| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: GPU Support Forum
Thread: GPU Support: Post Your Problems/Questions Here |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 195
|
|
| Author |
|
|
oldDirty
Cruncher Joined: Mar 10, 2009 Post Count: 21 Status: Offline Project Badges: 
|
And btw.: "its performance is comparably terrible in almost everything except heavily multi-threaded software." What is BOINC if not a heavily multithreaded workload? Multithreading is one thread using all cores ( like Ex Aqua@home ), in other way it's multitasking.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hey all,
Sorry if I've missed an answer to my question elsewhere; I've done a bit of digging in the forums but found nothing. I have recently started using my GPU for HCC WUs and after some trouble getting it working (although it was receiving WUs it would create a computational error within seconds of starting) involving driver updates I am now finding that all of the work I return is invalid; I am running an ATI radeon HD 6300m with an i5-2410m processor. The CPU is throttled by TThrottle to around 75 degrees and allowed to run 100% by boinc manager; I haven't noticed the GPU be throttled by TThrottle, only the CPU. If you need any more information just ask; I can post the log of an invalid result for example if that's of use. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Message log and Result Log are *Thé* number 1 useful pieces of information to troubleshoot, so what you're waiting for :)
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Result Log
Result Name: X0930115370185200911031316_ 1-- <core_client_version>7.0.28</core_client_version> <![CDATA[ <stderr_txt> Commandline: projects/www.worldcommunitygrid.org/wcg_hcc1_img_7.05_windows_intelx86__ati_hcc1 --zipfile X0930115370185200911031316.zip --imagelist images.txt --device 0 <app_init_data> <major_version>7</major_version> <minor_version>0</minor_version> <release>28</release> <app_version>705</app_version> <app_name>hcc1</app_name> <project_preferences> <color_scheme>Tahiti Sunset</color_scheme> <max_frames_sec>7</max_frames_sec> <max_gfx_cpu_pct>5.0</max_gfx_cpu_pct> </project_preferences> <project_dir>C:\ProgramData\BOINC/projects/www.worldcommunitygrid.org</project_dir> <boinc_dir>C:\ProgramData\BOINC</boinc_dir> <wu_name>X0930115370185200911031316</wu_name> <result_name>X0930115370185200911031316_1</result_name> <comm_obj_name>boinc_0</comm_obj_name> <slot>3</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>1414.803486</user_total_credit> <user_expavg_credit>186.908644</user_expavg_credit> <host_total_credit>1414.803486</host_total_credit> <host_expavg_credit>108.596592</host_expavg_credit> <resource_share_fraction>1.000000</resource_share_fraction> <checkpoint_period>60.000000</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>25684760337299.000000</rsc_fpops_est> <rsc_fpops_bound>1284238016864950.000000</rsc_fpops_bound> <rsc_memory_bound>78643200.000000</rsc_memory_bound> <rsc_disk_bound>50000000.000000</rsc_disk_bound> <computation_deadline>1362168243.000000</computation_deadline> <vbox_window>0</vbox_window> </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/X0930115370185200911031316_X0930115370185200911031316.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: Cedar CL_DEVICE_VENDOR: Advanced Micro Devices, Inc. CL_DEVICE_VERSION: CAL 1.4.1703 (VM) CL_DEVICE_MAX_COMPUTE_UNITS: CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS: 3 CL_DEVICE_MAX_WORK_ITEM_SIZES: 128 / 128 / 128 CL_DEVICE_MAX_WORK_GROUP_SIZE: 128 CL_DEVICE_MAX_CLOCK_FREQUENCY: 750 MHz CL_DEVICE_ADDRESS_BITS: 32 CL_DEVICE_MAX_MEM_ALLOC_SIZE: 512 MByte CL_DEVICE_GLOBAL_MEM_SIZE: 1024 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_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 cl_khr_d3d10_sharing cl_amd_meminfo Estimated kernel execution time = 0.60759 [sec] Starting analysis of X0930115370185200911031316.jp2... Extracting GLCM features... Total kernel time: 1299.918701 (1026 kernel executions) Total memory transfer time: 314.129700 Average kernel time: 1.266977 Min kernel time: 1.181261 (dx=23 dy=11 sample_dist=24 ) Max kernel time: 1.338824 dx=1 dy=1 sample_dist=0 INFO: GPU calculations complete. Total time for X0930115370185200911031316.jp2: 2377 seconds Finished Image #0, pctComplete = 0.500000 Starting analysis of X0930115370422200911031313.jp2... Extracting GLCM features... Total kernel time: 984.037292 (1026 kernel executions) Total memory transfer time: 319.048676 Average kernel time: 0.959101 Min kernel time: 0.886993 (dx=3 dy=25 sample_dist=24 ) Max kernel time: 1.054733 dx=3 dy=21 sample_dist=20 INFO: GPU calculations complete. Total time for X0930115370422200911031313.jp2: 1484 seconds Finished Image #1, pctComplete = 1.000000 CPU time used = 634.877270 14:20:58 (7628): called boinc_finish </stderr_txt> ]]> |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
This is the top of my message log, it runs since midnight this morning so I won't post it all unless you want more:
24/02/2013 00:34:05 | | Starting BOINC client version 7.0.28 for windows_x86_64 24/02/2013 00:34:05 | | log flags: file_xfer, sched_ops, task 24/02/2013 00:34:05 | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6 24/02/2013 00:34:05 | | Data directory: C:\ProgramData\BOINC 24/02/2013 00:34:05 | | Running under account Andy 24/02/2013 00:34:05 | | Processor: 4 GenuineIntel Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz [Family 6 Model 42 Stepping 7] 24/02/2013 00:34:05 | | Processor: 256.00 KB cache 24/02/2013 00:34:05 | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 syscall nx lm vmx tm2 popcnt pbe 24/02/2013 00:34:05 | | OS: Microsoft Windows 7: Home Premium x64 Edition, Service Pack 1, (06.01.7601.00) 24/02/2013 00:34:05 | | Memory: 5.98 GB physical, 11.96 GB virtual 24/02/2013 00:34:05 | | Disk: 654.69 GB total, 461.40 GB free 24/02/2013 00:34:05 | | Local time is UTC +0 hours 24/02/2013 00:34:05 | | ATI GPU 0: Cedar (CAL version 1.4.1703, 1024MB, 991MB available, 240 GFLOPS peak) 24/02/2013 00:34:05 | | OpenCL: ATI GPU 0: Cedar (driver version CAL 1.4.1703 (VM), device version OpenCL 1.1 AMD-APP (898.1), 1024MB, 991MB available) 24/02/2013 00:34:05 | World Community Grid | URL http://www.worldcommunitygrid.org/; Computer ID 2179996; resource share 100 24/02/2013 00:34:05 | World Community Grid | General prefs: from World Community Grid (last modified 01-Jan-1970 00:00:01) 24/02/2013 00:34:05 | World Community Grid | Host location: none 24/02/2013 00:34:05 | World Community Grid | General prefs: using your defaults 24/02/2013 00:34:05 | | Reading preferences override file 24/02/2013 00:34:05 | | Preferences: 24/02/2013 00:34:05 | | max memory usage when active: 3062.93MB 24/02/2013 00:34:05 | | max memory usage when idle: 4594.39MB 24/02/2013 00:34:05 | | max disk usage: 10.00GB 24/02/2013 00:34:05 | | suspend work if non-BOINC CPU load exceeds 75 % 24/02/2013 00:34:05 | | (to change preferences, visit the web site of an attached project, or select Preferences in the Manager) 24/02/2013 00:34:05 | | Not using a proxy 24/02/2013 00:34:06 | | Suspending computation - initial delay 24/02/2013 00:35:28 | World Community Grid | update requested by user 24/02/2013 00:35:31 | World Community Grid | Sending scheduler request: Requested by user. 24/02/2013 00:35:31 | World Community Grid | Reporting 1 completed tasks, not requesting new tasks 24/02/2013 00:35:35 | World Community Grid | Scheduler request completed 24/02/2013 00:36:05 | World Community Grid | Restarting task X0930116290760200912211840_1 using hcc1 version 705 (ati_hcc1) in slot 4 |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I have since changed my preferences so that boinc and tthrottle aren't doing the same job and boinc throttling the gpu unnecessarily (at least i think that makes sense) so cpu isbnow 100% on boinc manager
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
don't know where the problems are; but I have received over 200 repair jobs in 2 days now. wingman just erroring out
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
don't know where the problems are; but I have received over 200 repair jobs in 2 days now. wingman just erroring out Might have been me. I botched a driver update and probably lost hundreds.. or maybe a thousand. Enough to make the boinc manager not be able to connect to the client so i could stop things and minimize the damage.. oops:) @Andy: Try updating to the 7.0.52 client. I run that version on all of my systems right now and haven't had any problems, and I've noticed that GPU handling is getting better. Also, I've had problems when I try to get cute with my "integrated" graphics chip.. APU in my case. Try without any throttling and a modest 50% (or default) usage and see if you get valid results... then start pushing the limits:) |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Becay, thanks for your reply. I'll upgrade to 7.0.52 in the morning. I think I'd be correct in saying I'm running discrete graphics (potentially switchable to intel3000 which I know WCG won't utilise [or recognise?] but I won't claim to understand); Not meaning to be picky but what do you mean by 'get cute with'?
----------------------------------------Also can you confirm that you mean lose the throttling and set CPU time to 50% (or Core utilisation to 2/4 cores). Currently if I try running 3 CPU WUs and 1 CPU+GPU TThrottle cuts CPU (time?? Is that what it's cutting?) to 33% to keep heat in check. I think that to try to sort out my GPU issues I would be best just running the CPU and GPU WU on its own. Would that mean setting "On multiprocessors, use: " to 25% which would be one of the four cores and that core would be given to the CPU/GPU WU or could a CPU WU jump in when I'm not keeping an eye on things. If so would changing "Allow research to run on my CPU?" to NO allow CPU/GPU processing and not CPU or disallow both as GPU requires "feeding" from CPU Would it also be of use setting the GPU WUs to only be run when the computer is idle? Sorry for all of the questions, I'm trying to get my head around things by reading through other threads but I think it is just generating more questions. If anyone can answer any please dive in and don't wait for becay =) [Edit 1 times, last edit by Former Member at Feb 25, 2013 12:42:54 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Becay, thanks for your reply. I'll upgrade to 7.0.52 in the morning. I think I'd be correct in saying I'm running discrete graphics (potentially switchable to intel3000 which I know WCG won't utilise [or recognise?] but I won't claim to understand); Not meaning to be picky but what do you mean by 'get cute with'? In the newer versions of the Boinc client (I want to say after 1.0.40, I'm not sure) you are able to do things like run 0.25 cpu and 0.125 gpu per WU. I do this on my desktops to better utilize my hardware, but trying this on my laptop caused about 25% of WUs to error at some point, for some reason. Any gains I might have gotten with the successes were lost on the failures. Worth a shot:) Also can you confirm that you mean lose the throttling and set CPU time to 50% (or Core utilisation to 2/4 cores). 50% CPU time, 100% Cores. Defaults are defaults for a reason:) Currently if I try running 3 CPU WUs and 1 CPU+GPU TThrottle cuts CPU (time?? Is that what it's cutting?) to 33% to keep heat in check. I think that to try to sort out my GPU issues I would be best just running the CPU and GPU WU on its own. Would that mean setting "On multiprocessors, use: " to 25% which would be one of the four cores and that core would be given to the CPU/GPU WU or could a CPU WU jump in when I'm not keeping an eye on things. If so would changing "Allow research to run on my CPU?" to NO allow CPU/GPU processing and not CPU or disallow both as GPU requires "feeding" from CPU Would it also be of use setting the GPU WUs to only be run when the computer is idle? You might be over thinking this:) Start at the minimum then work yourself up. Manage Boinc through the boinc manager. It does a pretty good job... or at least 7.0.52 does. Sorry for all of the questions, I'm trying to get my head around things by reading through other threads but I think it is just generating more questions. If anyone can answer any please dive in and don't wait for becay =) Every machine is different. Mileage will vary... especially if you plan on actually using that machine for day-to-day stuff:) You'll have to play with a few configurations and see what works for you. |
||
|
|
|