| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: GPU Support Forum
Thread: GPU Work Units - Post Your Tech Support Questions Here |
| No member browsing this thread |
|
Thread Status: Active Thread Type: Sticky Thread Total posts in this thread: 290
|
|
| Author |
|
|
Hal Bregg
Cruncher Joined: Nov 20, 2018 Post Count: 5 Status: Offline Project Badges: 
|
From an earlier post in this thread: <use_all_gpus>1</use_all_gpus> in cc_config.xml Paul. I would like to use Nvidia card not Intel. I'm getting WUs for Intel GPU only. |
||
|
|
leifeinar@yahoo.no
Cruncher Joined: Aug 26, 2006 Post Count: 13 Status: Offline Project Badges: 
|
Havent got anny Gpu work for a day now, is the test over?
|
||
|
|
caitilarkin
Former World Community Grid Admin USA Joined: Nov 4, 2015 Post Count: 331 Status: Offline Project Badges: 
|
Havent got anny Gpu work for a day now, is the test over? Yes, the last work units were sent out on Tuesday. WCG tech team member kneed made an announcement in the GPU Stress Test thread. |
||
|
|
goben_2003
Advanced Cruncher Joined: Jun 16, 2006 Post Count: 145 Status: Offline Project Badges: 
|
Hi, during the stress test an issue with the WCG scheduler was noticed. I made a thread about it here so it is easier to discuss it without being mixed with other issues.
---------------------------------------- |
||
|
|
jay_Orlando
Senior Cruncher USA Joined: Jan 4, 2006 Post Count: 181 Status: Offline Project Badges:
|
Hey there!
----------------------------------------A couple of questions, please. 1) How do I tell if I have processed any GPU WU?? I did not see a distinction in My Contribution. 2) I have an (older) AMD 7750 GPU - bit it has 2 gig of memory - and is crunching EinsteinAtHom3 GPU WU successfully. I run linux - Ubuntu with an OpenCL driver but it is 1.1. This probably the problem. Any suggestions on how to get 2.1 - which package? Thanks!! Jay  |
||
|
|
goben_2003
Advanced Cruncher Joined: Jun 16, 2006 Post Count: 145 Status: Offline Project Badges:
|
Hey there! A couple of questions, please. 1) How do I tell if I have processed any GPU WU?? I did not see a distinction in My Contribution. 2) I have an (older) AMD 7750 GPU - bit it has 2 gig of memory - and is crunching EinsteinAtHom3 GPU WU successfully. I run linux - Ubuntu with an OpenCL driver but it is 1.1. This probably the problem. Any suggestions on how to get 2.1 - which package? Thanks!! Jay For 1: You can go to results status and filter by open pandemics. The result names that start with OPN1 are CPU, the ones that start with OPNG are GPU. It is tedious. I use BoincTasks, so I often look in it instead of on the results status page. |
||
|
|
PMH_UK
Veteran Cruncher UK Joined: Apr 26, 2007 Post Count: 766 Status: Offline Project Badges: 
|
For the future consider running WUProp.
----------------------------------------It captures elapsed run time by project and OPNG is separate entry. Paul.
Paul.
|
||
|
|
jay_Orlando
Senior Cruncher USA Joined: Jan 4, 2006 Post Count: 181 Status: Offline Project Badges:
|
An Update...
----------------------------------------I am trying to get the OpenCL 2.1 driver. 1.1 is the default for Ubuntu 20.4 - which I have. AMD says they will only support the Long Term Support builds. I read here about instructions to install OpenCL 2.1: https://rocmdocs.amd.com/en/latest/Installation_Guide/Installation-Guide.html One step was: sudo apt install rocm-dkms && sudo reboot But, no rocm-dkms package exists. Any suggestions? T H A N K S in advance!!! Jay |
||
|
|
jay_Orlando
Senior Cruncher USA Joined: Jan 4, 2006 Post Count: 181 Status: Offline Project Badges:
|
3 Updates:
----------------------------------------1) I tried the rocm update. A No-go. Although AMD says they support the LTS, they do not support an updated version of the 20.04 LTS of Ubuntu. I also tried to downgrade my updated 20.04 to earlier linux versions. That completely broke my system. 2) I wiped out / and /var, saving /home and installed Ubuntu 21.04 in the hopes that the Mesa OpenCL driver might be updated to 1.1. Nope. It is still at 1.1 3) The AMD instructions at https://rocmdocs.amd.com/en/latest/Installation_Guide/Installation-Guide.html were, to me misleading. The info said that Ubuntu LTS only supported Ubuntu LTS; then, had a Section Header for Debian. I *assumed* the The Debian section was mutually exclusive. Wrong. The sources of the post said that you had to add lines on your /etc/apt/sources.lst That action merged Debian xenial with Ubuntu. I tried it. See #1 above. [edit] in the above post, I did not find the rocm-dkms package. That is because it is found in the Debian repository. So. No WCG/Mip1 running on GPU for me - unless someone knows a linux distrubution that comes with an integrated OpenCL 2.1 Any info on this? Thanks in advance, Jay PS Meanwhile, I'll explore. But, if you have been there, sharing info would definitely help. [Edit 1 times, last edit by jay_Orlando at May 11, 2021 5:44:32 PM] |
||
|
|
mbotmiller
Cruncher Joined: Dec 1, 2005 Post Count: 47 Status: Offline Project Badges: 
|
I just wanted to see if anyone else has been seeing any odd behavior with OPNG GPU WUs recently. On a macOS 10.15.7 host that has both an Intel and AMD GPU, I'm seeing WUs for OPNG are failing almost immediately with a computation error after ~8-20 seconds. This host has been a reliable cruncher of OPNG GPU WUs in the past and I haven't made any recent system changes other than installing a recent macOS security update. I've included an example of both an AMD GPU OPNG WU failing and an Intel GPU OPNG WU failing below in case anybody might have any insight into why this would have started happening in the past day or two.
Host information: Wed May 12 16:17:19 2021 | | Starting BOINC client version 7.16.14 for x86_64-apple-darwin Wed May 12 16:17:19 2021 | | log flags: file_xfer, sched_ops, task Wed May 12 16:17:19 2021 | | Libraries: libcurl/7.58.0 OpenSSL/1.1.0g zlib/1.2.11 c-ares/1.13.0 Wed May 12 16:17:19 2021 | | Data directory: /Library/Application Support/BOINC Data Wed May 12 16:17:19 2021 | | OpenCL: AMD/ATI GPU 0: AMD Radeon Pro 5300M Compute Engine (driver version 1.2 (Apr 13 2021 00:41:20), device version OpenCL 1.2, 4080MB, 4080MB available, 2000 GFLOPS peak) Wed May 12 16:17:19 2021 | | OpenCL: Intel GPU 0: Intel(R) UHD Graphics 630 (driver version 1.2(Apr 13 2021 00:47:18), device version OpenCL 1.2, 1536MB, 1536MB available, 221 GFLOPS peak) Wed May 12 16:17:19 2021 | | OpenCL CPU: Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz (OpenCL driver vendor: Apple, driver version 1.1, device version OpenCL 1.2) Example of failing AMD OPNG WU: Result Name: OPNG_ 0009221_ 00016_ 0-- <core_client_version>7.16.14</core_client_version> <![CDATA[ <message> process exited with code 255 (0xff, -1)</message> <stderr_txt> wcgrid_opng_autodockgpu_7.28_x86_64-apple-darwin__opencl_ati_102 -jobs OPNG_0009221_00016.job -input OPNG_0009221_00016.zip -seed 229571665 -wcgruns 1450 -wcgdpf 29 INFO: Using gpu device from app init data 0 INFO:[15:56:31] Start AutoGrid... autogrid4: Successful Completion. INFO:[15:56:40] End AutoGrid... INFO:[15:56:41] Start AutoDock for ZINC000423498851-ACR2.11_RX1--fr824_001--CYS114.dpf(Job #0)... OpenCL device: AMD Radeon Pro 5300M Compute Engine </stderr_txt> ]]> Example of failing Intel OPNG WU: Result Name: OPNG_ 0009225_ 00365_ 2-- <core_client_version>7.16.14</core_client_version> <![CDATA[ <message> process exited with code 255 (0xff, -1)</message> <stderr_txt> wcgrid_opng_autodockgpu_7.28_x86_64-apple-darwin__opencl_intel_gpu_102 -jobs OPNG_0009225_00365.job -input OPNG_0009225_00365.zip -seed 1487642406 -wcgruns 750 -wcgdpf 15 INFO: Using gpu device from app init data 0 INFO:[16:27:56] Start AutoGrid... autogrid4: Successful Completion. INFO:[16:28:06] End AutoGrid... INFO:[16:28:06] Start AutoDock for ZINC000424153199-ACR2.13_RX1--fr824_001--CYS114.dpf(Job #0)... OpenCL device: Intel(R) UHD Graphics 630 </stderr_txt> ]]> |
||
|
|
|