Message boards : BOINC Manager : Sceduling system
Message board moderation
Author | Message |
---|---|
Send message Joined: 28 May 10 Posts: 52 |
Hi, not long ago I joined Asteroids@home. After running several wu's on different systems and waited for validation I decided to run this app on three systems. Two systems have included this app into the regular sceduling, the third does not load these apps. When I da a manual update, the message says 10.05.2014 15:44:51 | Asteroids@home | Not requesting tasks: don't need (CPU: buffer full; NVIDIA GPU: project preferences; AMD/ATI GPU: no applications; Intel GPU: no applications) I see this message for the third day now. When I put the other two projects I run on this pc on hold, I get Asteroid wu's. Resource share for the three projects is equal. This leads to my question: is it possible to visualize the sceduling system? Alexander |
Send message Joined: 5 Oct 06 Posts: 5129 |
It is possible to set a verbose logging option: <work_fetch_debug> but be warned that it generates a great deal of highly technical output. For details see Client configuration. |
Send message Joined: 28 May 10 Posts: 52 |
THX for the answer, tag is insertet to cc_config.xml and works. The thing is: Asteroid project preferences are set to run cpu-apps only. The project has cpu and nVidia apps. This setting works on the following pc's: https://asteroidsathome.net/boinc//show_host_detail.php?hostid=88982 https://asteroidsathome.net/boinc//show_host_detail.php?hostid=89536 https://asteroidsathome.net/boinc//show_host_detail.php?hostid=88967 it does not work here: https://asteroidsathome.net/boinc//show_host_detail.php?hostid=88967 The difference is, this system has all three types of gpu's Messages show: 12.05.2014 20:51:51 | Asteroids@home | [work_fetch] fetch share 0.000 (no apps) 12.05.2014 20:51:51 | | [work_fetch] ------- end work fetch state ------- 12.05.2014 20:51:51 | | [work_fetch] No project chosen for work fetch 12.05.2014 20:51:51 | Asteroids@home | [work_fetch] fetch share 0.000 (no apps) 12.05.2014 20:51:51 | | [work_fetch] --- state for Intel GPU --- 12.05.2014 20:51:51 | | [work_fetch] shortfall 864.00 nidle 1.00 saturated 0.00 busy 0.00 12.05.2014 20:51:51 | Asteroids@home | [work_fetch] fetch share 0.000 (blocked by prefs) 12.05.2014 20:51:51 | | [work_fetch] --- state for AMD/ATI GPU --- 12.05.2014 20:51:51 | | [work_fetch] shortfall 864.00 nidle 1.00 saturated 0.00 busy 0.00 12.05.2014 20:52:52 | Asteroids@home | [work_fetch] fetch share 0.258 12.05.2014 20:52:52 | | [work_fetch] --- state for NVIDIA GPU --- 12.05.2014 20:52:52 | | [work_fetch] shortfall 864.00 nidle 1.00 saturated 0.00 busy 0.00 12.05.2014 20:57:07 | Asteroids@home | [work_fetch] REC 2428.924 prio -1.603775 can req work 12.05.2014 20:57:07 | | [work_fetch] --- state for CPU --- 12.05.2014 20:57:07 | | [work_fetch] shortfall 0.00 nidle 0.00 saturated 984.67 busy 0.00 12.05.2014 21:10:00 | Asteroids@home | [work_fetch] REC 2427.420 prio -2.031742 can req work 12.05.2014 21:10:00 | | [work_fetch] --- state for CPU --- 12.05.2014 21:10:00 | | [work_fetch] shortfall 0.00 nidle 0.00 saturated 1325.20 busy 0.00 Resource share is set to equal values on all 4 running projects. I have no experiance in interpreting these messages, but why is something saturated whatfor no apps exist? BM 7.3.11 win7x64 i7-3770, GTX650ti, HD7870(Tahiti) Alexander |
Send message Joined: 5 Oct 06 Posts: 5129 |
Might be easier to refer to one of mine, with one complete 'cycle' of the log - the sequence in which the lines are shown is relevant. 13/05/2014 15:23:19 | | [work_fetch] ------- start work fetch state ------- 13/05/2014 15:23:19 | | [work_fetch] target work buffer: 25920.00 + 3456.00 sec 13/05/2014 15:23:19 | | [work_fetch] --- project states --- 13/05/2014 15:23:19 | Albert@Home | [work_fetch] REC 0.000 prio -0.000000 can't req work: "no new tasks" requested via Manager 13/05/2014 15:23:19 | Test4Theory@Home | [work_fetch] REC 0.000 prio -0.000000 can't req work: "no new tasks" requested via Manager 13/05/2014 15:23:19 | LHC@home 1.0 | [work_fetch] REC 0.000 prio -0.000000 can't req work: "no new tasks" requested via Manager 13/05/2014 15:23:19 | Milkyway@Home | [work_fetch] REC 0.000 prio -0.000000 can't req work: "no new tasks" requested via Manager 13/05/2014 15:23:19 | SETI@home Beta Test | [work_fetch] REC 0.159 prio -0.000000 can't req work: "no new tasks" requested via Manager 13/05/2014 15:23:19 | boincsimap | [work_fetch] REC 2344.479 prio -0.036650 can req work 13/05/2014 15:23:19 | NumberFields@home | [work_fetch] REC 2319.814 prio -0.037250 can req work 13/05/2014 15:23:19 | Einstein@Home | [work_fetch] REC 2263.506 prio -0.038438 can req work 13/05/2014 15:23:19 | GPUGRID | [work_fetch] REC 233705.920 prio -1.750106 can req work 13/05/2014 15:23:19 | SETI@home | [work_fetch] REC 241905.082 prio -1.775359 can't req work: some task is suspended via Manager 13/05/2014 15:23:19 | | [work_fetch] --- state for CPU --- 13/05/2014 15:23:19 | | [work_fetch] shortfall 6613.57 nidle 0.00 saturated 28004.59 busy 0.00 13/05/2014 15:23:19 | Albert@Home | [work_fetch] fetch share 0.000 (blocked by prefs) 13/05/2014 15:23:19 | Test4Theory@Home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | LHC@home 1.0 | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | Milkyway@Home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | SETI@home Beta Test | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | boincsimap | [work_fetch] fetch share 0.500 13/05/2014 15:23:19 | NumberFields@home | [work_fetch] fetch share 0.500 13/05/2014 15:23:19 | Einstein@Home | [work_fetch] fetch share 0.000 (blocked by prefs) 13/05/2014 15:23:19 | GPUGRID | [work_fetch] fetch share 0.000 (blocked by prefs) 13/05/2014 15:23:19 | SETI@home | [work_fetch] fetch share 0.000 (blocked by prefs) (no apps) 13/05/2014 15:23:19 | | [work_fetch] --- state for NVIDIA --- 13/05/2014 15:23:19 | | [work_fetch] shortfall 32410.61 nidle 0.00 saturated 2211.14 busy 0.00 13/05/2014 15:23:19 | Albert@Home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | Test4Theory@Home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | LHC@home 1.0 | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | Milkyway@Home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | SETI@home Beta Test | [work_fetch] fetch share 0.000 (blocked by prefs) 13/05/2014 15:23:19 | boincsimap | [work_fetch] fetch share 0.000 (no apps) 13/05/2014 15:23:19 | NumberFields@home | [work_fetch] fetch share 0.000 (no apps) 13/05/2014 15:23:19 | Einstein@Home | [work_fetch] fetch share 0.000 (blocked by prefs) 13/05/2014 15:23:19 | GPUGRID | [work_fetch] fetch share 1.000 13/05/2014 15:23:19 | SETI@home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | | [work_fetch] --- state for intel_gpu --- 13/05/2014 15:23:19 | | [work_fetch] shortfall 509.64 nidle 0.00 saturated 28866.36 busy 0.00 13/05/2014 15:23:19 | Albert@Home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | Test4Theory@Home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | LHC@home 1.0 | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | Milkyway@Home | [work_fetch] fetch share 0.000 13/05/2014 15:23:19 | SETI@home Beta Test | [work_fetch] fetch share 0.000 (blocked by prefs) 13/05/2014 15:23:19 | boincsimap | [work_fetch] fetch share 0.000 (no apps) 13/05/2014 15:23:19 | NumberFields@home | [work_fetch] fetch share 0.000 (no apps) 13/05/2014 15:23:19 | Einstein@Home | [work_fetch] fetch share 0.333 13/05/2014 15:23:19 | GPUGRID | [work_fetch] fetch share 0.667 13/05/2014 15:23:19 | SETI@home | [work_fetch] fetch share 0.000 (no apps) 13/05/2014 15:23:19 | | [work_fetch] ------- end work fetch state ------- Points to watch for: target work buffer: 25920.00 Work fetch will be triggered when saturation (we'll come to this later) falls below this value. Until then, nothing. --- project states --- Each attached project follows. You have Asteroids@home | can req work, which is all you need at this point. --- state for CPU --- (And similarly for AMD/ATI GPU, NVIDIA GPU, intel_gpu etc.) Everything from here on relates to that named resource, until the next 'state for' line. From your first message, you have "don't need (NVIDIA GPU: project preferences; AMD/ATI GPU: no applications; Intel GPU: no applications)", and those reasons are restated in the resource blocks. So, for CPU? Read it left to right: shortfall 0.00 nidle 0.00 saturated 984.67 busy 0.00 That means your CPU(s) have enough work on hand (from all attached projects, combined) to keep them busy for 984 seconds - a bit over a quarter of an hour. That's the number which is compared with 'target work buffer' at the top - and yours must be even smaller. If 'saturated' was smaller than 'target', there would be a shortfall figure greater than zero. Also, we see that your CPU(s) is/are active - "n(umber) idle zero". Work fetch should occur - assuming 'can req work' still applies - when either 'saturated' falls below 'target', or there's an idle CPU. When work is needed, the project with highest priority (negative numbers, so 'highest' means closest to zero) will be contacted first - assuming it's not blocked for some reason. The project in the first block are listed in priority order, so it's easy to see in my list that boincsimap will be the next to be contacted with a request for CPU work. Clear as mud? |
Send message Joined: 28 May 10 Posts: 52 |
First of all - thank you for taking so much time, it's realy appreciated! With help of your explanations I copied one work fetch cycle into an editor and removed the lines pointing to project that are not part of the game right now. Then I removed the lines pointing to gpu's - not part of the game. What's left over: 13.05.2014 20:33:01 | | [work_fetch] ------- start work fetch state ------- 13.05.2014 20:33:01 | | [work_fetch] target work buffer: 8640.00 + 864.00 sec 13.05.2014 20:33:01 | | [work_fetch] --- project states --- 13.05.2014 20:33:01 | WUProp@Home | [work_fetch] REC 0.007 prio -0.002937 can't req work: non CPU intensive 13.05.2014 20:33:01 | NRG | [work_fetch] REC 265.556 prio -0.245005 can req work 13.05.2014 20:33:01 | boincsimap | [work_fetch] REC 980.671 prio -0.895267 can req work 13.05.2014 20:33:01 | pogs | [work_fetch] REC 1044.700 prio -0.895423 can req work 13.05.2014 20:33:01 | Asteroids@home | [work_fetch] REC 2268.893 prio -1.928137 can req work 13.05.2014 20:33:01 | | [work_fetch] --- state for CPU --- 13.05.2014 20:33:01 | | [work_fetch] shortfall 877.89 nidle 0.00 saturated 8993.55 busy 0.00 13.05.2014 20:33:01 | NRG | [work_fetch] fetch share 0.242 13.05.2014 20:33:01 | boincsimap | [work_fetch] fetch share 0.242 13.05.2014 20:33:01 | pogs | [work_fetch] fetch share 0.258 13.05.2014 20:33:01 | Asteroids@home | [work_fetch] fetch share 0.258 13.05.2014 20:33:01 | | [work_fetch] ------- end work fetch state ------- What I see is that Asteroids is not the project with the highest priority also no wu was crunched for a couple of days. Only pogs, simap and nrg are loaded. How does it work to get higher priority? Just for clarification: saturated is in relationship with the local settings of work buffer and additional work buffer? I've set both to 0.01 days. One question is left over: whatfor stands REC ? Seems to be in relation with expected wu run time. Alexander |
Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License,
Version 1.2 or any later version published by the Free Software Foundation.