SETI Schedular Problems?


log in

Advanced search

Message boards : Number crunching : SETI Schedular Problems?

Author Message
strue87
Send message
Joined: 20 Mar 13
Posts: 13
Credit: 476,694
RAC: 0
United States
Message 1350943 - Posted: 26 Mar 2013, 20:37:14 UTC

Hi, I just wanted to ask if there is a problem with the scheduler. I have been having a heck of a time getting any work at all for my computer. I am running BOINC Version 7.0.58 and have a Core-I7-3630QM CPU with an Nvidia GTX 660M Video Card.

For the last 2 days everytime I check my computer is only processing 3 wu's at the max because there is no work for it to do. I just checked my results page and it said I had 19 wu's that I was working on which is what I show also but 17 of those are CUDA_Fermi wu's that only the GPU can work on.

I know the Astropulse V6 6.01 wu's are saying that the estimated time is like 228 hours and I'm going through them in about 9 or 10 hours. I'm guessing once I get through enough of them that it will correct itself but since it's only giving me 2 at a time to work on with my 8 CPU cores available that it's gonna take awhile (my other 6 cores have just been idle the last couple of days because I have no work for them to do).

I have been running the UPDATE command several times a day to try and get the schedular to give me some work and sometimes it does but only about 4 wu's and their usually CUDA wu's which my GPU is taking about 40 minutes to process.

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4087
Credit: 33,003,041
RAC: 5,808
United Kingdom
Message 1350949 - Posted: 26 Mar 2013, 20:51:29 UTC - in response to Message 1350943.
Last modified: 26 Mar 2013, 20:56:52 UTC

What are your cache settings?

According to your All tasks for computer 6946525 page you have 43 tasks, 8 of them are for the CPU, at moment your host hasn't completed it's 10 validations for Astropulse v6,
so those will be estimated to take a huge amount of time, once those 10 validations are done the Astropulse estimates will scale to something like actual.

Claggy

strue87
Send message
Joined: 20 Mar 13
Posts: 13
Credit: 476,694
RAC: 0
United States
Message 1350960 - Posted: 26 Mar 2013, 21:25:07 UTC - in response to Message 1350949.

What are your cache settings?

According to your All tasks for computer 6946525 page you have 43 tasks, 8 of them are for the CPU, at moment your host hasn't completed it's 10 validations for Astropulse v6,
so those will be estimated to take a huge amount of time, once those 10 validations are done the Astropulse estimates will scale to something like actual.

Claggy


Hmm, not sure what my cache settings are? I'm still a little new to all this but I set my BOINC settings for the maximum 10 days of work and my prefreneces are set to:


Suspend work while computer is on battery power?
Matters only for portable computers

yes

Suspend work while computer is in use?

no

Suspend GPU work while computer is in use?
Enforced by version 6.6.21+

no

'In use' means mouse/keyboard activity in last

3 minutes

Suspend work if no mouse/keyboard activity in last
Needed to enter low-power mode on some computers

--- minutes

Suspend work when non-BOINC CPU usage is above
0 means no restriction
Enforced by version 6.10.30+

50%

Do work only between the hours of
No restriction if equal

---

Leave tasks in memory while suspended?
Suspended tasks will consume swap space if 'yes'

yes

Switch between tasks every
Recommended: 60 minutes

120 minutes

On multiprocessors, use at most

16 processors

On multiprocessors, use at most
Enforced by version 6.1+

100% of the processors

Use at most
Can be used to reduce CPU heat

100% of CPU time

Disk and memory usage
Disk: use at most

10 GB

Disk: leave free at least
Values smaller than 0.001 are ignored

0.5 GB

Disk: use at most

50% of total

Tasks checkpoint to disk at most every

60 seconds

Swap space: use at most

50% of total

Memory: when computer is in use, use at most

50% of total

Memory: when computer is not in use, use at most

75% of total

Network usage
Maintain enough tasks to keep busy for at least
(max 10 days).

10 days

... and up to an additional

2 days

Confirm before connecting to Internet?
Matters only if you have a modem, ISDN or VPN connection

no

Disconnect when done?
Matters only if you have a modem, ISDN or VPN connection

no

Maximum download rate:

--- Kbytes/sec

Maximum upload rate:

--- Kbytes/sec

Use network only between the hours of

---

Transfer at most
Enforced by version 6.10.46+

--- Mbytes every --- days

Skip image file verification?
Check this ONLY if your Internet provider modifies image files (UMTS does this, for example). Skipping verification reduces the security of BOINC.

no

I'm not saying I'm getting no work at all but like currently I'm doing pretty good I have 6 of my 8 CPU's doing work 2 Astropulse, 4 Setiathome enhanced 6.03. The problem is the scheduler wont give me any to cache, and not even enough to keep all my cores busy. The Astropulses will take about 10 hours and the 6.03 wu's take about 40 minutes to process and I have none on hand for even my 2 core's that are idle.

Horacio
Send message
Joined: 14 Jan 00
Posts: 536
Credit: 73,404,716
RAC: 90,359
Argentina
Message 1350964 - Posted: 26 Mar 2013, 21:43:01 UTC

Network usage
Maintain enough tasks to keep busy for at least
(max 10 days).

10 days

... and up to an additional

2 days


This seems to be the issue, with 10+2 days of cache added to the long estimation of the tasks the scheduller will only give you tasks with deadlines longer than 12 days + the currently estimated times... But as the deadlines are ussually around 14 days, the scheduller will not send anything thinking that you wont be able to make it in time.

Try to set this settings to 3 days + 1 day (or less) and you may get better changes to get more work assigned.
____________

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4087
Credit: 33,003,041
RAC: 5,808
United Kingdom
Message 1350965 - Posted: 26 Mar 2013, 21:48:04 UTC - in response to Message 1350960.
Last modified: 26 Mar 2013, 21:51:26 UTC

What are your cache settings?

According to your All tasks for computer 6946525 page you have 43 tasks, 8 of them are for the CPU, at moment your host hasn't completed it's 10 validations for Astropulse v6,
so those will be estimated to take a huge amount of time, once those 10 validations are done the Astropulse estimates will scale to something like actual.

Claggy


Hmm, not sure what my cache settings are? I'm still a little new to all this but I set my BOINC settings for the maximum 10 days of work and my prefreneces are set to:

Network usage
Maintain enough tasks to keep busy for at least
(max 10 days).

10 days

... and up to an additional

2 days

I'm not saying I'm getting no work at all but like currently I'm doing pretty good I have 6 of my 8 CPU's doing work 2 Astropulse, 4 Setiathome enhanced 6.03. The problem is the scheduler wont give me any to cache, and not even enough to keep all my cores busy. The Astropulses will take about 10 hours and the 6.03 wu's take about 40 minutes to process and I have none on hand for even my 2 core's that are idle.


Having the cache set to 10 + 2 days is counter productive, Boinc 7 uses those two settings as High and Low Water marks, it'll try and keep 12 days cache, then every two days, once the amount of work drops below 10 days then it'll ask for work again,
The first setting is also used to say how many days offline you're going to be, so Boinc will try and do all the work 10 days before deadline, Probably the scheduler doesn't think you can finish any Shortie work in time, so it refuses to send any,
Best to set your Cache to something more reasonable, Like 4 + 0.01 With that Boinc will cache four point zero one days work, and very time it drops below four days it'll ask for more.
(with the present limits of 100 CPU and 100 GPU tasks you should easily get that with those settings, any higher setting is point less)

Claggy

strue87
Send message
Joined: 20 Mar 13
Posts: 13
Credit: 476,694
RAC: 0
United States
Message 1350968 - Posted: 26 Mar 2013, 21:53:54 UTC - in response to Message 1350965.


Having the cache set to 10 + 2 days is counter productive, Boinc 7 uses those two settings as High and Low Water marks, it'll try and keep 12 days cache, then every two days, once the amount of work drops below 10 days then it'll ask for work again,
The first setting is also used to say how many days offline you're going to be, so Boinc will try and do all the work 10 days before deadline, Probably the scheduler doesn't think you can finish any Shortie work in time, so it refuses to send any,
Best to set your Cache to something more reasonable, Like 4 + 0.01 With that Boinc will cache four point zero one days work, and very time it drops below four days it'll ask for more.
(with the present limits of 100 CPU and 100 GPU tasks you should easily get that with those settings, any higher setting is point less)

Claggy


OK thanks for the information I will try changing the settings to what you said. I maxed out the setting thinking it would give me my 100 CPU and 100 GPU tasks to work on during the outage but now all I get is dribbles of work.

strue87
Send message
Joined: 20 Mar 13
Posts: 13
Credit: 476,694
RAC: 0
United States
Message 1350978 - Posted: 26 Mar 2013, 22:17:01 UTC

I dont know if it's another symtom or not but I have a 24/7 Broadband connection through CenturyLink but my downloads are constantly stalling out and then not even retrying until hours later unless I go and manually restart them. Even the small 360KB Cuda files will stall out over and over again. The 8 Meg Astropulse files could take 10 retries or more to get the whole file.

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4087
Credit: 33,003,041
RAC: 5,808
United Kingdom
Message 1350985 - Posted: 26 Mar 2013, 22:25:02 UTC - in response to Message 1350978.

I dont know if it's another symtom or not but I have a 24/7 Broadband connection through CenturyLink but my downloads are constantly stalling out and then not even retrying until hours later unless I go and manually restart them. Even the small 360KB Cuda files will stall out over and over again. The 8 Meg Astropulse files could take 10 retries or more to get the whole file.

The Current Hurricane link is constantly at well over 100% Utilisation, causing slow downloads and dropped connections,

If you do the fix in this thread, it'll improve the connection reliability no end:

Windows TCP Settings - Follow up - Help with server communication

Claggy

strue87
Send message
Joined: 20 Mar 13
Posts: 13
Credit: 476,694
RAC: 0
United States
Message 1351063 - Posted: 27 Mar 2013, 3:16:40 UTC - in response to Message 1350985.



If you do the fix in this thread, it'll improve the connection reliability no end:

Windows TCP Settings - Follow up - Help with server communication

Claggy


Wow your a genius! It made a Vast Improvement to my downloads they're going like gangbusters now and I'm averaging 25KBps download speeds between the two downloads.

The only problem I ran into was that when I hit the 'apply' button after selecting the optimize button in the tcpoptimizer program that it asked me if I wanted to reboot and I clicked yes but it only did a soft reboot and then BOINC couldnt connect to the internet at all so I thought something went wrong but finally tried a full reboot and then it just took off!

Thanks so much for your help.

Message boards : Number crunching : SETI Schedular Problems?

Copyright © 2014 University of California