Author |
Message |
|
My computer is not exactly state of the art, but... I often see Boinc just sitting idly with no work to do.
When Boinc connects to get more work, I see something like this:
> Requesting 1728000 seconds of new work, and reporting 12 completed tasks
That means requesting 20 days of new work. That's fine. I set the "Connect to network about every" preference to "10 days" on purpose. I want to have a large cache of work items ready because my computer is sometimes disconnected from the network.
But then it downloads just 16 work items, worth about 10 hours on my computer, and in the project status, it says "Communication deferred 15:28:30" or something such. Basically seems to mean that Boinc will just sit idle for large portions of time.
Sounds like this thing is having problems counting time. All new work items get 1 hour 15 minutes in the "To Completion" column, and the number does seem accurate. And two of them run simultaneously.
From that message saying it'd be getting 20 days worth of work, it should be downloading 768 work items. That's a lot, especially considering that they typically have an expiration time of about 3 days.
Still... the main problem is that it won't connect until way after it's run out of work.
It's definitely not a disk-space problem. I have about 144 GB free.
I'm using version 5.4.9 on XP SP2.
____________
|
|
|
Actually, you know... It is even worse than that. I was browsing the message logs, and this is what I see (excerpt below).
Basically, that time it downloaded 12 work items at 12:14 PM, it finished them in just over 7 hours at 7:22 PM, and then it waited for 8 more hours until I started the next update manually (it still had 15 minutes to go until the next update).
7/20/2006 12:14:40 PM|malariacontrol.net beta|Finished download of file wu_2_30_478_1_224684020
7/20/2006 12:14:40 PM|malariacontrol.net beta|Throughput 40345 bytes/sec
7/20/2006 12:14:40 PM|malariacontrol.net beta|Finished download of file wu_2_28_478_1_249490253
7/20/2006 12:14:40 PM|malariacontrol.net beta|Throughput 25425 bytes/sec
7/20/2006 12:14:40 PM|malariacontrol.net beta|Started download of file wu_2_28_478_3_351416099
7/20/2006 12:14:40 PM|malariacontrol.net beta|Started download of file wu_2_29_478_0_1306021794
7/20/2006 12:14:41 PM||Rescheduling CPU: files downloaded
7/20/2006 12:14:41 PM||Rescheduling CPU: files downloaded
7/20/2006 12:14:42 PM|malariacontrol.net beta|Finished download of file wu_2_28_478_3_351416099
7/20/2006 12:14:42 PM|malariacontrol.net beta|Throughput 25406 bytes/sec
7/20/2006 12:14:42 PM|malariacontrol.net beta|Finished download of file wu_2_29_478_0_1306021794
7/20/2006 12:14:42 PM|malariacontrol.net beta|Throughput 24302 bytes/sec
7/20/2006 12:14:43 PM||Rescheduling CPU: files downloaded
7/20/2006 12:14:43 PM||Rescheduling CPU: files downloaded
7/20/2006 1:26:54 PM||Rescheduling CPU: application exited
7/20/2006 1:26:54 PM|malariacontrol.net beta|Computation for task wu_2_31_478_1_1690337849_2 finished
7/20/2006 1:26:54 PM|malariacontrol.net beta|Starting task wu_2_35_478_7_576311320_0 using malariacontrol version 533
7/20/2006 1:26:56 PM|malariacontrol.net beta|Started upload of file wu_2_31_478_1_1690337849_2_0
7/20/2006 1:26:59 PM|malariacontrol.net beta|Finished upload of file wu_2_31_478_1_1690337849_2_0
7/20/2006 1:26:59 PM|malariacontrol.net beta|Throughput 24608 bytes/sec
7/20/2006 1:49:59 PM||Rescheduling CPU: application exited
7/20/2006 1:49:59 PM|malariacontrol.net beta|Computation for task wu_2_35_478_6_1667203409_2 finished
7/20/2006 1:49:59 PM|malariacontrol.net beta|Starting task wu_2_49_478_0_435607853_0 using malariacontrol version 533
7/20/2006 1:50:01 PM|malariacontrol.net beta|Started upload of file wu_2_35_478_6_1667203409_2_0
7/20/2006 1:50:04 PM|malariacontrol.net beta|Finished upload of file wu_2_35_478_6_1667203409_2_0
7/20/2006 1:50:04 PM|malariacontrol.net beta|Throughput 20633 bytes/sec
7/20/2006 2:59:24 PM||Rescheduling CPU: application exited
7/20/2006 2:59:24 PM|malariacontrol.net beta|Computation for task wu_2_49_478_0_435607853_0 finished
7/20/2006 2:59:24 PM|malariacontrol.net beta|Starting task wu_2_29_478_7_2145253609_2 using malariacontrol version 533
7/20/2006 2:59:26 PM|malariacontrol.net beta|Started upload of file wu_2_49_478_0_435607853_0_0
7/20/2006 2:59:28 PM|malariacontrol.net beta|Finished upload of file wu_2_49_478_0_435607853_0_0
7/20/2006 2:59:28 PM|malariacontrol.net beta|Throughput 10278 bytes/sec
7/20/2006 3:00:49 PM||Rescheduling CPU: application exited
7/20/2006 3:00:49 PM|malariacontrol.net beta|Computation for task wu_2_35_478_7_576311320_0 finished
7/20/2006 3:00:49 PM|malariacontrol.net beta|Starting task wu_2_30_478_0_1542218101_1 using malariacontrol version 533
7/20/2006 3:00:52 PM|malariacontrol.net beta|Started upload of file wu_2_35_478_7_576311320_0_0
7/20/2006 3:00:55 PM|malariacontrol.net beta|Finished upload of file wu_2_35_478_7_576311320_0_0
7/20/2006 3:00:55 PM|malariacontrol.net beta|Throughput 20388 bytes/sec
7/20/2006 4:05:58 PM||Rescheduling CPU: application exited
7/20/2006 4:05:58 PM|malariacontrol.net beta|Computation for task wu_2_29_478_7_2145253609_2 finished
7/20/2006 4:05:58 PM|malariacontrol.net beta|Starting task wu_2_31_478_3_1859231568_0 using malariacontrol version 533
7/20/2006 4:06:00 PM|malariacontrol.net beta|Started upload of file wu_2_29_478_7_2145253609_2_0
7/20/2006 4:06:04 PM|malariacontrol.net beta|Finished upload of file wu_2_29_478_7_2145253609_2_0
7/20/2006 4:06:04 PM|malariacontrol.net beta|Throughput 18317 bytes/sec
7/20/2006 4:07:38 PM||Rescheduling CPU: application exited
7/20/2006 4:07:38 PM|malariacontrol.net beta|Computation for task wu_2_30_478_0_1542218101_1 finished
7/20/2006 4:07:38 PM|malariacontrol.net beta|Starting task wu_2_49_478_1_1716541225_1 using malariacontrol version 533
7/20/2006 4:07:40 PM|malariacontrol.net beta|Started upload of file wu_2_30_478_0_1542218101_1_0
7/20/2006 4:07:43 PM|malariacontrol.net beta|Finished upload of file wu_2_30_478_0_1542218101_1_0
7/20/2006 4:07:43 PM|malariacontrol.net beta|Throughput 18013 bytes/sec
7/20/2006 5:15:51 PM||Rescheduling CPU: application exited
7/20/2006 5:15:51 PM|malariacontrol.net beta|Computation for task wu_2_31_478_3_1859231568_0 finished
7/20/2006 5:15:51 PM|malariacontrol.net beta|Starting task wu_2_30_478_1_224684020_0 using malariacontrol version 533
7/20/2006 5:15:53 PM|malariacontrol.net beta|Started upload of file wu_2_31_478_3_1859231568_0_0
7/20/2006 5:15:57 PM|malariacontrol.net beta|Finished upload of file wu_2_31_478_3_1859231568_0_0
7/20/2006 5:15:57 PM|malariacontrol.net beta|Throughput 24672 bytes/sec
7/20/2006 5:17:47 PM||Rescheduling CPU: application exited
7/20/2006 5:17:47 PM|malariacontrol.net beta|Computation for task wu_2_49_478_1_1716541225_1 finished
7/20/2006 5:17:47 PM|malariacontrol.net beta|Starting task wu_2_28_478_1_249490253_0 using malariacontrol version 533
7/20/2006 5:17:49 PM|malariacontrol.net beta|Started upload of file wu_2_49_478_1_1716541225_1_0
7/20/2006 5:17:52 PM|malariacontrol.net beta|Finished upload of file wu_2_49_478_1_1716541225_1_0
7/20/2006 5:17:52 PM|malariacontrol.net beta|Throughput 10321 bytes/sec
7/20/2006 6:14:21 PM||Rescheduling CPU: application exited
7/20/2006 6:14:21 PM|malariacontrol.net beta|Computation for task wu_2_28_478_1_249490253_0 finished
7/20/2006 6:14:22 PM|malariacontrol.net beta|Starting task wu_2_28_478_3_351416099_1 using malariacontrol version 533
7/20/2006 6:14:24 PM|malariacontrol.net beta|Started upload of file wu_2_28_478_1_249490253_0_0
7/20/2006 6:14:28 PM|malariacontrol.net beta|Finished upload of file wu_2_28_478_1_249490253_0_0
7/20/2006 6:14:28 PM|malariacontrol.net beta|Throughput 18294 bytes/sec
7/20/2006 6:22:04 PM||Rescheduling CPU: application exited
7/20/2006 6:22:04 PM|malariacontrol.net beta|Computation for task wu_2_30_478_1_224684020_0 finished
7/20/2006 6:22:04 PM|malariacontrol.net beta|Starting task wu_2_29_478_0_1306021794_2 using malariacontrol version 533
7/20/2006 6:22:06 PM|malariacontrol.net beta|Started upload of file wu_2_30_478_1_224684020_0_0
7/20/2006 6:22:09 PM|malariacontrol.net beta|Finished upload of file wu_2_30_478_1_224684020_0_0
7/20/2006 6:22:09 PM|malariacontrol.net beta|Throughput 18360 bytes/sec
7/20/2006 7:10:02 PM||Rescheduling CPU: application exited
7/20/2006 7:10:02 PM|malariacontrol.net beta|Computation for task wu_2_28_478_3_351416099_1 finished
7/20/2006 7:10:04 PM|malariacontrol.net beta|Started upload of file wu_2_28_478_3_351416099_1_0
7/20/2006 7:10:07 PM|malariacontrol.net beta|Finished upload of file wu_2_28_478_3_351416099_1_0
7/20/2006 7:10:07 PM|malariacontrol.net beta|Throughput 18011 bytes/sec
7/20/2006 7:22:35 PM||Rescheduling CPU: application exited
7/20/2006 7:22:35 PM|malariacontrol.net beta|Computation for task wu_2_29_478_0_1306021794_2 finished
7/20/2006 7:22:35 PM||Resuming round-robin CPU scheduling.
7/20/2006 7:22:37 PM|malariacontrol.net beta|Started upload of file wu_2_29_478_0_1306021794_2_0
7/20/2006 7:22:41 PM|malariacontrol.net beta|Finished upload of file wu_2_29_478_0_1306021794_2_0
7/20/2006 7:22:41 PM|malariacontrol.net beta|Throughput 18339 bytes/sec
7/21/2006 3:03:28 AM|malariacontrol.net beta|Sending scheduler request to https://malariacontrol.net/malariacontrol_cgi/cgi
7/21/2006 3:03:28 AM|malariacontrol.net beta|Reason: Requested by user
7/21/2006 3:03:28 AM|malariacontrol.net beta|Requesting 1728000 seconds of new work, and reporting 12 completed tasks
7/21/2006 3:03:38 AM|malariacontrol.net beta|Scheduler request succeeded
7/21/2006 3:03:40 AM|malariacontrol.net beta|Started download of file wu_2_31_491_6_919404844
____________
|
|
|
That means requesting 20 days of new work. That's fine. I set the "Connect to network about every" preference to "10 days" on purpose. I want to have a large cache of work items ready because my computer is sometimes disconnected from the network.
Wus last about 3 days. After this period, they are sent to other users. If you connect every 10 days, you will give results too late and not get a lot of credit. You should try to connect more often.
The number of seconds of new work requested seems a bit huge, can you restart the benchmark? Or the duration_correction_factor is wrong...
____________
|
|
|
Wus last about 3 days. After this period, they are sent to other users. If you connect every 10 days, you will give results too late and not get a lot of credit. You should try to connect more often.
Results seem to be uploaded immediately after they complete, so the 10 days wouldn't cause my WUs to be reassigned. (otherwise, I'd suggest changing the maximum in the preferences page)
I can connect often no problem, but sometimes I'll be disconnected for hours at a time, so I just wanted to keep Boinc busy during those times (if I'm using it, I might as well use it all the time, right?)
The number of seconds of new work requested seems a bit huge, can you restart the benchmark? Or the duration_correction_factor is wrong...
You mean shut Boinc down and then restart it? Yes, I'll try that, and I'll fiddle with this preference, see what happens.
I'm just saying there's a bug in the system which can cause users to set preferences that waste computer cycles by allowing the machine to go idle for long periods of time.
Ideally, the time between connections would have nothing to do with the size of the work queue. I'd like to have queued up 24 hours worth of work, but have Boinc connect maybe every hour, report the completed WUs and get a few new ones to keep the queue full.
____________
|
|
|
Rather similar experience with malariacontrol. The reason is not Boinc, but the application, because for example Einstein runs very fine, it is able to use 100% from prosessor power when malariacontrol does not accept it (offering 50/50 for both processes). The reason seems to be, that it notices, that the working application takes 100% when it is loading and does not allow to load more (malariacontrol is able to load only one package at a time and waits after the calculation is ready some hours before it accepts new package) in two processor environment.
Eino
Wus last about 3 days. After this period, they are sent to other users. If you connect every 10 days, you will give results too late and not get a lot of credit. You should try to connect more often.
Results seem to be uploaded immediately after they complete, so the 10 days wouldn\'t cause my WUs to be reassigned. (otherwise, I\'d suggest changing the maximum in the preferences page)
I can connect often no problem, but sometimes I\'ll be disconnected for hours at a time, so I just wanted to keep Boinc busy during those times (if I\'m using it, I might as well use it all the time, right?)
The number of seconds of new work requested seems a bit huge, can you restart the benchmark? Or the duration_correction_factor is wrong...
You mean shut Boinc down and then restart it? Yes, I\'ll try that, and I\'ll fiddle with this preference, see what happens.
I\'m just saying there\'s a bug in the system which can cause users to set preferences that waste computer cycles by allowing the machine to go idle for long periods of time.
Ideally, the time between connections would have nothing to do with the size of the work queue. I\'d like to have queued up 24 hours worth of work, but have Boinc connect maybe every hour, report the completed WUs and get a few new ones to keep the queue full.
|
|
|
You need to set your preferences, the 10 days does not work because the work has a 3 day deadline. What is happening is you are getting less, becasue if the server sent you 10 days worth of work, after 3 days, the other 7 days worth would be past due.
Set the connect interval to a lower number of days, like 1 or even .75 and the additional work to an even lower numebr like .33 for 8 additional hours. This should improve BOINC\'s performance for you.
Computer is connected to the Internet about every
(Leave blank or 0 if always connected.
BOINC will try to maintain at least this much work.) 0.416 days
Maintain enough work for an additional
Enforced by version 5.10+
Notice that in the preferences, it says BOINC will try. There are many factors that limit how much you get, your preferences, task deadline, amount of work already on computer, availablility of work (Sometimes there just is none), and more.
|
|
|
I just signed on to the project and I received more work than I can possibly complete before they are due on 11-9-2009 I set my preferences to keep one day's worth of work & I am always connected to a network. I'm running Climate Prediction & sometimes SETI too. Do I do this right to not get too many WUs?
|
|