| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: GPU Support Forum
Thread: Debian wheezy amd64 on ATI: which packages do I need to run 32-bit workunits successfully? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 8
|
|
| Author |
|
|
rww
Cruncher Joined: Nov 14, 2009 Post Count: 4 Status: Offline Project Badges: 
|
Howdy,
I'm using Debian testing/wheezy, amd64/x86_64 architecture, on a computer with an ATI Radeon HD 7870 GPU. The vast majority of my GPU tasks complete successfully. However, a small number of them fail with the following error: <core_client_version>7.0.27</core_client_version> <![CDATA[ <message> process exited with code 127 (0x7f, -129) </message> <stderr_txt> ../../projects/www.worldcommunitygrid.org/wcg_hcc1_img_7.08_i686-pc-linux-gnu__ati_hcc1: error while loading shared libraries: libOpenCL.so.1: cannot open shared object file: No such file or directory </stderr_txt> ]]> I assume that this is happening on 32-bit work units because my system only has 64-bit libraries installed, not 32-bit ones. I am familiar with how to use multiarch on Debian testing. Does someone have a complete list of i686 packages I need to install to get HCC i686 GPU workunits to process correctly? Additionally (and less importantly), if someone has a complete list of i686 packages I need to install to get all other active World Community Grid projects that might send me 32-bit work units to process correctly, that would be excellent. Thanks! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
|
||
|
|
rww
Cruncher Joined: Nov 14, 2009 Post Count: 4 Status: Offline Project Badges:
|
Update: I installed the i386 version of amd-libopencl1, to sort out the error I was getting, and am now getting this one instead. I'm not sure what VerifyGPU.cpp is or where I can get a copy of it to see what's going on on line 114 :\
<core_client_version>7.0.27</core_client_version> <![CDATA[ <message> process exited with code 231 (0xe7, -25) </message> <stderr_txt> Commandline: ../../projects/www.worldcommunitygrid.org/wcg_hcc1_img_7.08_i686-pc-linux-gnu__ati_hcc1 --zipfile X0900092891432200709250952.zip --imagelist images.txt --device 0 <app_init_data> <major_version>7</major_version> <minor_version>0</minor_version> <release>27</release> <app_version>708</app_version> <app_name>hcc1</app_name> <project_preferences> <color_scheme>Tahiti Sunset</color_scheme> <max_frames_sec>3</max_frames_sec> <max_gfx_cpu_pct>5.0</max_gfx_cpu_pct> </project_preferences> <project_dir>/var/lib/boinc-client/projects/www.worldcommunitygrid.org</project_dir> <boinc_dir>/var/lib/boinc-client</boinc_dir> <wu_name>X0900092891432200709250952</wu_name> <result_name>X0900092891432200709250952_0</result_name> <shm_key>-1</shm_key> <slot>0</slot> <wu_cpu_time>0.000000</wu_cpu_time> <starting_elapsed_time>0.000000</starting_elapsed_time> <using_sandbox>0</using_sandbox> <user_total_credit>1128786.368093</user_total_credit> <user_expavg_credit>11634.489508</user_expavg_credit> <host_total_credit>187706.337527</host_total_credit> <host_expavg_credit>11580.675395</host_expavg_credit> <resource_share_fraction>1.000000</resource_share_fraction> <checkpoint_period>600.000000</checkpoint_period> <fraction_done_start>0.000000</fraction_done_start> <fraction_done_end>1.000000</fraction_done_end> <gpu_type>ATI</gpu_type> <gpu_device_num>0</gpu_device_num> <gpu_opencl_dev_index>0</gpu_opencl_dev_index> <ncpus>1.000000</ncpus> <rsc_fpops_est>25414572031581.000000</rsc_fpops_est> <rsc_fpops_bound>508291440631620.000000</rsc_fpops_bound> <rsc_memory_bound>78643200.000000</rsc_memory_bound> <rsc_disk_bound>50000000.000000</rsc_disk_bound> <computation_deadline>1357934408.000000</computation_deadline> </app_init_data> INFO: gpu_type set in init_data.xml to ATI INFO: gpu_device_num set in init_data.xml to 0 Boinc requested ATI gpu device number0 Unzipping input images ../../projects/www.worldcommunitygrid.org/X0900092891432200709250952_X0900092891432200709250952.zip Processing jobdescription Number of Images defined in image list is 2 Found compute platform Advanced Micro Devices, Inc. Selecting this platform ERROR: VerifyGPU.cpp:114 Device not found. 18:37:10 (11023): called boinc_finish </stderr_txt> ]]> |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello rww,
----------------------------------------Update: I installed the i386 version of amd-libopencl1, to sort out the error I was getting, and am now getting this one instead. I'm not sure what VerifyGPU.cpp is or where I can get a copy of it to see what's going on on line 114 I am not current on Linux, but Linus Torvalds recently celebrated the removal of archaic 386 support from the Linux kernel. I think you should look for a more advanced library.Lawrence Added after rww replied:  Good explanation. Sorry I can't help.[Edit 1 times, last edit by Former Member at Jan 6, 2013 3:02:29 PM] |
||
|
|
rww
Cruncher Joined: Nov 14, 2009 Post Count: 4 Status: Offline Project Badges:
|
I am not current on Linux, but Linus Torvalds recently celebrated the removal of archaic 386 support from the Linux kernel. I think you should look for a more advanced library. Debian uses "i386" to refer to the 32-bit x86 architecture. This is an increasingly-inaccurate name (Debian itself hasn't supported i386 itself for a while), but it is still the name it uses ;) [Edit 1 times, last edit by rww at Jan 6, 2013 8:42:55 AM] |
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: 
|
That error occurs when it tries to get a list of the GPU cards available. Not sure why it is failing on your machine but I have seen other posts in the forums of users periodically experiencing the same issue. It could be related to the version of the OpenCL 32 bit library you installed. I am looking into modifying the code on how it finds the right GPU to compute on, at the same time I will look into creating a 64 bit version of the gpu app so the 32 bit libraries are not necessary.
Thanks, armstrdj |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
[rhetorical] Was there not already a 64 bit HCC-CPU for Linux to serve the "pure" 64 bit systems :D... The GPU ones that came to my Linux 64 sys kept borging with one of those VerifyGPU.cpp errors, though it happily processes a mix of 32 and 64 bit WCG CPU sciences sans issue...
----------------------------------------Several months into HCC GPU crunching and not succeeded to complete a single HCC-GPU task on my fringe capable GT220, though it is [still] not listed as banished. This is of course no surprise as it was posted that having a 10-15 percent success rate for a card version is enough to not lock out these borderline cards. Fair enough, but if users don't stop these coming you could have long serial WU fail generation being funneled back to reliable hosts. Not causing bottlenecks? [Edit 1 times, last edit by Former Member at Jan 7, 2013 5:14:11 PM] |
||
|
|
rww
Cruncher Joined: Nov 14, 2009 Post Count: 4 Status: Offline Project Badges:
|
That error occurs when it tries to get a list of the GPU cards available. Not sure why it is failing on your machine but I have seen other posts in the forums of users periodically experiencing the same issue. It could be related to the version of the OpenCL 32 bit library you installed. According to my package manager, that's 1:12-6+point-1, which seems to be sourced from fglrx (AMD driver) "internal version 8.982.13" (Debian changelog). Debian tends to be rather slow at getting new stuff, especially right now (wheezy's frozen for release some time this year (hopefully)), so that could very well be it. I might spin up an Ubuntu install at some point and see if it's any better. I am looking into modifying the code on how it finds the right GPU to compute on, at the same time I will look into creating a 64 bit version of the gpu app so the 32 bit libraries are not necessary. I assumed that that already exists. I have a wcg_hcc1_img_7.08_x86_64-pc-linux-gnu__ati_hcc1 in my projects/www.worldcommunitygrid.org directory, and it seems to be used for the vast majority of my workunits (which, as I said, tend to compute just fine). Edit: speaking of which, an example run with the x86_64 version, in case it helps you or people working on this in the future: <core_client_version>7.0.27</core_client_version> <![CDATA[ <stderr_txt> Commandline: ../../projects/www.worldcommunitygrid.org/wcg_hcc1_img_7.08_x86_64-pc-linux-gnu__ati_hcc1 --zipfile X0960097551428200803041732.zip --imagelist images.txt --device 0 <app_init_data> <major_version>7</major_version> <minor_version>0</minor_version> <release>27</release> <app_version>708</app_version> <app_name>hcc1</app_name> <project_preferences> <color_scheme>Tahiti Sunset</color_scheme> <max_frames_sec>3</max_frames_sec> <max_gfx_cpu_pct>5.0</max_gfx_cpu_pct> </project_preferences> <project_dir>/var/lib/boinc-client/projects/www.worldcommunitygrid.org</project_dir> <boinc_dir>/var/lib/boinc-client</boinc_dir> <wu_name>X0960097551428200803041732</wu_name> <result_name>X0960097551428200803041732_0</result_name> <shm_key>-1</shm_key> <slot>0</slot> <wu_cpu_time>0.000000</wu_cpu_time> <starting_elapsed_time>0.000000</starting_elapsed_time> <using_sandbox>0</using_sandbox> <user_total_credit>1210985.095084</user_total_credit> <user_expavg_credit>15331.728483</user_expavg_credit> <host_total_credit>269905.064518</host_total_credit> <host_expavg_credit>15282.447640</host_expavg_credit> <resource_share_fraction>1.000000</resource_share_fraction> <checkpoint_period>600.000000</checkpoint_period> <fraction_done_start>0.000000</fraction_done_start> <fraction_done_end>1.000000</fraction_done_end> <gpu_type>ATI</gpu_type> <gpu_device_num>0</gpu_device_num> <gpu_opencl_dev_index>0</gpu_opencl_dev_index> <ncpus>1.000000</ncpus> <rsc_fpops_est>25423512781853.000000</rsc_fpops_est> <rsc_fpops_bound>508470255637060.000000</rsc_fpops_bound> <rsc_memory_bound>78643200.000000</rsc_memory_bound> <rsc_disk_bound>50000000.000000</rsc_disk_bound> <computation_deadline>1358186408.000000</computation_deadline> </app_init_data> INFO: gpu_type set in init_data.xml to ATI INFO: gpu_device_num set in init_data.xml to 0 Boinc requested ATI gpu device number0 Unzipping input images ../../projects/www.worldcommunitygrid.org/X0960097551428200803041732_X0960097551428200803041732.zip Processing jobdescription Number of Images defined in image list is 2 Found compute platform Advanced Micro Devices, Inc. Selecting this platform CL_DEVICE_NAME: Pitcairn CL_DEVICE_VENDOR: Advanced Micro Devices, Inc. CL_DEVICE_VERSION: CAL 1.4.1741 (VM) CL_DEVICE_MAX_COMPUTE_UNITS: CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS: 3 CL_DEVICE_MAX_WORK_ITEM_SIZES: 256 / 256 / 256 CL_DEVICE_MAX_WORK_GROUP_SIZE: 256 CL_DEVICE_MAX_CLOCK_FREQUENCY: 0 MHz CL_DEVICE_ADDRESS_BITS: 32 CL_DEVICE_MAX_MEM_ALLOC_SIZE: 500 MByte CL_DEVICE_GLOBAL_MEM_SIZE: 1810 MByte CL_DEVICE_ERROR_CORRECTION_SUPPORT: no CL_DEVICE_LOCAL_MEM_TYPE: local CL_DEVICE_LOCAL_MEM_SIZE: 32 KByte CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE: 64 KByte CL_DEVICE_QUEUE_PROPERTIES: CL_QUEUE_PROFILING_ENABLE CL_DEVICE_EXTENSIONS: cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_gl_sharing cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_popcnt Estimated kernel execution time = 0.22372 [sec] Starting analysis of X0960097551428200803041732.jp2... Extracting GLCM features... Total kernel time: 27.737017 (1026 kernel executions) Total memory transfer time: 4.155187 Average kernel time: 0.027034 Min kernel time: 0.025741 (dx=5 dy=25 sample_dist=24 ) Max kernel time: 0.028758 dx=1 dy=1 sample_dist=0 INFO: GPU calculations complete. Total time for X0960097551428200803041732.jp2: 91 seconds Finished Image #0, pctComplete = 0.500000 Starting analysis of X0960097550748200803041742.jp2... Extracting GLCM features... Total kernel time: 26.307125 (1026 kernel executions) Total memory transfer time: 8.295090 Average kernel time: 0.025640 Min kernel time: 0.024491 (dx=25 dy=5 sample_dist=24 ) Max kernel time: 0.027336 dx=1 dy=1 sample_dist=0 INFO: GPU calculations complete. Total time for X0960097550748200803041742.jp2: 88 seconds Finished Image #1, pctComplete = 1.000000 CPU time used = 116.227263 00:41:40 (4462): called boinc_finish </stderr_txt> ]]> Thanks for your insights  [Edit 1 times, last edit by rww at Jan 8, 2013 2:08:44 PM] |
||
|
|
|