| Author |
Message |
|---|
|
|
|
Hi,
For SR5 i'm using this as my app_config.xml on my host SEPDE
<app_config>
<app>
<name>llrSR5</name>
<fraction_done_exact/>
<max_concurrent>1</max_concurrent>
</app>
<app_version>
<app_name>llrSR5</app_name>
<cmdline>-t 6</cmdline>
<avg_ncpus>6</avg_ncpus>
</app_version>
</app_config>
And as per the instructions i've seen on say the 321 challenge page I fully expected i'd get 6 tasks the first time after
But I've done the six tasks and got given six more each time. I was under the impression I should only be getting 1 new task each time?
Am I being particularly stupid?
I've read the config file in and restarted the host.
____________
 |
|
|
 Crun-chi Volunteer tester
Send message
Joined: 25 Nov 09
Posts: 3008
ID: 50683
Credit: 59,175,698
RAC: 1,834
                    
|
|
Did you set resources share to 0% !
Did you reduce number of core to use to 12.5%?
____________
92*10^1439761-1 REPDIGIT PRIME :) :) :)
314187728^131072+1 GENERALIZED FERMAT
31*332^367560+1 CRUS PRIME
Proud member of team Aggie The Pew. Go Aggie! |
|
|
|
|
|
Resources shares is 0. I hadn't realised that I needed to reduce the cores though. I'll set what changing to 12.5 does.
____________
|
|
|
|
|
|
The problem with reducing it to 12.5% is that my gpu tasks then stop running.
____________
|
|
|
|
|
The problem with reducing it to 12.5% is that my gpu tasks then stop running.
Try 25%, that should let 2 cpu's crunch, one for the cpu tasks and one for the gpu. |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 236,922,854
RAC: 3,199
                       
|
Resources shares is 0.
That's the problem. This will force the client to wait until it's completely out of work before requesting more work. That's not what you want to happen. Resource share needs to be greater than 0. If it's 0, it's treating PrimeGrid as a backup project, and that will force it to load a full set of (too many) tasks each time.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
Actually now that I engage my brain and check properly the resource share is 100.
At any rate we shall see what happens in 3 hours when the remaining three tasks finish of the latest 6.
____________
|
|
|
|
|
|
Update. Still not working correctly, getting 6 tasks each time.
Resource share is 100 and im not running any other project.
If I set the core usage to anything less than 75% then the GPU tasks stop running.
75% being 6 of the 8 processors.
____________
|
|
|
|
|
|
Do you have an app_config for your GPU task?
What did the log say when it failed to get another GPU task? |
|
|
|
|
|
I don't have an app_config for the gpu tasks. I wasn't aware that I needed to do that.
____________
|
|
|
KeithVolunteer tester
Send message
Joined: 8 Dec 13
Posts: 436
ID: 284516
Credit: 412,432,810
RAC: 6,580
               
|
|
You don't need an app_config for the GPU tasks.
"75% of the CPUs" is a good setting for using 6 of 8 cores.
Is the task finishing and sending the result in before the new tasks are downloaded or are the new tasks downloaded before the in progress one is done?
The only time I've see this behaviour is when I run out of tasks completely. As long as the new task downloads before the previous one is finished I only get one task at a time.
Grasping at straws here, but you wouldn't have set any non-standard disk usage settings or low disk space; would you?
____________
My Primes
Badge Score: 2*1 + 3*1 + 4*2 + 6*4 + 7*9 + 9*1 + 10*2 = 129
 |
|
|
|
|
|
Thanks to everyone that's replied.
Its now working. I haven't changed anything but i'm only getting one task at a time now.
____________
|
|
|
|
|
I don't have an app_config for the gpu tasks. I wasn't aware that I needed to do that.
You don't unless you want to run multiple simultaneous tasks which isn't useful outside of PPS Sieve. Just thought maybe you had specified 1 cpu for your gpu task and that's what was keeping them from downloading (log file would have said something like cpu tasks full). |
|
|
|
|
|
And now i've randomly been sent 6 tasks again. Oh well.
____________
|
|
|
axnVolunteer developer
Send message
Joined: 29 Dec 07
Posts: 285
ID: 16874
Credit: 28,027,106
RAC: 0
        
|
And now i've randomly been sent 6 tasks again. Oh well.
Try this.
1. Set the CPUs to use as 12.5%
2. Modify the app config to do some trickery.
<app_config>
<app_version>
<app_name>llrSR5</app_name>
<cmdline>-t 4</cmdline>
<avg_ncpus>1</avg_ncpus>
</app_version>
</app_config>
Note the changes in cmdline and avg_ncpus. Yours is a HT 8 core, so only 4 genuine cores. -t 4 is good enough (however you can play with this parameter to see if there is a better setting). Also, we're lying to BOINC that it uses only 1 core.
With this setting, it should only get 1 CPU task at a time. There should be no issue running a GPU task with this setting, but if it doesn't, try reducing the avg_ncpus to 0.95
I have thrown out all unwanted cruft from the app_config setting. What you see here is all you need. |
|
|
|
|
|
Thanks Axn. Im using your config and will see what happens once the remaining tasks have processed.
But the processor % is set at 12.5 and the gpu tasks are still processing as well which is progress.
____________
|
|
|
|
|
|
I'm having a problem with a computer in that it's losing internet connection, and when it regains internet connection it downloads multiple tasks, it looks like it depends on if it's processing a task, eg: if it's working on a 321 and regains internet it downloads only 1 task, if it loses connection and finishes a 321 multi threaded t4 then regains it downloads 4 new tasks but runs it a single one at t4 and the others wait |
|
|
axnVolunteer developer
Send message
Joined: 29 Dec 07
Posts: 285
ID: 16874
Credit: 28,027,106
RAC: 0
|
I'm having a problem with a computer in that it's losing internet connection, and when it regains internet connection it downloads multiple tasks, it looks like it depends on if it's processing a task, eg: if it's working on a 321 and regains internet it downloads only 1 task, if it loses connection and finishes a 321 multi threaded t4 then regains it downloads 4 new tasks but runs it a single one at t4 and the others wait
Same problem, same solution. Use the approach in my post with parameters modified for your situation. |
|
|
|
|
|
Update - thanks AXN its now working as expected.
One SR5 task at a time running on multiple processors.
____________
|
|
|
