Welcome to MilkyWay@home

CUDA Application Available for 64 Bit Linux/GNU

Message boards : Number crunching : CUDA Application Available for 64 Bit Linux/GNU
Message board moderation

To post messages, you must log in.

AuthorMessage
Anthony Waters

Send message
Joined: 16 Jun 09
Posts: 85
Credit: 172,476
RAC: 0
Message 32331 - Posted: 12 Oct 2009, 17:26:04 UTC

The CUDA application for 64 Bit Linux/GNU has been released on BOINC the 32 Bit version will follow soon. Please see http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1083&nowrap=true#29742 for important information about the application. If Computation Errors are being experienced please make sure the BOINC client is from the 6.10.x series, the 6.6.x series has an issue relating to executing the CUDA client. With a GeForce GTX 285 that was donated from NVIDIA it takes about 2:55 with the work units being distributed.

Thanks to Crunch3r for pointing out the 6.10.x clients fixed the issues with the Linux/GNU CUDA clients.
ID: 32331 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile koubi

Send message
Joined: 14 Sep 08
Posts: 2
Credit: 322,080
RAC: 0
Message 32414 - Posted: 16 Oct 2009, 0:32:11 UTC

it work but some Wu go to error:

de_11_1s_const_1_119006_1255652113_0
Workunit 132576486
Created 16 Oct 2009 0:15:19 UTC
Sent 16 Oct 2009 0:15:49 UTC
Received 16 Oct 2009 0:28:41 UTC
Server state Over
Outcome Client error
Client state Compute error
Exit status 255 (0xff)
Computer ID 108916
Report deadline 19 Oct 2009 0:15:49 UTC
Run time 0.886608
stderr out

<core_client_version>6.10.13</core_client_version>
<![CDATA[
<message>
process exited with code 255 (0xff, -1)
</message>
<stderr_txt>
Device index specified on the command line was 0
Looking for a Double Precision capable NVIDIA GPU
The device GeForce GTX 260 specified on the command line can be used
cudaSafeCall() Runtime API error in file <../astronomy_gpu/evaluation_gpu6_double.cu>, line 172 : invalid argument.

</stderr_txt>
]]>

Validate state Invalid
Claimed credit 0.00376889624255691
Granted credit 0
application version 0.20


an idea?

os: debian lenny 5.03 x86_64
kernel:2.6.30.9
nvidia drivers: 190.32
boinc version: 6.10.13
gpu: nvidia gtx260 @ 756/1096/1512
ID: 32414 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile koubi

Send message
Joined: 14 Sep 08
Posts: 2
Credit: 322,080
RAC: 0
Message 32415 - Posted: 16 Oct 2009, 0:37:56 UTC - in response to Message 32414.  
Last modified: 16 Oct 2009, 0:38:33 UTC

other wu without error:

Name de_11_2s_const_1_92135_1255650118_0
Workunit 132549615
Created 15 Oct 2009 23:42:01 UTC
Sent 15 Oct 2009 23:42:35 UTC
Received 16 Oct 2009 0:05:16 UTC
Server state Over
Outcome Success
Client state Done
Exit status 0 (0x0)
Computer ID 108916
Report deadline 18 Oct 2009 23:42:35 UTC
Run time 1.239918
stderr out

<core_client_version>6.10.13</core_client_version>
<![CDATA[
<stderr_txt>
Device index specified on the command line was 0
Looking for a Double Precision capable NVIDIA GPU
The device GeForce GTX 260 specified on the command line can be used
called boinc_finish

</stderr_txt>
]]>

Validate state Valid
Claimed credit 0.00527078741820362
Granted credit 39.038531
application version 0.20


average 2:57 per wu
ID: 32415 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile [AF>Libristes] Dudumomo

Send message
Joined: 2 Apr 08
Posts: 9
Credit: 2,881,370
RAC: 0
Message 32420 - Posted: 16 Oct 2009, 7:01:45 UTC - in response to Message 32415.  

Hi here !
I do have the same problem, only WU starting by "gs_11_2s", "ps_11_2s" or "de_11_2s" works. But WU starting by "gs_11_1s","ps_11_1s" or "de_11_1s" doesn't works.

Always the same error :
<core_client_version>6.10.13</core_client_version>
<![CDATA[
<message>
process exited with code 255 (0xff, -1)
</message>
<stderr_txt>
Device index specified on the command line was 0
Looking for a Double Precision capable NVIDIA GPU
The device GeForce GTX 275 specified on the command line can be used
cudaSafeCall() Runtime API error in file <../astronomy_gpu/evaluation_gpu6_double.cu>, line 172 : invalid argument.

</stderr_txt>
]]>


Any ideas why ?

Thank you
ID: 32420 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Paul D. Buck

Send message
Joined: 12 Apr 08
Posts: 621
Credit: 161,934,067
RAC: 0
Message 32423 - Posted: 16 Oct 2009, 9:35:49 UTC

There seems to be a batch of these tasks which fail on the CUDA version of the MW application, but passes on the ATI version .20b ...

I am not sure why, but I got a rather long stack of them on my system with GTX295 cards ...

Perhaps CP can advise ...
ID: 32423 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile [AF>Libristes] Dudumomo

Send message
Joined: 2 Apr 08
Posts: 9
Credit: 2,881,370
RAC: 0
Message 32425 - Posted: 16 Oct 2009, 9:59:21 UTC - in response to Message 32423.  

Yep same thing here with an other member : http://milkyway.cs.rpi.edu/milkyway/results.php?hostid=113524&offset=0&show_names=0&state=5
ID: 32425 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Anthony Waters

Send message
Joined: 16 Jun 09
Posts: 85
Credit: 172,476
RAC: 0
Message 32473 - Posted: 17 Oct 2009, 3:46:29 UTC

Thanks for reporting the issue, I was able to get one of the 'broken' WU and fixed the issue that was occuring. The newest version 0.21 contains a fix, please report if it actually works (it worked for me).
ID: 32473 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile [AF>Libristes>Jip] Elgrande71
Avatar

Send message
Joined: 21 Mar 08
Posts: 20
Credit: 184,001,371
RAC: 715
Message 32480 - Posted: 17 Oct 2009, 7:31:02 UTC - in response to Message 32473.  
Last modified: 17 Oct 2009, 7:32:01 UTC

I always have same errors with this new application (0.21).
Sorry.
Example : http://milkyway.cs.rpi.edu/milkyway/result.php?resultid=136797218
ID: 32480 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile [AF>Libristes>Jip] Elgrande71
Avatar

Send message
Joined: 21 Mar 08
Posts: 20
Credit: 184,001,371
RAC: 715
Message 32485 - Posted: 17 Oct 2009, 9:00:36 UTC - in response to Message 32480.  

I always have same errors with this new application (0.21).
Sorry.
Example : http://milkyway.cs.rpi.edu/milkyway/result.php?resultid=136797218

Sorry but the problem seems to be fixed.
These errors has been producing by the 0.20 version of the milkyway application.
ID: 32485 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Number crunching : CUDA Application Available for 64 Bit Linux/GNU

©2024 Astroinformatics Group