When reporting CUDA errors...

Questions and Answers : GPU applications : When reporting CUDA errors...
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 843857 - Posted: 22 Dec 2008, 23:06:11 UTC

When you (the user) do run into Cuda specific errors, such as (e.g.):
"Cuda error 'cudaAcc_GetPowerSpectrum_kernel' in file 'c:/sw/gpgpu/seti/seti_boinc/client/cuda/cudaAcc_PowerSpectrum.cu' in line 56 : unknown error."

Could you please also post:
1. a link to the task that had the error.
2. the driver version that you are using.
3. OS type (XP/Vista) 32 or 64-bit.

This makes it easier for those of us reporting all these errors to the developers and easier for the developers to come up with a fix.
ID: 843857 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 844973 - Posted: 25 Dec 2008, 12:50:19 UTC

A further request from the developers. When you report errors, could you also please:

a) state if you overclock the GPU and if so by how much?
b) tell if you have tried to clock the GPU to default speeds and still see the problem?
c) did you set the fan speed manually to anything else than default?
d) which program(s) do you use to overclock the GPU, fan etc.?
e) which program(s) do you use to keep check of the GPU?
ID: 844973 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 848656 - Posted: 3 Jan 2009, 12:18:07 UTC

Please report errors with the stock application only in this forum. When you're using Raistmer's modified application (or that of anyone else), report about its errors in the Number Crunching forum where the modifiers/optimizers can answer you.

You can recognize if you are using Raistmer's application when you see the following in the stderr.txt output of the task:

Rise priority modification by Raistmer based on rev380 of SETI@home sources
Priority of worker thread rised successfully


The stock CUDA application (6.06) will not output this message to stderr.txt
ID: 848656 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 854002 - Posted: 16 Jan 2009, 0:52:42 UTC

Please people, show some common sense on those things you cannot do when you run Seti on the GPU are all those other things that use the GPU in full.

- If you have a dedicated gaming rig, then do not run BOINC/CUDA at the same time you are playing your game, or only run it when you aren't gaming.
- When you love to watch DVDs, do not use BOINC/CUDA at the same time. This will impact your viewing experience. (stutters, slow-downs)
- When using a screen saver, remember that the GPU is used to render the screen saver's graphics. Why not just not use a screen saver and turn off your monitor before you leave, or let it go to blank?

Seti will at all times try to use the whole GPU, in full saturation for as long as there are CUDA tasks running and waiting to run. This will have an impact on anything else you do that uses the GPU.

Ways around it?
First and foremost is to exit BOINC before you start gaming or doing your other GPU intensive thing.
BOINC Manager->File->Exit, check "stop running of science applications when exiting BOINC Manager" and click OK.
Restarting BOINC Manager will also restart the BOINC service (if you installed it that way).


Second, you can suspend BOINC before you start gaming or your other GPU intensive thing. (*)
BOINC Manager->Advanced view->Activity->Suspend.
To resume folow the same path and set it back to Run based on preferences.


Third, you can have a cc_config.xml in your BOINC Data directory that will suspend your BOINC (under Windows only, I'm afraid) automatically when any of the exclusive applications you told it to look out for is detected in main memory. (*)

To do so, set up the cc_config.xml file with the following options (example of my cc_config.xml):
<cc_config>
  <options>
     <exclusive_app>Speed.exe</exclusive_app>
     <exclusive_app>Crysis.exe</exclusive_app>
     <exclusive_app>FarCry.exe</exclusive_app>
     <exclusive_app>NFSHP2.exe</exclusive_app>
     <exclusive_app>FarCry2.exe</exclusive_app>
     <exclusive_app>Fallout3.exe</exclusive_app>
     <exclusive_app>RA3.exe</exclusive_app>
     <no_gpus>0</no_gpus>
  </options>
</cc_config> 


You only need to tell what application name it needs to look out for, no need for the full path to the name.
If you already have one, you can add the <options></options> part to your existing cc_config.xml file.

Watch out with launcher files. For example, Fallout 3 uses the Fallout 3 Launcher, which does not stay in memory after you launched the game. So do not use the launcher file as the exclusive app, but the game executable's name instead.

(*) There is a problem when you have your preferences set to Leave applications in memory while suspended.
Any task still in the video card's RAM will stay there until you resume BOINC. Although it will not use your GPU, the amount of memory that a Seti CUDA tasks takes up (200MB+ of the VRAM) may still have impact on your other activities.

After a discussion I had with the developers tonight, it looks like this behaviour is set to change with a future version of BOINC. Then upon suspension of BOINC the task will be removed from the video card's memory after it has checkpointed at least once.
ID: 854002 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 865063 - Posted: 13 Feb 2009, 16:00:54 UTC

I have put together a FAQ with some questions and answers. I will be adding to it in the coming days, but at least you can already read it, here.

Have fun.
ID: 865063 · Report as offensive
Profile (_KoDAk_)
Volunteer tester

Send message
Joined: 12 Jul 08
Posts: 5
Credit: 65,783
RAC: 0
Ukraine
Message 884510 - Posted: 12 Apr 2009, 9:11:31 UTC
Last modified: 12 Apr 2009, 10:02:56 UTC

Exit status 1 (0x1)
2008_x64 182.08 overclocked (only shaders ) by EVGA Precision 1.6.0 fun+
http://setiathome.berkeley.edu/result.php?resultid=1203112217
http://setiathome.berkeley.edu/result.php?resultid=1203112249
http://setiathome.berkeley.edu/result.php?resultid=1203112234
http://setiathome.berkeley.edu/result.php?resultid=1203112249
GPUGrid have no errors
ID: 884510 · Report as offensive
Profile (_KoDAk_)
Volunteer tester

Send message
Joined: 12 Jul 08
Posts: 5
Credit: 65,783
RAC: 0
Ukraine
Message 884531 - Posted: 12 Apr 2009, 11:16:05 UTC

Exit status 1 (0x1)
http://setiathome.berkeley.edu/result.php?resultid=1203117327
http://setiathome.berkeley.edu/result.php?resultid=1203117332
http://setiathome.berkeley.edu/result.php?resultid=1203117574
ID: 884531 · Report as offensive
Profile (_KoDAk_)
Volunteer tester

Send message
Joined: 12 Jul 08
Posts: 5
Credit: 65,783
RAC: 0
Ukraine
Message 884864 - Posted: 13 Apr 2009, 5:44:42 UTC

my be problem in that i have 2 GPU (but they both 9600GSO and shader =1728)
ID: 884864 · Report as offensive
Profile (_KoDAk_)
Volunteer tester

Send message
Joined: 12 Jul 08
Posts: 5
Credit: 65,783
RAC: 0
Ukraine
Message 884937 - Posted: 13 Apr 2009, 14:31:18 UTC

and second problem is
that when add's new task(from another project) one GPU task become USE 100% CPU core !!!
if restart boinc = OK
but until NO add's new task(from another project)
ID: 884937 · Report as offensive
Profile Leopoldo
Volunteer tester
Avatar

Send message
Joined: 4 Aug 99
Posts: 102
Credit: 3,051,091
RAC: 0
Russia
Message 885025 - Posted: 13 Apr 2009, 18:35:42 UTC - in response to Message 884937.  

Cuda error 'cufftExecC2C' in file 'c:/sw/gpgpu/seti/seti_boinc/client/cuda/cudaAcc_fft.cu' in line 63 : unknown error.


fft stands for "Fast Fourier Transformation", AFAIK. IMHO, it's very vulnerable for overcloking. So, can You switch off overclocking and watch into results from new workunits?

___________
WBW, Leonid
ID: 885025 · Report as offensive
Stacey Baird
Avatar

Send message
Joined: 8 Nov 04
Posts: 41
Credit: 407,924
RAC: 0
Philippines
Message 891174 - Posted: 4 May 2009, 14:13:19 UTC

I'm using a new intel quad, an nvidia gforce 8400 video card that works with Cuda and the latest cuda driver and winXP Rev3. and the 6.6.26.

The problem is that Cuda is not running continuously, I would think I would have four processors rotating in and out of various projects and the one Cuda sticking with all the SETI 6.08 projects. Instead it processes very well but then stops and comes back an hour later.

Maybe a settings problem?

Thanks for any help.


Stacey Baird
Davao City, Philippines


The welfare of the people is the highest law - Cicero

If no one complains, the people must be satisfied.
ID: 891174 · Report as offensive
Ivailo Bonev
Volunteer tester
Avatar

Send message
Joined: 26 Jun 00
Posts: 247
Credit: 35,864,461
RAC: 2
Bulgaria
Message 891499 - Posted: 5 May 2009, 5:42:24 UTC

Cuda error 'cudaMemcpy(dev_cx_DataArray, cx_DataArray, NumDataPoints * sizeof(*cx_DataArray), cudaMemcpyHostToDevice)' in file 'c:/sw/gpgpu/seti/seti_boinc/client/cuda/cudaAcceleration.cu' in line 259 : unspecified launch failure.

Link to task:Error
Driver: ver.182.50
OS: Win XP SP3 32-bit
ID: 891499 · Report as offensive

Questions and Answers : GPU applications : When reporting CUDA errors...


 
©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.