GPU Issues Mega Thread
log in

Advanced search

Message boards : News : GPU Issues Mega Thread

Previous · 1 . . . 5 · 6 · 7 · 8
Author Message
ErickGriffin
Send message
Joined: 15 Jul 12
Posts: 2
Credit: 7,060,015
RAC: 7,849

Message 66092 - Posted: 7 Jan 2017, 13:49:52 UTC

Seeing a BSOD which I think is being caused by MilkyWay GPU s/w. I have several Win10 minidumps of the issue (I hope), or whatever windows takes when it goes blue. The reason why I think its GPU is I have that set only to run when I am away from the computer and when screensaver is running, that's when it always happens.

CPU = AMD FX-8370 Eight-Core Processor, 4013 MHz, 4 Kern, 8 logische Prozessors
Mem = 16GB
Graphics = GeForce GTX 1080
Driver = Nvidia 376.33
OS = Microsoft Windows 10 Pro, 10.0.14393 Build 14393

Boinc = 7.5.33 (x64), wxWidgets = 3.0.1

Not sure if you get informed by Microsoft on issues related to your product or not. Please advise what you need next.

captainjack
Send message
Joined: 22 Jun 13
Posts: 40
Credit: 34,529,178
RAC: 13,351

Message 66093 - Posted: 7 Jan 2017, 14:21:38 UTC

wb8ili,

Glad you got it fixed.

Just in case you are still wondering, most of the client files are in /var/lib/boinc-client (projects folder and slots folder included). The procedure that starts/restarts the boinc client is in /etc/init.d and is called boinc-client.

Happy crunching.

ErickGriffin
Send message
Joined: 15 Jul 12
Posts: 2
Credit: 7,060,015
RAC: 7,849

Message 66094 - Posted: 7 Jan 2017, 17:20:42 UTC - in response to Message 66092.

I think this is actually a BOINC issue, I stopped MilkyWay and loaded a different project that uses GPU, and hit same issue even faster. Will post on their forum.

Chris Rampson
Send message
Joined: 14 May 11
Posts: 6
Credit: 27,418,790
RAC: 21,069

Message 66139 - Posted: 25 Jan 2017, 0:53:59 UTC

Speaking about the "#define Q_INV_SQR inf" Milkyway error:

I have been running BOINC/SETI for over 20 years. I have projects in Einstein, SETI, GPUGrid, Rosetta and Milkyway. Of the Milkyway project, there are 3 different types of jobs that I noticed. ONLY ONE OF THOSE IS FAILING.

I do not subscribe to the idea that there is a configuration problem when all my other projects are screaming along (including alternate Milkyway jobs).

The most reasonable explanation is a TYPO in the project files. Someone fatfingered a "t" into a "f" - which is not too hard to do since they are next to each other on the keyboard.

What needs to happen is for a Milkyway developer to explain why this happens - and hopefully fix his own code.

captainjack
Send message
Joined: 22 Jun 13
Posts: 40
Credit: 34,529,178
RAC: 13,351

Message 66141 - Posted: 25 Jan 2017, 15:12:04 UTC
Last modified: 25 Jan 2017, 15:48:26 UTC

Chris Rampson,

If you believe that there is a problem with the code, you might ask yourself this question: Why are other Milkyway users able to successfully complete those tasks using Linux and NVIDIA GPU's?

Edit:

Re: this thread http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=4087&postid=66127

Sebastian*
Send message
Joined: 8 Apr 09
Posts: 64
Credit: 4,181,538,754
RAC: 3,670,959

Message 66172 - Posted: 10 Feb 2017, 18:26:52 UTC

Hello again, and a happy new year to everybody.

I still got issues when running several WUs in parallel on the Hawaii bases GPUs. One WU at a time still runs fine.

Could someone look at my invalid WUs with the Validate errors. I can't make anything out of the text.

https://milkyway.cs.rpi.edu/milkyway/results.php?hostid=705276&offset=0&show_names=0&state=5&appid=

Maybe some can help me figure out what is broken.

AMD drivers have improved, the WUs don't hang any longer, but there are still Validate errors.

Previous · 1 . . . 5 · 6 · 7 · 8
Post to thread

Message boards : News : GPU Issues Mega Thread


Main page · Your account · Message boards


Copyright © 2017 AstroInformatics Group