| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Beta Testing
Forum: Beta Test Support Forum
Thread: HCC GPU Beta Test 9/5/2012 Post problems here |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 96
|
|
| Author |
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
It is still doing it The stuttering and juddering of anything on the screen including the mouse movements A nvidia_hcc1 work unit renders a 6 Core 1055T and 4Gb of Ram with a GeForce GTX 560 Ti on Windows 7 Home Premium 64 bit version ... into a quivering wreck Surprisingly I got through relatively unscathed with HD 7770s (see my rig above). This was the first time I caught the betas in the act, and was expecting problems. But I was even able to edit some videos without seeing lag or audio sync problems, though it was still a short test and not exhaustive. But I often had lag/sync problems when I was doing Folding on an Nvidia display card (especially a GTX 560, which was worse than a GT 440 for some reason). So I may be able to use both my AMD HD 7770s for HCC instead of only the non-display card, as I had assumed. Part of it may be that Folding uses 99% of the GPU power on any of the the Nvidia cards I have used, whereas HCC used only about 85% on the 7770s. So I guess it varies a lot by the system. |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges:
|
Just noticed something unusual. I have 2 machines that are running other GPU apps but received none of the betas. I have a profile set to receive HCC tasks with the run CPU tasks option set to no. I would have thought that I would receive GPU tasks with that profile setting but I didn't. I have given up trying to even request WCG work while also running POEM on my cards. As soon as I heard about the beta, I suspended POEM and manually updated WCG. Voila! Got the work units, but that was at 3 AM EST (07:00 GMT) this morning. I think there is a bug in BOINC, and have some hope that the next version (after 7.0.33) may address it, but am not sure about that. EDIT: It started picking up the Betas last night while still running POEM, so it probably helps to have something in the pipeline for it to get. (But somehow the requests either don't happen, or else don't get recorded in the log properly when there is no work to be gotten.) [Edit 1 times, last edit by Jim1348 at Sep 6, 2012 10:50:57 AM] |
||
|
|
katoda
Senior Cruncher Poland Joined: Apr 28, 2007 Post Count: 170 Status: Offline Project Badges: 
|
Got three WU, errored with almost the same output as Ezrabynx's (https://secure.worldcommunitygrid.org/forums/wcg/viewpostinthread?post=390845 ), but on a machine running Windows XP , nVidia drivers version 301.42 and nVidia 8600 GTS.
---------------------------------------- [Edit 1 times, last edit by katoda at Sep 5, 2012 7:43:57 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Seem to be getting same error as some others where GPU work finishes but task sticks at 99.415%.
I've had 4 WUs all complete as far as 99.415% and then stop, left them to run for a while but after one had sat for almost an hour (gets to 99.415% in 5 mins) aborted. I'm running a Radeon HD5770 win 7 64 bit Intel i5 750 CPU. Stderr as follows: Commandline: projects/www.worldcommunitygrid.org/wcg_beta3_img_6.49_windows_intelx86__ati_hcc1 X0000038190794200409181620.jp2 --device 0 <app_init_data> <major_version>6</major_version> <minor_version>10</minor_version> <release>58</release> <app_version>649</app_version> <app_name>beta3</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>E:\Documents and Settings\WCGData/projects/www.worldcommunitygrid.org</project_dir> <boinc_dir>E:\Documents and Settings\WCGData</boinc_dir> <wu_name>BETA_X0000038190794200409181620</wu_name> <comm_obj_name>boinc_4</comm_obj_name> <slot>5</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>704232.610836</user_total_credit> <user_expavg_credit>1247.498782</user_expavg_credit> <host_total_credit>704017.800577</host_total_credit> <host_expavg_credit>1244.887542</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></gpu_type> <gpu_device_num>-1</gpu_device_num> <gpu_opencl_dev_index>-1</gpu_opencl_dev_index> <ncpus>0.000000</ncpus> <rsc_fpops_est>13814991649664.000000</rsc_fpops_est> <rsc_fpops_bound>276299832993280.000000</rsc_fpops_bound> <rsc_memory_bound>78643200.000000</rsc_memory_bound> <rsc_disk_bound>50000000.000000</rsc_disk_bound> <computation_deadline>1346990141.136000</computation_deadline> <vbox_window>0</vbox_window> </app_init_data> INFO: gpu_type not found in init_data.xml. INFO: GPU device not specified in init_data.xml. Checking Commandline. Boinc requested ATI gpu device number0 Found compute platform Advanced Micro Devices, Inc. Selecting this platform CL_DEVICE_NAME: Juniper 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: 850 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 Estimated kernel execution time = 0.29251 [sec] Starting analysis of ../../projects/www.worldcommunitygrid.org/BETA_X0000038190794200409181620_X0000038190794200409181620.jp2... Extracting GLCM features... Total kernel time: 201.400314 (1026 kernel executions) Total memory transfer time: 2.039096 Average kernel time: 0.196297 Min kernel time: 0.183528 (dx=3 dy=25 sample_dist=24 ) Max kernel time: 0.227449 dx=1 dy=2 sample_dist=1 |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
Just noticed something unusual. I have 2 machines that are running other GPU apps but received none of the betas. I have a profile set to receive HCC tasks with the run CPU tasks option set to no. I would have thought that I would receive GPU tasks with that profile setting but I didn't. I have given up trying to even request WCG work while also running POEM on my cards. As soon as I heard about the beta, I suspended POEM and manually updated WCG. Voila! Got the work units, but that was at 3 AM EST (07:00 GMT) this morning. I think there is a bug in BOINC, and have some hope that the next version (after 7.0.33) may address it, but am not sure about that. If I have a mix of POEM and WCG tasks running I get betas. If I'm running POEM only I get nada. 
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
  |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges: 
|
To those whose work units appear to be sticking at 99.415% give them chance before aborting
----------------------------------------It's seems to run on the cpu for approx 28 seconds which I think the cPU must be being used to upack the work unit - then the GPU get's its chance and then once this section is complete at 99.415% the cpu looks like it packs it up ready to be send across your internet connection this takes around 30 seconds. I can't find tilda on this linux keyboard - I have had to abandon the windoes box while it is completing the beta's Is this the case or are they really stuck  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi Guys,
----------------------------------------I have two machines with not very impressive GPUs that have, up until now, not received any GPU beta WUs because of their mediocrity. In the last few minutes both machines have started to get them, but they're failing almost immediately with "Computation error" and the message log says e.g.: 05/09/2012 21:44:53 World Community Grid Output file BETA_X0000038270039200409250029--201209051802130_3_0 for task BETA_X0000038270039200409250029--201209051802130_3 absent Now, even more interestingly, just before the first one of these on my lappie, I got a retread for a CFSW WU which claims it is "Running, high priority (0.25CPUs + 1.00 NVIDIA GPUs)". Surely that can't be right... Comments? More info: The failing GPU beta WU's are down to: Error: cl_khr_local_int32_base_atomics extension required by this program is not supported Like I said, mediocre GPUs. The CFSW WU finished normally. I haven't seen any more claiming to be using the GPU. [Edit 1 times, last edit by Former Member at Sep 5, 2012 9:09:47 PM] |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges:
|
... and another thing
---------------------------------------- On GPUGrid I can run 7 WU's on a 6 core machine. With HCC Beta's it seems to consume a whole CPU Core for the duration (maybe this is the source of the juddering) .. It also confuses the BOINC Scheduler on 6.10.58 many a DSFL WU that was running is now "waiting to run" and others running instead. This was running happliy before the Beta\s arrived |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges:
|
On GPUGrid I can run 7 WU's on a 6 core machine. With HCC Beta's it seems to consume a whole CPU Core for the duration (maybe this is the source of the juddering) That is more than likely your problem. You shouldn't try to run more tasks than you have CPU cores. If you're running 6 the GPU app will pause one while it crunches.
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
----------------------------------------[Edit 1 times, last edit by nanoprobe at Sep 5, 2012 8:56:16 PM] |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges:
|
keep 'em coming - my points score for the day will be approx 100,000 just from running the HCC Beta's - that's with a duty cycle on the GPU of about 80%
----------------------------------------When this goes live will the WU's be longer running ? [Edit 1 times, last edit by David Autumns at Sep 5, 2012 9:02:29 PM] |
||
|
|
|