| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Least capable machine gets WU, most capable machine doesn't. Why? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 10
|
|
| Author |
|
|
1dark1
Cruncher Joined: Dec 7, 2005 Post Count: 7 Status: Offline Project Badges: 
|
I have two devices currently in production, a two year old dell laptop with an Nvidia 3050 stuffed into it, and a beast of a gaming rig with an Nvidia 4090. The laptop stays flush with WUs; the gaming rig hasn't seen a WU unit in days. why is this? Both devices are using the same preferences, coexist peacefully on the same local network, and both are Windows boxes. When the gaming rig has WUs, I can return hundreds and hundreds of results per day, even if I'm in a marathon gaming session. I'm creeping into the top 20 on my WCG team (slashdot users), and it is beyond frustrating to see my best device starved for work. Here's the last 50 lines (more or less) from the BOINC client event logs on the gaming rig. TBH, this is pretty much what the event log says for the past two days.
12/6/2023 4:19:53 PM | World Community Grid | Scheduler request completed: got 0 new tasks 12/6/2023 4:19:53 PM | World Community Grid | No tasks sent 12/6/2023 4:19:53 PM | World Community Grid | No tasks are available for OpenPandemics - COVID 19 12/6/2023 4:19:53 PM | World Community Grid | No tasks are available for OpenPandemics - COVID-19 - GPU 12/6/2023 4:19:53 PM | World Community Grid | No tasks are available for Africa Rainfall Project 12/6/2023 4:19:53 PM | World Community Grid | No tasks are available for Help Stop TB 12/6/2023 4:19:53 PM | World Community Grid | No tasks are available for Mapping Cancer Markers 12/6/2023 4:19:53 PM | World Community Grid | No tasks are available for Smash Childhood Cancer 12/6/2023 4:19:55 PM | World Community Grid | Project requested delay of 121 seconds 12/6/2023 4:22:01 PM | World Community Grid | Requesting new tasks for CPU and NVIDIA GPU and NVIDIA GeForce RTX 4090 12/6/2023 4:22:03 PM | World Community Grid | Scheduler request completed: got 0 new tasks 12/6/2023 4:22:03 PM | World Community Grid | No tasks sent 12/6/2023 4:22:03 PM | World Community Grid | No tasks are available for OpenPandemics - COVID 19 12/6/2023 4:22:03 PM | World Community Grid | No tasks are available for OpenPandemics - COVID-19 - GPU 12/6/2023 4:22:03 PM | World Community Grid | No tasks are available for Africa Rainfall Project 12/6/2023 4:22:03 PM | World Community Grid | No tasks are available for Help Stop TB 12/6/2023 4:22:03 PM | World Community Grid | No tasks are available for Mapping Cancer Markers 12/6/2023 4:22:03 PM | World Community Grid | No tasks are available for Smash Childhood Cancer 12/6/2023 4:22:03 PM | World Community Grid | Project requested delay of 121 seconds 12/6/2023 4:24:07 PM | World Community Grid | Requesting new tasks for CPU and NVIDIA GPU and NVIDIA GeForce RTX 4090 12/6/2023 4:24:09 PM | World Community Grid | Scheduler request completed: got 0 new tasks 12/6/2023 4:24:09 PM | World Community Grid | No tasks sent 12/6/2023 4:24:09 PM | World Community Grid | No tasks are available for OpenPandemics - COVID 19 12/6/2023 4:24:09 PM | World Community Grid | No tasks are available for OpenPandemics - COVID-19 - GPU 12/6/2023 4:24:09 PM | World Community Grid | No tasks are available for Africa Rainfall Project 12/6/2023 4:24:09 PM | World Community Grid | No tasks are available for Help Stop TB 12/6/2023 4:24:09 PM | World Community Grid | No tasks are available for Mapping Cancer Markers 12/6/2023 4:24:09 PM | World Community Grid | No tasks are available for Smash Childhood Cancer 12/6/2023 4:24:09 PM | World Community Grid | Project requested delay of 121 seconds 12/6/2023 4:26:11 PM | World Community Grid | Requesting new tasks for CPU and NVIDIA GPU and NVIDIA GeForce RTX 4090 12/6/2023 4:26:13 PM | World Community Grid | Scheduler request completed: got 0 new tasks 12/6/2023 4:26:13 PM | World Community Grid | No tasks sent 12/6/2023 4:26:13 PM | World Community Grid | No tasks are available for OpenPandemics - COVID 19 12/6/2023 4:26:13 PM | World Community Grid | No tasks are available for OpenPandemics - COVID-19 - GPU 12/6/2023 4:26:13 PM | World Community Grid | No tasks are available for Africa Rainfall Project 12/6/2023 4:26:13 PM | World Community Grid | No tasks are available for Help Stop TB 12/6/2023 4:26:13 PM | World Community Grid | No tasks are available for Mapping Cancer Markers 12/6/2023 4:26:13 PM | World Community Grid | No tasks are available for Smash Childhood Cancer 12/6/2023 4:26:13 PM | World Community Grid | Project requested delay of 121 seconds 12/6/2023 4:28:19 PM | World Community Grid | Sending scheduler request: To fetch work. 12/6/2023 4:28:19 PM | World Community Grid | Requesting new tasks for CPU and NVIDIA GPU and NVIDIA GeForce RTX 4090 12/6/2023 4:28:22 PM | World Community Grid | Scheduler request completed: got 0 new tasks 12/6/2023 4:28:22 PM | World Community Grid | No tasks sent 12/6/2023 4:28:22 PM | World Community Grid | No tasks are available for OpenPandemics - COVID 19 12/6/2023 4:28:22 PM | World Community Grid | No tasks are available for OpenPandemics - COVID-19 - GPU 12/6/2023 4:28:22 PM | World Community Grid | No tasks are available for Africa Rainfall Project 12/6/2023 4:28:22 PM | World Community Grid | No tasks are available for Help Stop TB 12/6/2023 4:28:22 PM | World Community Grid | No tasks are available for Mapping Cancer Markers 12/6/2023 4:28:22 PM | World Community Grid | No tasks are available for Smash Childhood Cancer If I am interpreting the log correctly, the last wu it uploaded would be this entry: 12/4/2023 10:50:13 PM | World Community Grid | Finished upload of MCM1_0208919_2825_0_r1364957339_0 12/4/2023 10:50:21 PM | World Community Grid | Sending scheduler request: To fetch work. 12/4/2023 10:50:21 PM | World Community Grid | Reporting 7 completed tasks 12/4/2023 10:50:21 PM | World Community Grid | Requesting new tasks for CPU and NVIDIA GPU and NVIDIA GeForce RTX 4090 12/4/2023 10:50:23 PM | World Community Grid | Scheduler request completed: got 0 new tasks 12/4/2023 10:50:23 PM | World Community Grid | No tasks sent 12/4/2023 10:50:23 PM | World Community Grid | No tasks are available for OpenPandemics - COVID 19 12/4/2023 10:50:23 PM | World Community Grid | No tasks are available for OpenPandemics - COVID-19 - GPU 12/4/2023 10:50:23 PM | World Community Grid | No tasks are available for Africa Rainfall Project 12/4/2023 10:50:23 PM | World Community Grid | No tasks are available for Help Stop TB 12/4/2023 10:50:23 PM | World Community Grid | No tasks are available for Mapping Cancer Markers 12/4/2023 10:50:23 PM | World Community Grid | No tasks are available for Smash Childhood Cancer 12/4/2023 10:50:23 PM | World Community Grid | Project requested delay of 121 seconds Is there some reason why my most capable machine is not getting WUs, while my least capable machine is? |
||
|
|
thunder7
Senior Cruncher Netherlands Joined: Mar 6, 2013 Post Count: 238 Status: Offline Project Badges: 
|
I see that too. My 40 core and 48 core machines have a full queue (1000 WU's). The 88 core machine is limping along, never having more than 100 WU's in the queue, often as low as 30. I need to babysit it pretty carefully, updating the project or detaching/attaching it multiple times a week to make sure it gets enough. Tiresome to say the least.
Luckily, I run on linux and can automate the 'get me extra tasks please' command to a certain extent, but it doesn't help 100%. It would be nice to know why powerful machines don't get enough WU's. |
||
|
|
TonyEllis
Senior Cruncher Australia Joined: Jul 9, 2008 Post Count: 286 Status: Offline Project Badges: 
|
See this type of pattern as well with the systems here considered "reliable". Urgent "_2" or greater suffix WUs are more likly to be sent to the slower machines than the faster.
----------------------------------------
Run Time Stats https://grassmere-productions.no-ip.biz/
|
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7846 Status: Offline Project Badges: 
|
That is puzzling to say the least. I am wondering if there is a scheduling algorithm which looks at the open cache size and fills machines requesting small numbers of work units first and then filling machines requesting larger numbers of work units later. Thus if there is a large number of machines with small requests this may exhaust the available number of work units in the scheduler before the requests for the bigger requests are polled. This would have the effect of keeping the all the smaller machines filled, but would short the bigger machines. There may also be some capacity restraints on the
----------------------------------------If this is the case, it would involve some modification of the system from the BOINC administrators. It may be possible to reduce the cache size of the machine so it is requesting work units in smaller increments, but right now that is just a theory. Cheers
Sgt. Joe
----------------------------------------*Minnesota Crunchers* [Edit 1 times, last edit by Sgt.Joe at Dec 8, 2023 12:29:18 AM] |
||
|
|
thunder7
Senior Cruncher Netherlands Joined: Mar 6, 2013 Post Count: 238 Status: Offline Project Badges:
|
60639: 08-Dec-2023 11:34:58 (low) [World Community Grid] Scheduler request completed: got 0 new tasks
----------------------------------------60654: 08-Dec-2023 11:37:04 (low) [World Community Grid] Reporting 1 completed tasks 60656: 08-Dec-2023 11:37:07 (low) [World Community Grid] Scheduler request completed: got 2 new tasks 60667: 08-Dec-2023 11:39:14 (low) [World Community Grid] Reporting 1 completed tasks 60669: 08-Dec-2023 11:39:16 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60681: 08-Dec-2023 11:51:29 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60693: 08-Dec-2023 12:01:42 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60705: 08-Dec-2023 12:27:02 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60720: 08-Dec-2023 13:19:22 (low) [World Community Grid] Reporting 1 completed tasks 60722: 08-Dec-2023 13:19:24 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60737: 08-Dec-2023 13:22:56 (low) [World Community Grid] Reporting 1 completed tasks 60739: 08-Dec-2023 13:22:58 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60766: 08-Dec-2023 13:25:04 (low) [World Community Grid] Reporting 5 completed tasks 60768: 08-Dec-2023 13:25:07 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60791: 08-Dec-2023 13:27:09 (low) [World Community Grid] Reporting 4 completed tasks 60793: 08-Dec-2023 13:27:14 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60828: 08-Dec-2023 13:29:15 (low) [World Community Grid] Reporting 8 completed tasks 60830: 08-Dec-2023 13:29:19 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60862: 08-Dec-2023 13:31:21 (low) [World Community Grid] Reporting 7 completed tasks 60864: 08-Dec-2023 13:31:30 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60883: 08-Dec-2023 13:33:31 (low) [World Community Grid] Reporting 5 completed tasks 60885: 08-Dec-2023 13:33:33 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60915: 08-Dec-2023 13:35:39 (low) [World Community Grid] Reporting 6 completed tasks 60919: 08-Dec-2023 13:35:42 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60942: 08-Dec-2023 13:37:49 (low) [World Community Grid] Reporting 5 completed tasks 60944: 08-Dec-2023 13:37:52 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60975: 08-Dec-2023 13:39:58 (low) [World Community Grid] Reporting 6 completed tasks 60978: 08-Dec-2023 13:40:00 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 60995: 08-Dec-2023 13:42:06 (low) [World Community Grid] Reporting 3 completed tasks 60997: 08-Dec-2023 13:42:08 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61017: 08-Dec-2023 13:44:14 (low) [World Community Grid] Reporting 3 completed tasks 61019: 08-Dec-2023 13:44:17 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61039: 08-Dec-2023 13:46:23 (low) [World Community Grid] Reporting 3 completed tasks 61041: 08-Dec-2023 13:46:28 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61082: 08-Dec-2023 13:48:30 (low) [World Community Grid] Reporting 10 completed tasks 61084: 08-Dec-2023 13:48:33 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61102: 08-Dec-2023 13:50:34 (low) [World Community Grid] Reporting 2 completed tasks 61106: 08-Dec-2023 13:50:36 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61120: 08-Dec-2023 14:01:57 (low) [World Community Grid] Reporting 2 completed tasks 61122: 08-Dec-2023 14:01:59 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61136: 08-Dec-2023 14:04:05 (low) [World Community Grid] Reporting 1 completed tasks 61138: 08-Dec-2023 14:04:07 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61158: 08-Dec-2023 14:06:09 (low) [World Community Grid] Reporting 3 completed tasks 61160: 08-Dec-2023 14:06:14 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61193: 08-Dec-2023 14:08:15 (low) [World Community Grid] Reporting 7 completed tasks 61198: 08-Dec-2023 14:08:19 (low) [World Community Grid] Scheduler request completed: got 14 new tasks 61251: 08-Dec-2023 14:10:20 (low) [World Community Grid] Reporting 4 completed tasks 61253: 08-Dec-2023 14:10:24 (low) [World Community Grid] Scheduler request completed: got 26 new tasks 61340: 08-Dec-2023 14:12:26 (low) [World Community Grid] Reporting 2 completed tasks 61342: 08-Dec-2023 14:12:30 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61354: 08-Dec-2023 14:23:43 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61366: 08-Dec-2023 14:37:57 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61380: 08-Dec-2023 15:06:11 (low) [World Community Grid] Reporting 1 completed tasks 61382: 08-Dec-2023 15:06:13 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61399: 08-Dec-2023 15:08:19 (low) [World Community Grid] Reporting 2 completed tasks 61401: 08-Dec-2023 15:08:22 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61413: 08-Dec-2023 15:23:37 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61425: 08-Dec-2023 15:38:52 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61441: 08-Dec-2023 15:48:52 (low) [World Community Grid] Reporting 1 completed tasks 61444: 08-Dec-2023 15:48:53 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61496: 08-Dec-2023 15:50:55 (low) [World Community Grid] Reporting 14 completed tasks 61500: 08-Dec-2023 15:50:57 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61582: 08-Dec-2023 15:52:58 (low) [World Community Grid] Reporting 25 completed tasks 61584: 08-Dec-2023 15:53:04 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61596: 08-Dec-2023 15:59:13 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61608: 08-Dec-2023 16:23:34 (low) [World Community Grid] Scheduler request completed: got 0 new tasks 61620: 08-Dec-2023 17:05:04 (low) [World Community Grid] Scheduler request completed: got 0 new tasks all of which means, it wasn´t running anything when I returned home again. Time to detach and attach again :-( I´m debating what would happen if I ran 11 docker images, each running 8 threads. [Edit 1 times, last edit by thunder7 at Dec 8, 2023 4:16:06 PM] |
||
|
|
bfmorse
Senior Cruncher US Joined: Jul 26, 2009 Post Count: 442 Status: Offline Project Badges: 
|
I seem to recall someone mentioning that a single site swallowed several hundred WU's in one sitting - leaving the rest of the volunteers with nothing.
Unfortunately I do not recall the thread nor precisely how long ago it happened. That event, most likely, was the catalyst for the current apparent restrictions on Work Unit release. It also does not help when volunteers, not wanting to run out of work, have queues in excess of five (5) days. As we know in most cases, after exactly six (6.0) days from the initial WU release, the system will automatically send out a "resend" on that WU: For. Each. And. Every. One. And, the timeout for resends is three (3.0) days. And I seem to have a magnetism for those resends. [off topic - as of about a minute ago, attempts to edit and save Profiles still crash] |
||
|
|
adriverhoef
Master Cruncher The Netherlands Joined: Apr 3, 2009 Post Count: 2346 Status: Offline Project Badges:
|
That event, most likely, was the catalyst for the current apparent restrictions on Work Unit release. bfmorse, do you perhaps mean the Linux 5.15.107+ cluster that scooped up thousands and thousands of tasks? (see post 689642)Adri |
||
|
|
thunder7
Senior Cruncher Netherlands Joined: Mar 6, 2013 Post Count: 238 Status: Offline Project Badges:
|
The current WU's only taking 95 minutes of computing time compared to 3 hours earlier doesn´t help any, BTW.
|
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7846 Status: Offline Project Badges:
|
That event, most likely, was the catalyst for the current apparent restrictions on Work Unit release. bfmorse, do you perhaps mean the Linux 5.15.107+ cluster that scooped up thousands and thousands of tasks? (see post 689642)Adri Makes me wonder what happened to that site. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
bfmorse
Senior Cruncher US Joined: Jul 26, 2009 Post Count: 442 Status: Offline Project Badges:
|
Adri,
Yes. That was the one. |
||
|
|
|