blanked AP - faster or slower

Message boards : Number crunching : blanked AP - faster or slower
Message board moderation

To post messages, you must log in.

AuthorMessage
merle van osdol

Send message
Joined: 23 Oct 02
Posts: 809
Credit: 1,980,117
RAC: 0
United States
Message 1572418 - Posted: 15 Sep 2014, 21:55:13 UTC

Does it make any difference if you give a highly blanked AP to your faster or slower gpu? I would think not, but?
ID: 1572418 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 1572423 - Posted: 15 Sep 2014, 22:12:06 UTC

At present, blanking is done on the CPU, which is considerably slower than a GPU. Blanked AP GPU tasks will take longer the more blanking there is.

Soon, though, with the next release of the AP apps, blanking won't be an issue for GPUs. For now, yes, the more blanking there is, the longer the task will take.
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 1572423 · Report as offensive
tbret
Volunteer tester
Avatar

Send message
Joined: 28 May 99
Posts: 3380
Credit: 296,162,071
RAC: 40
United States
Message 1572435 - Posted: 15 Sep 2014, 22:57:24 UTC - in response to Message 1572418.  

Does it make any difference



HUH?

The short answer is "yes."

The long answer is that your faster GPU will always be faster than the slow GPU, so any part of the blanked WU done on the GPU will go by faster...

So if you are running two highly blanked AP units at the same time, it doesn't matter. If you are running one highly blanked and the other not, then the not-blanked will be quicker on the quicker card.
ID: 1572435 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1572443 - Posted: 15 Sep 2014, 23:21:51 UTC

I'm not sure you could actually force BOINC to only run highly blanked tasks on one of two GPUs. Well not without constantly micromanaging the system task by task.
Any efficiency gains would probably be lost in the constant stopping and starting of tasks to try and move them around.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1572443 · Report as offensive
merle van osdol

Send message
Joined: 23 Oct 02
Posts: 809
Credit: 1,980,117
RAC: 0
United States
Message 1572448 - Posted: 15 Sep 2014, 23:29:26 UTC - in response to Message 1572443.  
Last modified: 15 Sep 2014, 23:30:16 UTC

That's what I did. I moved the highly blanked task to the slower gpu and let the faster one run through many unblanked tasks. I don't know if it makes any difference in the sum total of 'points efficiency' or not.
ID: 1572448 · Report as offensive

Message boards : Number crunching : blanked AP - faster or slower


 
©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.