| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Beta Testing
Forum: Beta Test Support Forum
Thread: OpenPandemics GPU Beta Test - Feb 18 2021 [Issues Thread] |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 272
|
|
| Author |
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3294 Status: Offline Project Badges: 
|
Just finished my first task from this batch:
----------------------------------------BETA_OPN1_0020010_00198_1 Ryzen-5-1400 Pending Validation 20/02/21 16:09:59 20/02/21 16:53:03 0,09 / 0,71 72,8 / 0,0 46 Jobs completed. AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
maeax
Advanced Cruncher Joined: May 2, 2007 Post Count: 142 Status: Offline Project Badges:
|
NVIDIA must be set to YES.
----------------------------------------Otherwhise no tasks for AMD and Intel! Have only AMD and Intel GPU's. The message is: Tasks for NVIDIA GPU are available, but your preferences are set to not accept them. Radeon Pro WX 7100 finished 40 jobs in one task in 8 Minutes.
AMD Ryzen Threadripper PRO 3995WX 64-Cores/ AMD Radeon (TM) Pro W6600. OS Win11pro
----------------------------------------[Edit 3 times, last edit by maeax at Feb 20, 2021 5:20:57 PM] |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3294 Status: Offline Project Badges:
|
bikeaddict,
----------------------------------------I looked up your GPU and it sounds about right that it will take several hours to finish 1 task. AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
Richard Haselgrove
Senior Cruncher United Kingdom Joined: Feb 19, 2021 Post Count: 360 Status: Offline Project Badges: 
|
Seem to be having a problem with intel_gpu version, under Windows 7
----------------------------------------BOINC reports OpenCL: Intel GPU 0: Intel(R) HD Graphics 4600 (driver version 10.18.10.3621, device version OpenCL 1.2, 1298MB, 1298MB available, 192 GFLOPS peak) Task has reached 1.316% progress, but has stalled at CPU time 00:00:22 CPU time since checkpoint 00:00:03 Elapsed time 00:52:41 Estimated time remaining 2d 17:51:47 Fraction done 1.316% Currently showing absolutely 0.0000% CPU usage, despite being set to real-time priority (a trick learned at Einstein, where this intel_gpu usually lives). I know how to fetch working files out of the slot directory, if that would help. Edit - a second (identical) machine is making progress: CPU time 00:00:18 CPU time since checkpoint 00:00:03 Elapsed time 00:11:13 Estimated time remaining 00:11:09 Fraction done 50.158% Will investigate. [Edit 1 times, last edit by Richard Haselgrove at Feb 20, 2021 5:20:32 PM] |
||
|
|
goben_2003
Advanced Cruncher Joined: Jun 16, 2006 Post Count: 145 Status: Offline Project Badges:
|
Hi,
----------------------------------------Is this expected to fully load the gpu? It is loading in pulses(0-99%). I also noticed that it does not reserve a CPU slot. Meaning that with my 6 core/12 thread i7, it will run 12 cpu jobs and 1 .947CPU + 1 Nvidia GPU. The below numbers are with cpu jobs reduced by 1 to ensure a core free for the GPU. For my Nvidia RTX 2060 Average load: ~26%. Pulses are a lot shorter than on the Intel and seem to align with starting and ending jobs (Job #0,1,2,3,etc) For my Intel UHD 620: Average load: ~82% Pulses seem to align with starting and ending jobs (Job #0,1,2,3,etc) My theory for why the average load is higher on the intel is that the jobs take longer to run, which brings up the average load. They both alternate periods of heavy activity and zero activity. The heavy activity periods are significantly longer on the intel. -goben_2003 Note: edited for pulse alignment on nvidia  [Edit 1 times, last edit by goben_2003 at Feb 20, 2021 7:12:53 PM] |
||
|
|
pututu
Senior Cruncher United States Joined: Jan 3, 2016 Post Count: 241 Status: Recently Active Project Badges: 
|
So far running fine on my GTX 750 ti, windows 10, 456.71 driver.
----------------------------------------GPU utilization not 100% and I suppose this is normal. Could possibly run 2 tasks per gpu if monitor lagging is not an issue. Looking forward for the production WUs soon...   [Edit 1 times, last edit by pututu at Feb 20, 2021 5:38:39 PM] |
||
|
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges: 
|
Hello, I have a question: Are there any WUs for Linux and Intel GPUs? I got none on my linux-machine. The messages from Boinc: Libraries: libcurl/7.58.0 OpenSSL/1.1.1 zlib/1.2.11 libidn2/2.0.4 libpsl/0.19.1 (+libidn2/2.0.4) nghttp2/1.30.0 librtmp/2.3 OpenCL: Intel GPU 0: Intel(R) HD Graphics Skylake Desktop GT1 (driver version 1.3, device version OpenCL 2.0 beignet 1.3, 1900MB, 1900MB available, 96 GFLOPS peak) [libc detection] gathered: 2.27, Ubuntu GLIBC 2.27-3ubuntu1.4 Processor: 2 GenuineIntel Intel(R) Pentium(R) CPU G4400 @ 3.30GHz [Family 6 Model 94 Stepping 3] 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 ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperf All I get is the following answer: foldingpc Requesting new tasks for Intel GPU Scheduler request completed: got 0 new tasks Can someone help me? Thank you ! Howdy Umlauf, There is work for linux and intel gpu. The feeder has say 20 slots and depending on who grabs the first result in the pair, then the second one has to go to the same class. So windows with nvidia grabs the first result, then second result has to go to windows with nvidia as well. If you could though, please find your sched_request*.xml and sched_reply*xml and send them to support@worldcommunitygrid.org with ATTN: Uplinger in the subject. This will allow me to check things on the scheduler in more detail. There could be a problem with intel going out in general as others have mentioned an issue. We did apply an update to the scheduler a few days ago to help send work to intel gpu. There could be a bug in that code that didn't show up during my testing. Thanks, -Uplinger |
||
|
|
|