Author Topic: Nvidia Titan error on Astropulse (Read 21097 times)

Pizzadude · « **on:** 15 Mar 2013, 03:28:02 pm »

I am having trouble getting my new Titan card to run Astropulse it seems to get into a repeating loop and does not do any work on the unit. Here is the error and my setup - any ideas ?

15/03/2013 19:22:22 | SETI@home | Starting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11
15/03/2013 19:22:32 | SETI@home | Task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 exited with zero status but no 'finished' file
15/03/2013 19:22:32 | SETI@home | If this happens repeatedly you may need to reset the project.
15/03/2013 19:22:32 | SETI@home | Restarting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11
15/03/2013 19:22:32 | SETI@home | Starting task ap_04au12ac_B2_P0_00150_20130314_08598.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 12
15/03/2013 19:22:32 | SETI@home | Starting task ap_04au12ac_B1_P1_00305_20130314_07459.wu_0 using astropulse_v6 version 604 (cuda_opencl_100) in slot 13
15/03/2013 19:22:36 | SETI@home | Task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 exited with zero status but no 'finished' file
15/03/2013 19:22:36 | SETI@home | If this happens repeatedly you may need to reset the project.
15/03/2013 19:22:36 | SETI@home | Restarting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11
15/03/2013 19:22:36 | SETI@home | Starting task 20dc11aa.17029.18577.206158430221.10.84_1 using setiathome_enhanced version 610 (cuda_fermi) in slot 14
15/03/2013 19:22:36 | SETI@home | Starting task 20dc11aa.17029.18577.206158430221.10.145_1 using setiathome_enhanced version 610 (cuda_fermi) in slot 15
15/03/2013 19:22:41 | SETI@home | Task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 exited with zero status but no 'finished' file
15/03/2013 19:22:41 | SETI@home | If this happens repeatedly you may need to reset the project.
15/03/2013 19:22:41 | SETI@home | Restarting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11
15/03/2013 19:22:46 | SETI@home | Task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 exited with zero status but no 'finished' file
15/03/2013 19:22:46 | SETI@home | If this happens repeatedly you may need to reset the project.
15/03/2013 19:22:46 | SETI@home | Restarting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11
15/03/2013 19:22:49 | SETI@home | Starting task 16fe12ad.14612.366667.206158430221.10.4_0 using setiathome_enhanced version 610 (cuda_fermi) in slot 16
15/03/2013 19:22:49 | SETI@home | Starting task 20dc11aa.17029.18577.206158430221.10.167_0 using setiathome_enhanced version 610 (cuda_fermi) in slot 17
15/03/2013 19:22:51 | SETI@home | Task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 exited with zero status but no 'finished' file
15/03/2013 19:22:51 | SETI@home | If this happens repeatedly you may need to reset the project.
15/03/2013 19:22:51 | SETI@home | Restarting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11
15/03/2013 19:22:56 | SETI@home | Task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 exited with zero status but no 'finished' file
15/03/2013 19:22:56 | SETI@home | If this happens repeatedly you may need to reset the project.
15/03/2013 19:22:56 | SETI@home | Restarting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11
15/03/2013 19:23:01 | SETI@home | Task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 exited with zero status but no 'finished' file
15/03/2013 19:23:01 | SETI@home | If this happens repeatedly you may need to reset the project.
15/03/2013 19:23:01 | SETI@home | Restarting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11
15/03/2013 19:23:07 | SETI@home | Task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 exited with zero status but no 'finished' file
15/03/2013 19:23:07 | SETI@home | If this happens repeatedly you may need to reset the project.
15/03/2013 19:23:07 | SETI@home | Restarting task ap_04au12ac_B2_P1_00073_20130314_11177.wu_1 using astropulse_v6 version 604 (cuda_opencl_100) in slot 11

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8b2_win_x64_SSE3.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
   <platform>windows_intelx86</platform>
<file_ref>
<file_name>AK_v8b2_win_x64_SSE3.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
   <platform>windows_x86_64</platform>
<file_ref>
<file_name>AK_v8b2_win_x64_SSE3.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>astropulse_v6</name>
</app>
<file_info>
<name>AP6_win_x86_SSE2_OpenCL_NV_r1761.exe</name>
<executable/>
</file_info>
<file_info>
<name>libfftw3f-3.dll</name>
<executable/>
</file_info>
<file_info>
<name>ap_cmdline_win_x86_SSE2_OpenCL_NV.txt</name>
</file_info>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>604</version_num>
<platform>windows_intelx86</platform>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<plan_class>cuda_opencl_100</plan_class>
   <cmdline>-ffa_block 6144 -ffa_block_fetch 1536 -unroll 10</cmdline>
<coproc>
<type>CUDA</type>
<count>0.3</count>
</coproc>
<file_ref>
<file_name>AP6_win_x86_SSE2_OpenCL_NV_r1761.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>libfftw3f-3.dll</file_name>
</file_ref>
<file_ref>
<file_name>ap_cmdline_win_x86_SSE2_OpenCL_NV.txt</file_name>
<open_name>ap_cmdline.txt</open_name>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v6</app_name>
<version_num>604</version_num>
<platform>windows_intelx86</platform>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<plan_class>opencl_nvidia_100</plan_class>
   <cmdline>-ffa_block 6144 -ffa_block_fetch 1536 -unroll 10</cmdline>
<coproc>
<type>CUDA</type>
<count>0.3</count>
</coproc>
<file_ref>
<file_name>AP6_win_x86_SSE2_OpenCL_NV_r1761.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>libfftw3f-3.dll</file_name>
</file_ref>
<file_ref>
<file_name>ap_cmdline_win_x86_SSE2_OpenCL_NV.txt</file_name>
<open_name>ap_cmdline.txt</open_name>
</file_ref>
</app_version>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>Lunatics_x41g_win32_cuda32.exe</name>
<executable/>
</file_info>
<file_info>
<name>cudart32_32_16.dll</name>
<executable/>
</file_info>
<file_info>
<name>cufft32_32_16.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>610</version_num>
   <platform>windows_intelx86</platform>
<plan_class>cuda_fermi</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>0.3</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
     <file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>609</version_num>
   <platform>windows_intelx86</platform>
<plan_class>cuda23</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>0.3</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
     <file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
   <platform>windows_intelx86</platform>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>0.3</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
     <file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>610</version_num>
   <platform>windows_x86_64</platform>
<plan_class>cuda_fermi</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>0.3</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
     <file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>609</version_num>
   <platform>windows_x86_64</platform>
<plan_class>cuda23</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>0.3</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
     <file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
   <platform>windows_x86_64</platform>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>0.3</count>
</coproc>
<file_ref>
<file_name>Lunatics_x41g_win32_cuda32.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
     <file_name>cudart32_32_16.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft32_32_16.dll</file_name>
</file_ref>
</app_version>
</app_info>

Raistmer · « **Reply #1 on:** 15 Mar 2013, 03:36:15 pm »

stderr?

Claggy · « **Reply #2 on:** 15 Mar 2013, 03:39:31 pm »

Quote from: Pizzadude on 15 Mar 2013, 03:28:02 pm

I am having trouble getting my new Titan card to run Astropulse it seems to get into a repeating loop and does not do any work on the unit. Here is the error and my setup - any ideas ?

How many CPU cores are you reserving? you're got to reserve at least one or two cores.

Have you tried running only one instance at a time?, rather than three.

If you remove the cmd line parameters from the app_info, any improvement?
(put them in the ap_cmdline_win_x86_SSE2_OpenCL_NV.txt file instead, then you can make changes without restarting Boinc)

Can we have a link to the host too, please.

Claggy

Claggy · « **Reply #3 on:** 15 Mar 2013, 04:00:13 pm »

Quote from: Pizzadude on 15 Mar 2013, 03:28:02 pm

<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>Lunatics_x41g_win32_cuda32.exe</name>
<executable/>
</file_info>
<file_info>
<name>cudart32_32_16.dll</name>
<executable/>
</file_info>
<file_info>
<name>cufft32_32_16.dll</name>
<executable/>
</file_info>

While you're at it, you might want to upgrade to x41zc_cuda5 from x41g_cuda32, x41zc is more Optimised for Fermi's and especially Kepler's,

Claggy

Pizzadude · « **Reply #4 on:** 15 Mar 2013, 04:20:42 pm »

There are 4 cores free on the CPU.

Running one instance, two or three makes no difference.
Removed cmdline parameters but still the same.
Inserted parameters into ap_cmdline_win_x86_SSE2_OpenCL_NV.txt, but still getting the same issue.

http://setiathome.berkeley.edu/results.php?userid=15792

Quote from: Claggy on 15 Mar 2013, 03:39:31 pm

Quote from: Pizzadude on 15 Mar 2013, 03:28:02 pm
I am having trouble getting my new Titan card to run Astropulse it seems to get into a repeating loop and does not do any work on the unit. Here is the error and my setup - any ideas ?
How many CPU cores are you reserving? you're got to reserve at least one or two cores.

Have you tried running only one instance at a time?, rather than three.

If you remove the cmd line parameters from the app_info, any improvement?
(put them in the ap_cmdline_win_x86_SSE2_OpenCL_NV.txt file instead, then you can make changes without restarting Boinc)

Can we have a link to the host too, please.

Claggy

Claggy · « **Reply #5 on:** 15 Mar 2013, 04:39:38 pm »

Quote from: Pizzadude on 15 Mar 2013, 04:20:42 pm

http://setiathome.berkeley.edu/results.php?userid=15792

That is a url only you can view, the rest of us get:

Unable to handle request

No access

This one works:

http://setiathome.berkeley.edu/show_user.php?userid=15792

Even better, this one goes straight to your host:

http://setiathome.berkeley.edu/show_host_detail.php?hostid=6758290

Claggy

Claggy · « **Reply #6 on:** 15 Mar 2013, 04:52:28 pm »

On one Wu, you were getting the following a couple of times:

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: OpenCL kernel/call 'ReadBuf:gpu_results->CPU_result' call failed (-5) in file ..\..\ap_client_main.cpp near line 3238.

then:

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: WriteBuffer(gpu_data): -4
Error in ap oclFFT_1: -4
ERROR: clFFT_Execute: -4
ERROR: Enqueueing kernel onto command queue. (dechirp_range_kernel)
ERROR code: CL_MEM_OBJECT_ALLOCATION_FAILURE

On at least two other Wu's, it was:

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: clEnqueueNDRangeKernel: GPU_coadd_kernel_cl(inner): -4

You can try just running Nvidia Astropulse on it's own (don't let any Cuda Wu's run from any project) and see if it's the same.

Claggy

Pizzadude · « **Reply #7 on:** 15 Mar 2013, 04:57:42 pm »

Sorry, I sat here looking at the link thinking it did not look right - but it did not click with me what was wrong

Quote from: Claggy on 15 Mar 2013, 04:39:38 pm

Quote from: Pizzadude on 15 Mar 2013, 04:20:42 pm

http://setiathome.berkeley.edu/results.php?userid=15792
That is a url only you can view, the rest of us get:

Unable to handle request

No access

This one works:

http://setiathome.berkeley.edu/show_user.php?userid=15792

Even better, this one goes straight to your host:

http://setiathome.berkeley.edu/show_host_detail.php?hostid=6758290

Claggy

Raistmer · « **Reply #8 on:** 15 Mar 2013, 06:01:36 pm »

It's second time I get reports about AP not functioning on TITAN NV GPU.
Perhaps, some incompatibility there. I have no hardware to check.

Pizzadude · « **Reply #9 on:** 15 Mar 2013, 06:57:00 pm »

Ok I have setup the machine to run nothing but Astropulse and downloaded fresh work, however this has made no difference all work units are failing.

Quote from: Claggy on 15 Mar 2013, 04:52:28 pm

On one Wu, you were getting the following a couple of times:

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: OpenCL kernel/call 'ReadBuf:gpu_results->CPU_result' call failed (-5) in file ..\..\ap_client_main.cpp near line 3238.

then:

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: WriteBuffer(gpu_data): -4
Error in ap oclFFT_1: -4
ERROR: clFFT_Execute: -4
ERROR: Enqueueing kernel onto command queue. (dechirp_range_kernel)
ERROR code: CL_MEM_OBJECT_ALLOCATION_FAILURE

On at least two other Wu's, it was:

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: clEnqueueNDRangeKernel: GPU_coadd_kernel_cl(inner): -4

You can try just running Nvidia Astropulse on it's own (don't let any Cuda Wu's run from any project) and see if it's the same.

Claggy

Claggy · « **Reply #10 on:** 16 Mar 2013, 06:44:28 am »

For a long shot, you could try the compatibility grid setting laid out in the Sticky post in Number crunching, I doubt it'll work, but worth a try.

Claggy

Pizzadude · « **Reply #11 on:** 16 Mar 2013, 08:39:56 am »

Already tried that last night out of desperation, but it makes no difference. I am going back to CPU until the problem gets sorted.

Quote from: Claggy on 16 Mar 2013, 06:44:28 am

For a long shot, you could try the compatibility grid setting laid out in the Sticky post in Number crunching, I doubt it'll work, but worth a try.

Claggy

Claggy · « **Reply #12 on:** 16 Mar 2013, 08:43:53 am »

Can you post the ouput from CLinfo please

http://boinc.berkeley.edu/dl/clinfo.zip

Claggy

Raistmer · « **Reply #13 on:** 16 Mar 2013, 04:13:10 pm »

Quote from: Pizzadude on 16 Mar 2013, 08:39:56 am

Already tried that last night out of desperation, but it makes no difference. I am going back to CPU until the problem gets sorted.

Quote from: Claggy on 16 Mar 2013, 06:44:28 am
For a long shot, you could try the compatibility grid setting laid out in the Sticky post in Number crunching, I doubt it'll work, but worth a try.

Claggy

Cause I have no TITAN hardware I'm very doubt that problem will be ever solved w/o active testers.

Pizzadude · « **Reply #14 on:** 16 Mar 2013, 09:44:35 pm »

Updating to driver 314.21(beta) seems to allow astropulse processing, however as this is a beta driver that has been out less than 24 hours we will have to wait and see if there are any downsides to this step forward

Quote from: Raistmer on 16 Mar 2013, 04:13:10 pm

Quote from: Pizzadude on 16 Mar 2013, 08:39:56 am
Already tried that last night out of desperation, but it makes no difference. I am going back to CPU until the problem gets sorted.

Quote from: Claggy on 16 Mar 2013, 06:44:28 am
For a long shot, you could try the compatibility grid setting laid out in the Sticky post in Number crunching, I doubt it'll work, but worth a try.

Claggy
Cause I have no TITAN hardware I'm very doubt that problem will be ever solved w/o active testers.

Author Topic: Nvidia Titan error on Astropulse (Read 21097 times)

Pizzadude

Nvidia Titan error on Astropulse

Raistmer

Re: Nvidia Titan error on Astropulse

Claggy

Re: Nvidia Titan error on Astropulse

Claggy

Re: Nvidia Titan error on Astropulse

Pizzadude

Re: Nvidia Titan error on Astropulse

Claggy

Re: Nvidia Titan error on Astropulse

Claggy

Re: Nvidia Titan error on Astropulse

Pizzadude

Re: Nvidia Titan error on Astropulse

Raistmer

Re: Nvidia Titan error on Astropulse

Pizzadude

Re: Nvidia Titan error on Astropulse

Claggy

Re: Nvidia Titan error on Astropulse

Pizzadude

Re: Nvidia Titan error on Astropulse

Claggy

Re: Nvidia Titan error on Astropulse

Raistmer

Re: Nvidia Titan error on Astropulse

Pizzadude

Re: Nvidia Titan error on Astropulse