resrictiong the Nvidia GPU usage to opencl_nvidia_SoG

Message boards : Number crunching : resrictiong the Nvidia GPU usage to opencl_nvidia_SoG
Message board moderation

To post messages, you must log in.

AuthorMessage
Aarron
Volunteer tester

Send message
Joined: 11 Oct 06
Posts: 1
Credit: 2,470,042
RAC: 10
United Kingdom
Message 1991032 - Posted: 22 Apr 2019, 2:40:25 UTC

Hi all,
I've been crunching for a while and have recently got a new graphics card, an RTX 2070, running the SoG application with s@h take between 3 and 5 mins to process a typical WU but I'm finding the cuda 42 and cuda 50 applications are taking 15-19 mins per WU. this is a lot slower. is there anything I can do besides installing the lunatic's optimized apps to make the project only use the SoG application?


Many thanks,

Aarron
ID: 1991032 · Report as offensive
Profile Bill G Special Project $75 donor
Avatar

Send message
Joined: 1 Jun 01
Posts: 1138
Credit: 170,011,824
RAC: 85,772
United States
Message 1991034 - Posted: 22 Apr 2019, 3:25:22 UTC - in response to Message 1991032.  

Just give it a bit of time and it will settle on the fastest one.
This is normal operation. Or you could just install Lunatics and it would go with your pick..

SETI@home classic workunits 4,019
SETI@home classic CPU time 34,348 hours
ID: 1991034 · Report as offensive
Profile Wiggo "Democratic Socialist"
Avatar

Send message
Joined: 24 Jan 00
Posts: 16695
Credit: 226,354,133
RAC: 162,839
Australia
Message 1991048 - Posted: 22 Apr 2019, 6:03:47 UTC

Even though the servers will eventually settle on the best app they will on occasions revisit those older apps.

The only way to stop this is to use the Lunatics Installer.

The installer will automatically select the much quicker AVX CPU MB8 app and you just have to select the SoG MB8 app for your Nvidia GPU's, make your selections to do Astropulse CPU and Nvidia apps and skip the Intel iGPU and AMD/ATi sections.

Cheers.
ID: 1991048 · Report as offensive

Message boards : Number crunching : resrictiong the Nvidia GPU usage to opencl_nvidia_SoG


 
©2019 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.