Problems with newer Nvidia drivers

Message boards : Number crunching : Problems with newer Nvidia drivers
Message board moderation

To post messages, you must log in.

AuthorMessage
BWX

Send message
Joined: 31 May 03
Posts: 36
Credit: 156,754,993
RAC: 24
United States
Message 1872689 - Posted: 13 Jun 2017, 1:48:23 UTC

Ever later than driver version 376.33 (last good driver), cuda units error out. the latest message is:

SETI@home | task postponed 180.000000 sec: Cuda runtime, memory related failure, threadsafe temporary Exit

Going back to 376.33 works fine.

I thought there was a solution to this but couldn't find one easily.

What am I missing?
ID: 1872689 · Report as offensive
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 34744
Credit: 261,360,520
RAC: 489
Australia
Message 1872690 - Posted: 13 Jun 2017, 2:02:40 UTC

Why not update to Lunatics 0.45 Beta 6 and select the the SoG GPU app instead of using any of those old CUDA apps?

You're using the oldest Cuda32 ATM which is likely causing your problem when Cuda50 would've been better, but the SoG app will be even more productive than any of those even if you have to reserve a CPU core for each task that you run on that GTX 970 (you should be able to run 2 tasks at a time on that GPU meaning that you have to reserve 2 CPU cores to do that).

Cheers.
ID: 1872690 · Report as offensive
Profile Darrell
Volunteer tester
Avatar

Send message
Joined: 14 Mar 03
Posts: 267
Credit: 1,418,681
RAC: 0
United States
Message 1872852 - Posted: 14 Jun 2017, 7:24:20 UTC

Looking thru your task results (And I am not an NVidia person), the error looks alot like the kind I get on my AMD when I use too high an -sbs setting in the command line. You might want to get with Wiggo and check your settings.
... and still I fear, and still I dare not laugh at the Mad Man!

Queen - The Prophet's Song
ID: 1872852 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1872930 - Posted: 14 Jun 2017, 15:47:27 UTC - in response to Message 1872852.  

Mike is the go to person for setting
ID: 1872930 · Report as offensive
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 34744
Credit: 261,360,520
RAC: 489
Australia
Message 1872976 - Posted: 14 Jun 2017, 22:17:22 UTC - in response to Message 1872930.  

Mike is the go to person for setting

+1
ID: 1872976 · Report as offensive
Profile Jimbocous Project Donor
Volunteer tester
Avatar

Send message
Joined: 1 Apr 13
Posts: 1853
Credit: 268,616,081
RAC: 1,349
United States
Message 1873313 - Posted: 16 Jun 2017, 10:50:47 UTC - in response to Message 1872689.  

Ever later than driver version 376.33 (last good driver), cuda units error out. the latest message is:

SETI@home | task postponed 180.000000 sec: Cuda runtime, memory related failure, threadsafe temporary Exit

Going back to 376.33 works fine.

I thought there was a solution to this but couldn't find one easily.

What am I missing?

Fwiw, I've been fighting this issue for quite some time, on 5 different machines, and pretty well came to the same conclusion that much past 376 there's an incompatibility. Initially, I thought it was a hardware problem with PCI-E extenders I was using to add external GPUs, so I wasted a ton of time chasing that. Coupling that with Win10's insistence on updating drivers without permission and it becomes a real issue.
As with some other suggestions in this thread, I'm retiring Cuda50 in favor of SoG apps included in the Lunatics Beta 6. Perhaps now I can put those extra 980s back to work.
[feeling extraordinarily stupid about not checking this message base about this earlier.]
ID: 1873313 · Report as offensive

Message boards : Number crunching : Problems with newer Nvidia drivers


 
©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.