V9 of modified SETI MB CUDA + opt AP package for full GPU+CPU utilization

[elec999]

My Boinc freezes as soon I putted in the new multi GPU files.
Thank you

[keeleysam]

Any VLAR kill team version for SSSE3?

[n0xlf]

VERY cool - My two GTX280s are crunching away, and I have 1 AP unit running on the CPU - I assume the other 3 cores will fill with more AP units when I can actually get them (things seem very plugged up on their end ATM)...

Hope this isn't too far off topic, but is there a reason why Seti doesn't use optimized versions by default?  I would think they would want the most computing bang for their buck they could get, and it isn't very difficult to detect processor capabilities...

[Raistmer]

Any VLAR kill team version for SSSE3?

It's VLAR kill one.

[Raistmer]

Hope this isn't too far off topic, but is there a reason why Seti doesn't use optimized versions by default?  I would think they would want the most computing bang for their buck they could get, and it isn't very difficult to detect processor capabilities...

They need much wider compatibility and portability than we have with opt apps.

[Raistmer]

My Boinc freezes as soon I putted in the new multi GPU files.
Thank you

Try to reboot. And if you want some more productive advice, please, post more info next time.

[K Rider]

I have three GTX280's with the V9 multi GPU version installed and it only uses two cards. Is there a setting I need to change or is this a limitation of CUDA/BOINC?
Thanks

[rpont]

I downloaded and installed the AMD version. Running under Windows 7 with a Geforce 9800 GT and I'm getting errors with the multibeam part.

I did back up the complete Boinc data folders and have restored and back running as before and I did reboot before starting the new version.

No doubt I've done something stupid but here is the error log for one failed workunit grabbed from client_state.xml:

<result>
    <name>16ja09aa.27910.34725.6.8.199_1</name>
    <final_cpu_time>290.224200</final_cpu_time>
    <exit_status>-1073741819</exit_status>
    <state>3</state>
    <platform>windows_intelx86</platform>
    <version_num>608</version_num>
    <fpops_cumulative>58866180000000.000000</fpops_cumulative>
<stderr_out>
<![CDATA[
<message>
 - exit code -1073741819 (0xc0000005)
</message>
<stderr_txt>
setiathome_CUDA: Found 1 CUDA device(s):
   Device 1 : GeForce 9800 GT
           totalGlobalMem = 536870912
           sharedMemPerBlock = 16384
           regsPerBlock = 8192
           warpSize = 32
           memPitch = 262144
           maxThreadsPerBlock = 512
           clockRate = 1500000
           totalConstMem = 65536
           major = 1
           minor = 1
           textureAlignment = 256
           deviceOverlap = 0
           multiProcessorCount = 14
setiathome_CUDA: CUDA Device 1 specified, checking...
   Device 1: GeForce 9800 GT is okay
SETI@home using CUDA accelerated device GeForce 9800 GT
setiathome_enhanced 6.03 Visual Studio/Microsoft C++
libboinc: 6.3.22

Work Unit Info:
...............
WU true angle range is :  0.432681
Optimal function choices:
-----------------------------------------------------
name                
-----------------------------------------------------
              v_BaseLineSmooth (no other)
            v_GetPowerSpectrum 0.00081 0.00000
                   v_ChirpData 0.02663 0.00000
                  v_Transpose4 0.01660 0.00000
               FPU opt folding 0.00658 0.00000
setiathome_CUDA: Found 1 CUDA device(s):
   Device 1 : GeForce 9800 GT
           totalGlobalMem = 536870912
           sharedMemPerBlock = 16384
           regsPerBlock = 8192
           warpSize = 32
           memPitch = 262144
           maxThreadsPerBlock = 512
           clockRate = 1500000
           totalConstMem = 65536
           major = 1
           minor = 1
           textureAlignment = 256
           deviceOverlap = 0
           multiProcessorCount = 14
setiathome_CUDA: CUDA Device 1 specified, checking...
   Device 1: GeForce 9800 GT is okay
SETI@home using CUDA accelerated device GeForce 9800 GT
Restarted at 17.66 percent.
setiathome_CUDA: Found 1 CUDA device(s):
Cuda error 'cudaGetDeviceProperties( &cDevProp, i )' in file 'd:/BTR/SETI6/SETI_MB_CUDA/client/cuda/cudaAcceleration.cu' in line 135 : invalid argument.
ERROR: Invalid parameter detected in function (null). File: (null) Line: 0
ERROR: Expression: (null)

</stderr_txt>
]]>
</stderr_out>
    <ready_to_report/>
    <completed_time>1235300353.341991</completed_time>
    <wu_name>16ja09aa.27910.34725.6.8.199</wu_name>
    <report_deadline>1237157056.000000</report_deadline>
    <file_ref>
        <file_name>16ja09aa.27910.34725.6.8.199_1_0</file_name>
        <open_name>result.sah</open_name>
    </file_ref>
</result>

[Richard Haselgrove]

COULD be a driver issue. What nVidia driver do you have installed?

[rpont]

COULD be a driver issue. What nVidia driver do you have installed?

The latest one from nVidia (Prerelease - WDDM 1.1) installed on 19 Feb by Windows Update.

[Richard Haselgrove]

Hmmm. I stick to the finalised release ones from nVidia - can't help with that one.

Probably best to suspend processing until Raistmer checks in and can advise further.

[rpont]

That's what I thought :-)

Prerelease is all I get for Windows 7 at the moment.

There was another thread about VLAR units but I can't find that now. Usually they complete in 20-25 mins but the VLAR ones took about an hour and a half which seemed to be a lot quicker than was expected on the other thread but they went through on the 18-19 Feb so no records of them now apart from whats in stdoutdae.txt they had an angle of 0.01 something. Two of them are still on the seti page WU 142908101 and 102 if anyone wants to look at them.

[popandbob]

I have three GTX280's with the V9 multi GPU version installed and it only uses two cards. Is there a setting I need to change or is this a limitation of CUDA/BOINC?
Thanks

It is not a limitation of boinc/cuda. Others have gotten such a setup to work.
Have you installed the drivers for each card via device manager?

[K Rider]

Yes, all three cards show up in device manager with the newest NV driver 182.06 installed. They are recognized by the SLI control panel and by GPU-Z.

[Mach]

Have you set the number of cpu's in the cc_config file as described in post http://lunatics.kwsn.net/gpu-crunching/modified-seti-mb-cuda-opt-ap-package-for-full-gpu-utilize.0.html ?