PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise
11) Message boards : Generalized Fermat Prime Search : Help needed debugging! (Message 59687)
Posted 3229 days ago by Profile Seth Johnson
Mike,

If I'm barging into the wrong thread with this, I'll go elsewhere and post this...

I'm on Mac OS X 10.7

I previously had this problem where the GeneferCUDA would sit in a 'ready to work' state and never run. Even when the computer is rebooted and BOINC is the only thing running. You changed something on the server when I reported this before.

I had to reinstall the OS due to my previous graphic card dying on me. The fans locked up. The vendor has replaced it with a newer model, and my OS wasn't friendly to just swapping out the old card for the new card.

Anyway, I'm back to 'ready to work' state, even though the log shows the card being recognized.


What do do?


Mon Nov 19 01:26:26 2012 | | OpenCL: NVIDIA GPU 0: GeForce GTX 560 Ti (driver version CLH 1.0, device version OpenCL 1.1, 1024MB, 842MB available)



12) Message boards : General discussion : Is there a log file available -- Mac OS X clients (Message 53396)
Posted 3440 days ago by Profile Seth Johnson
Hi Seth,

you can find most config files, log files and stderr* and stdout* files related to BOINC in the following directory:

/Library/Application Support/BOINC Data/

Let me know if you are looking for anything more specific.

HTH,
Phil



Thanks! That's what I was looking for.

Appreciatively,

Seth
13) Message boards : General discussion : Is there a log file available -- Mac OS X clients (Message 53215)
Posted 3444 days ago by Profile Seth Johnson
What is the location of those files? I did a search on my entire computer and couldn't find any of those three.

thanks,

Seth
14) Message boards : General discussion : Is there a log file available -- Mac OS X clients (Message 53170)
Posted 3445 days ago by Profile Seth Johnson
Was wondering if there is a log file available that I can actively parse and send out notifications when problems are encountered. Where would it be located?

Any help on this appreciated.

thanks,

Seth
15) Message boards : Problems and Help : still no CUDA work units available - Mac OS X (Message 52954)
Posted 3450 days ago by Profile Seth Johnson
My GPU finished its first world-record Genefer work unit and seemed to report it correctly. Now it's working on its second. Total time is estimated at a bit over 300 hours.

thanks,

Seth
16) Message boards : Number crunching : The Marie Sophie Germain Challenge 2012 (Message 52544)
Posted 3459 days ago by Profile Seth Johnson
Can you provide a link to where I can see the stats and standings of the Marie Sophie Germain challenge? I've read through every forum post I can find and see no link to the stats for this challenge.

thanks,

Seth
17) Message boards : Problems and Help : still no CUDA work units available - Mac OS X (Message 52276)
Posted 3463 days ago by Profile Seth Johnson
I'll let you know in about a hundred hours.

-- Seth Johnson
18) Message boards : Problems and Help : still no CUDA work units available - Mac OS X (Message 52223)
Posted 3465 days ago by Profile Seth Johnson
I don't know if someone switched a flip, but now my weaker-GPU machine is receiving GCWsieve wu's. I won't know if there's an improvement on my 460 machine because it's bit into one of those world record Genefer wu's that will take over a hundred hours. I'll post up here after that's complete.

thanks,

Seth
19) Message boards : Problems and Help : still no CUDA work units available - Mac OS X (Message 52219)
Posted 3465 days ago by Profile Seth Johnson
I have a GTX 460 1gb video card and previously was only getting Genefer CUDA work units on this machine even though my preferences were for every Mac OS X CUDA project available in the preferences.

I had computation errors today on my Genefer short CUDA work units, and now the server won't give me any CUDA work units. This is what it says in the log:

Wed Mar 28 23:56:28 2012 | PrimeGrid | update requested by user Wed Mar 28 23:56:29 2012 | PrimeGrid | Sending scheduler request: Requested by user. Wed Mar 28 23:56:29 2012 | PrimeGrid | Requesting new tasks for NVIDIA GPU Wed Mar 28 23:56:31 2012 | PrimeGrid | Scheduler request completed: got 0 new tasks Wed Mar 28 23:56:31 2012 | PrimeGrid | No work sent Wed Mar 28 23:56:31 2012 | PrimeGrid | Message from server: No work available for the applications you have selected. Please check your project preferences on the web site.


I have toggled the CUDA check boxes a few different ways and updated the project and still don't get any work units. I finally changed to the world record Genefer Work Units, and it sent me one of those.

How can I get other types of CUDA work units? Also, I have another Mac OS X Lion machine here with a much weaker GPU, and it gets nothing but PPS Sieve 1.39 cuda32 work units.

Finally, on the Mac OS X Boinc client, is there any log file I can use for monitoring the client?

thanks,

Seth
20) Message boards : Problems and Help : No CUDA work units for GTX470 on OSX (Message 52020)
Posted 3469 days ago by Profile Seth Johnson

CUFFT error 2 is "out of memory", which is strange because you should have enough. I've never seen that happen at n=22 before, but I have seen it happen at n=23, which uses twice as much video memory for the calculations (and twice as much memory on the main computer, too.)

One thing I learned when getting that error is that there appear to be things of unknown origin that may use video memory and not release it. That leaves the GPU with insufficient memory for crunching large numbers. Or there's another CUDA program already running, which is using some of the video memory

Rebooting might solve the problem, so I would suggest rebooting and trying another long WU.



Michael,

I've rebooted this machine and it's now processed several Genefer 1.07 Cuda32_13 work units. I am not seeing errors on my end, so I hope the memory issue has been resolved.

The other remaining issue here is that although in the preferences I've added checks for every cuda application, I only receive the Genefer work units on either of my two Mac OS X machines with nvidia cards. Is there still a server issue preventing me from receiving the Cullen/Woodall Prime Search (Sieve), Proth Prime Search (Sieve), and Proth Prime Search (LLR) Cuda work units?

thanks,

Seth


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2021 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 0.00, 0.00, 0.00
Generated 23 Sep 2021 | 7:59:32 UTC