To Many ERRORS

Message boards : Number crunching : To Many ERRORS
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4

AuthorMessage
Profile Boong

Send message
Joined: 25 Nov 08
Posts: 7
Credit: 16,943,808
RAC: 0
Australia
Message 1276690 - Posted: 28 Aug 2012, 14:24:57 UTC

Running BOINC client version 7.0.28 for windows_x86_64 , using anonymous platform , AMD Athlon II X4 640 Processor , ATI Radeon HD 5870 , Catalyst Version 12.8

Running 1x mb wu on each core of cpu and 1x mb wu on gpu

clicked on properties and dcf is on 1.

Please help i keep getting computation errors on gpu wu's

28/08/2012 7:46:08 PM | SETI@home | Aborting task 14my12ab.21457.20108.13.10.152_1: exceeded elapsed time limit 665.45 (1902.40G/2.86G)

my brothers computer is getting the same error , two other comps in the house are ok but they running a different client version and cat version

Thanks

Andrew

Melbourne AU
ID: 1276690 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34249
Credit: 79,922,639
RAC: 80
Germany
Message 1276700 - Posted: 28 Aug 2012, 14:32:24 UTC
Last modified: 28 Aug 2012, 14:35:51 UTC

Free at least one core of your CPU.
I will check later again.

Your APR is out of order.

Do you have flops included in your appinfo.xml ?


With each crime and every kindness we birth our future.
ID: 1276700 · Report as offensive
Profile Boong

Send message
Joined: 25 Nov 08
Posts: 7
Credit: 16,943,808
RAC: 0
Australia
Message 1276914 - Posted: 29 Aug 2012, 6:39:18 UTC - in response to Message 1276700.  

Free at least one core of your CPU.
I will check later again.

Your APR is out of order.

Do you have flops included in your appinfo.xml ?


no flops in my appinfo.xml .

do you know the flops for my card and how to put it in pls

how do i free up one core if i only run three wu it shares it across the four cores still
ID: 1276914 · Report as offensive
Profile Boong

Send message
Joined: 25 Nov 08
Posts: 7
Credit: 16,943,808
RAC: 0
Australia
Message 1276917 - Posted: 29 Aug 2012, 6:58:30 UTC - in response to Message 1267499.  

@Fred

I have calculated something for you.
Please add this line into your appinfo.

<flops>509408724.212160</flops>

Below your command line.

This should increase estimates, reduce your DCF and limit the errors.

Beware its very low value on purpose.



will this be the same for mine ill do this and see how i go
ID: 1276917 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34249
Credit: 79,922,639
RAC: 80
Germany
Message 1276930 - Posted: 29 Aug 2012, 8:30:35 UTC
Last modified: 29 Aug 2012, 8:32:49 UTC

Yes, you can use it also.
Those are very low calculated not to get other errors.

Close Boinc.
Open in notepad or something similar.
Put it after the cmdline.

<plan_class>ati13ati</plan_class>
<cmdline>-period_iterations_num 50 -instances_per_device 2 -no_cpu_lock</cmdline>
<flops>509408724.212160</flops>
<file_ref>

Make sure you save it as txt.
Restart boinc.


With each crime and every kindness we birth our future.
ID: 1276930 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34249
Credit: 79,922,639
RAC: 80
Germany
Message 1276937 - Posted: 29 Aug 2012, 9:14:01 UTC - in response to Message 1276914.  
Last modified: 29 Aug 2012, 9:14:34 UTC


how do i free up one core if i only run three wu it shares it across the four cores still


Go to tools/computingpreferences in your boinc manager.



On multiprecessors use 75 out of 100 cause you have a quad core CPU.


With each crime and every kindness we birth our future.
ID: 1276937 · Report as offensive
Profile Boong

Send message
Joined: 25 Nov 08
Posts: 7
Credit: 16,943,808
RAC: 0
Australia
Message 1276947 - Posted: 29 Aug 2012, 10:01:21 UTC - in response to Message 1276937.  

ok Thanks for that Mike

now after freeing up 1 core do i change the ncpus from

<max_ncpus>0.05</max_ncpus>

thanks
ID: 1276947 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34249
Credit: 79,922,639
RAC: 80
Germany
Message 1276976 - Posted: 29 Aug 2012, 11:02:26 UTC - in response to Message 1276947.  

ok Thanks for that Mike

now after freeing up 1 core do i change the ncpus from

<max_ncpus>0.05</max_ncpus>

thanks


No, no further changes necessary.

Lets see how it goes.



With each crime and every kindness we birth our future.
ID: 1276976 · Report as offensive
Profile Boong

Send message
Joined: 25 Nov 08
Posts: 7
Credit: 16,943,808
RAC: 0
Australia
Message 1278776 - Posted: 1 Sep 2012, 10:29:00 UTC
Last modified: 1 Sep 2012, 10:49:08 UTC

Things are running great now Mike no more ERRORS thanks . after freeing up one cpu core my gpu load dropped and stayed around 50% so now im running two wu on the single gpu but i have been noticing something strange , both wu run great and gpu loads sits around 95%-99% and both wu take about 15 minutes to complete from an estimate of 21 minutes but after the first two or three wu finish and the next one loads on the gpu the load drops and stays around 75%-80% and one unit takes around an hour to an hour and a half while the second wu still takes the 15 min with both wu with the same estimate , but again if i close boince and restart it both units take 15 min again for the first few then the same thing happens even if the wu have already been started and previously sed an hour it then takes 15 min after a restart of boince until the few have run then one wu slows down , do u have any ideas pls

also if i only run one wu on the gpu it still takes 15 min to complete at 55% load

Andrew
ID: 1278776 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34249
Credit: 79,922,639
RAC: 80
Germany
Message 1278788 - Posted: 1 Sep 2012, 10:54:07 UTC

So you are still suffering from low GPU usage bug.

You have to decide to free another core like i have.
On my FX CPU i need to free 2 CPU cores to get full utilisation.
That depends on over all load of the system.

If you can live with it its fine so long you dont get errors again.
If i were you i would try to free another core and see if the full GPU load benefits the second core.

Mike




With each crime and every kindness we birth our future.
ID: 1278788 · Report as offensive
Profile Fred J. Verster
Volunteer tester
Avatar

Send message
Joined: 21 Apr 04
Posts: 3252
Credit: 31,903,643
RAC: 0
Netherlands
Message 1278875 - Posted: 1 Sep 2012, 14:39:00 UTC - in response to Message 1278788.  

So you are still suffering from low GPU usage bug.

You have to decide to free another core like i have.
On my FX CPU i need to free 2 CPU cores to get full utilisation.
That depends on over all load of the system.

If you can live with it its fine so long you dont get errors again.
If i were you i would try to free another core and see if the full GPU load benefits the second core.

Mike



The errors stopped after freeing up 2 cores and also doing 2 MB WUs per GPU.

But the APR is still too high!
SETI@home Enhanced (anonymous platform, ATI GPU)
Number of tasks completed	18
Max tasks per day	2097
Number of tasks today	0
Consecutive valid tasks	17
Average processing rate	681.7520469167
Average turnaround time	0.52 day

AstroPulse v6 (anonymous platform, ATI GPU)
Number of tasks completed	72
Max tasks per day	306
Number of tasks today	0
Consecutive valid tasks	206
Average processing rate	658.96699180625
Average turnaround time	2.38 days. 



I7-2600 + 2 HD5870 GPUs.



ID: 1278875 · Report as offensive
Previous · 1 · 2 · 3 · 4

Message boards : Number crunching : To Many ERRORS


 
©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.