...bunghole! |
Woyteck - Boinc Busters Poland Send message Joined: 3 Jun 99 Posts: 49 Credit: 3,203,845 RAC: 0 |
I have noticed that, compared to the last version, the 6.08 does the seti units extremely slow. So not sure what is happening. -- Get up, stand up! Don\'t give up the fight! Credits will make everybody feel high! ;-) |
Morten Ross Send message Joined: 30 Apr 01 Posts: 183 Credit: 385,664,915 RAC: 0 |
Every now and then I get a stuck workunit with 6.08. I have a feeling it's a bug in the Boinc Manager however, as restarting the client seems to restart any stuck workunits. I think you have misunderstood my question - I'm asking about CPU utilization. In windows Task Manager - what is the CPU utilization of "setiathome_6.08_windows_intelx86__cuda.exe" duing the non-progressing/"stuck" Work Unit? You say the CPU tasks don't change - on a quad core you'll have 4 CPUs chrunching (the .exe-files do not have *cuda* in their name), and one feeding the GPU - this is the task of "setiathome_6.08_windows_intelx86__cuda.exe". If your GPU is idle, "setiathome_6.08_windows_intelx86__cuda.exe" is either using exessive CPU or none at all - what is your situation? Morten Morten Ross |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
Every now and then I get a stuck workunit with 6.08. I have a feeling it's a bug in the Boinc Manager however, as restarting the client seems to restart any stuck workunits. Are you sure it's not dropping into CPU fallback mode?, does the task still progress?, slowly?, Claggy |
Borgholio Send message Joined: 2 Aug 99 Posts: 654 Credit: 18,623,738 RAC: 45 |
Every now and then I get a stuck workunit with 6.08. I have a feeling it's a bug in the Boinc Manager however, as restarting the client seems to restart any stuck workunits. I meant that my CPU utilization does not change at all, only my GPU utilization on the specific task that has stalled. You will be assimilated...bunghole! |
©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.