Advanced search

Message boards : Number crunching : Do all the tasks fails CUDA 4.2 and CUDA 3.1...?

Author Message
Profile Carlesa25
Avatar
Send message
Joined: 13 Nov 10
Posts: 328
Credit: 72,619,453
RAC: 0
Level
Thr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 26180 - Posted: 5 Jul 2012 | 9:50:17 UTC
Last modified: 5 Jul 2012 | 10:13:03 UTC

Hello: I are failing all tasks 4.2 CUDA now also the CUDA 3.1...?

UBUNTU 12.04 GTX 295

"<core_client_version>7.0.28</core_client_version>
<![CDATA[
<message>
process exited with code 98 (0x62, -158)
</message>
<stderr_txt>
# Using device 0
# There are 2 devices supporting CUDA
# Device 0: "GeForce GTX 295"
# Clock rate: 1.24 GHz
# Total amount of global memory: 939327488 bytes
# Number of multiprocessors: 30
# Number of cores: 240
# Device 1: "GeForce GTX 295"
# Clock rate: 1.24 GHz
# Total amount of global memory: 939196416 bytes
# Number of multiprocessors: 30
# Number of cores: 240
ERROR: Failed to parse input file
11:59:55 (16677): called boinc_finish

</stderr_txt>
]]> "

Ken Florian
Send message
Joined: 4 May 12
Posts: 56
Credit: 1,832,989,878
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 26185 - Posted: 5 Jul 2012 | 15:00:29 UTC - in response to Message 26180.

I am in a similar situation.

I added a 550TI to my machine and set fans to maintain temps of <=72.

Tasks started failing with "failed to parse input file".

Had it no been for Carlesa25's post, I would have attributed it to the addition of the second card but perhaps this is not the problem??

Ken

Profile Carlesa25
Avatar
Send message
Joined: 13 Nov 10
Posts: 328
Credit: 72,619,453
RAC: 0
Level
Thr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 26186 - Posted: 5 Jul 2012 | 15:57:51 UTC - in response to Message 26185.

After many failures, CUDA 3.1 and 4.2 CUDA have entered two CUDA 3.1 tasks that are running well... we'll see if it lasts.

Is proving very confusing and irritating the theme of several CUDAs

mynis
Send message
Joined: 31 May 12
Posts: 8
Credit: 12,361,387
RAC: 0
Level
Pro
Scientific publications
watwatwat
Message 26187 - Posted: 5 Jul 2012 | 17:04:51 UTC

I noticed that you are running Ubunutu 12.04, any chance that you updated to the 302.17 NVidia driver? I all ready posted about this in a few other threads, but I was experiencing similar issues and reverting to 295.59 allowed me to continue crunching.

Profile Carlesa25
Avatar
Send message
Joined: 13 Nov 10
Posts: 328
Credit: 72,619,453
RAC: 0
Level
Thr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 26188 - Posted: 5 Jul 2012 | 17:44:28 UTC - in response to Message 26187.

I noticed that you are running Ubunutu 12.04, any chance that you updated to the 302.17 NVidia driver? I all ready posted about this in a few other threads, but I was experiencing similar issues and reverting to 295.59 allowed me to continue crunching.



Hello: Thanks for the comment, I'm using 295.59 yesterday worked perfectly with CUDA 3.1 (CUDA 4.2 always fails)now seems returns to work.

JAMES DORISIO
Send message
Joined: 6 Sep 10
Posts: 8
Credit: 3,207,346,852
RAC: 4,660,146
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 26190 - Posted: 5 Jul 2012 | 19:27:35 UTC

I also started getting failures on Boinc 6.10.17 Ubuntu 10.04 with nividia driver 302.17 on 2 GTX550ti computers , the first few work units after the driver update went thru ok then a lot of failures most cuda 4.2 but some 3.1 as well, both computers suspended for now.

A 3rd one a GTS450 running Ubuntu 10.04 with 302.17 has been receiving all cuda 3.1 and is still running ok.

Is anybody having success with Ubuntu and 302.17 with cuda 4.2?
Jim

mynis
Send message
Joined: 31 May 12
Posts: 8
Credit: 12,361,387
RAC: 0
Level
Pro
Scientific publications
watwatwat
Message 26205 - Posted: 7 Jul 2012 | 1:46:00 UTC - in response to Message 26190.
Last modified: 7 Jul 2012 | 1:48:15 UTC

Based on what I've read on the forums here, a few people seem to be running CUDA 4.2 ok with 302.17 but more often than not people are having issues. Myself, I had to downgrade to 295.59 to avoid getting "output file not found" errors. We should probably make a separate thread about it or something. The most interesting part is that people seem to be getting a large variety of different errors, but each machine only outputs one specific error every time.

Post to thread

Message boards : Number crunching : Do all the tasks fails CUDA 4.2 and CUDA 3.1...?

//