| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: No GPU WUs , settings |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 36
|
|
| Author |
|
|
squid
Advanced Cruncher Germany Joined: May 15, 2020 Post Count: 56 Status: Offline Project Badges: 
|
AutoDock-GPU runs fine on the Supercomputer IBM Summit (also called OLCF-4) with 27,648 NVIDIA GPUs.
https://blogs.nvidia.com/blog/2020/05/26/covid-autodock-summit-ornl/ |
||
|
|
Ian-n-Steve C.
Senior Cruncher United States Joined: May 15, 2020 Post Count: 180 Status: Offline Project Badges: 
|
AutoDock-GPU runs fine on the Supercomputer IBM Summit (also called OLCF-4) with 27,648 NVIDIA GPUs. https://blogs.nvidia.com/blog/2020/05/26/covid-autodock-summit-ornl/ Sounds like they got even better results with porting the OpenCL code to CUDA. With help from NVIDIA and Scripps Research, the team ported the code to CUDA so it could run on Summit, delivering an added benefit of another 2.8x speedup. Another researcher, Aaron Scheinberg of Jubilee Development, accelerated the work another 3x when he found a way to use OpenMP to speed up feeding data to the GPUs.  EPYC 7V12 / [5] RTX A4000 EPYC 7B12 / [5] RTX 3080Ti + [2] RTX 2080Ti EPYC 7B12 / [6] RTX 3070Ti + [2] RTX 3060 [2] EPYC 7642 / [2] RTX 2080Ti |
||
|
|
mgpointner
Advanced Cruncher Argentina Joined: Nov 16, 2009 Post Count: 55 Status: Offline Project Badges: 
|
I'm really wondering why they don't migrate all tasks from CPU to GPU. Or is there a reason why everything can't be done on GPU?
---------------------------------------- |
||
|
|
Speedy51
Veteran Cruncher New Zealand Joined: Nov 4, 2005 Post Count: 1295 Status: Offline Project Badges: 
|
I'm really wondering why they don't migrate all tasks from CPU to GPU. Or is there a reason why everything can't be done on GPU? To me the most logical reason why everything cannot be transferred over is because it would be too much pressure on the system to cope with the amount of work units returned on a daily basis. As an example I have one RTX 2070 result can run anywhere from just over a minute to 3 or 4 minutes a four-minute work unit is rare. These are just my thoughts  |
||
|
|
spRocket
Senior Cruncher Joined: Mar 25, 2020 Post Count: 274 Status: Offline Project Badges: 
|
I sometimes have days where I get nothing, and others where I'll get several nice blocks of Nvidia WUs. I just consider it to be a nice bonus to my usual daily crunching, and don't consider it worth going overboard trying to get them when the chances are slim no matter what. If they can manage to get the firehose going again, great, but until then, I'll stick with 15 CPU+1 GPU task instead of 14+2.
|
||
|
|
zdnko
Senior Cruncher Joined: Dec 1, 2005 Post Count: 225 Status: Offline Project Badges: 
|
I'm really wondering why they don't migrate all tasks from CPU to GPU. Or is there a reason why everything can't be done on GPU? I don't even understand the point of developing the app for 3 different gpu and 3 different os and then releasing a few wu for gpu  |
||
|
|
jeffwy
Advanced Cruncher Taiwan Joined: Nov 17, 2004 Post Count: 77 Status: Offline Project Badges: 
|
Ok it's now been over three weeks and I have gotten Zero (0) new GPU WUs, have they stopped sending them out?
---------------------------------------- |
||
|
|
nexuss
Cruncher Joined: May 9, 2009 Post Count: 1 Status: Offline Project Badges: 
|
Ok it's now been over three weeks and I have gotten Zero (0) new GPU WUs, have they stopped sending them out? no i receive 20- 55 Wu´s on Day. is not much but always something |
||
|
|
jeffwy
Advanced Cruncher Taiwan Joined: Nov 17, 2004 Post Count: 77 Status: Offline Project Badges:
|
I have gotten zero! Almost for a month now.
---------------------------------------- |
||
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 2186 Status: Offline Project Badges: 
|
Well, it's pretty easy to get tasks for my iGPU HD4600, mostly because other iGPU's that gets work, are not reliable, or fail immediately.
---------------------------------------- Work for my GTX980 though, is not that easy to get, but I get a few now and then.I'm only running my computer for a couple of hours/day, and only GPU work. I feel no need to waste electricity on CPU tasks, when just a few GPU WU's during an hour, or even shorter total time, can crunch the same amount of "jobs", as the CPU does during an entire day. [Edit 4 times, last edit by Grumpy Swede at May 26, 2021 6:36:13 PM] |
||
|
|
|