| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: Mandatory "Request for Project to be Ported to GPU" Thread |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 133
|
|
| Author |
|
|
Aurum
Master Cruncher The Great Basin Joined: Dec 24, 2017 Post Count: 2391 Status: Offline Project Badges: 
|
On the other hand it is also true that if a task performs best as a single thread then it should stay a single threaded. Tasks like these include LHC@Home's Sixtrack project... and LHC ATLAS.  ...KRI please cancel all shadow-banning |
||
|
|
Aurum
Master Cruncher The Great Basin Joined: Dec 24, 2017 Post Count: 2391 Status: Offline Project Badges:
|
The GPUs that can do CUDA 11 are all from Nvidia, and do not include graphical interfaces. In other words, you cannot find them on any graphics cards, only on cards that can do CUDA or OpenCL, but not graphics. And yet all my 1080 Ti & 2080 Ti GPUs have:NVIDIA-SMI 450.57 Driver Version: 450.57 CUDA Version: 11.0 including the one I'm reading this web page on. I'm confused, maybe it's the subtle distinction between being CUDA 11 capable and actually being used. I know I have numerous projects that run CUDA 10.x WUs. ...KRI please cancel all shadow-banning |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
Over at F@H they have had a unified work load for years. All of the WU can run on CPU or GPU. Yes, they use OpenCL to do this. It works fine and there are not bugs thanks to extensive testing. They use OpenCL for their Core 22 work units on the GPU, but are close to releasing a CUDA version. It is based on "OpenMM", and not necessarily tied to OpenCL. For the CPUs, they use an entirely different "core A7", soon to be replaced with "core A8", based on GROMACs, which does not rely on OpenCL at all. (Though maybe a bit confusingly GROMACs also allows the use of GPUs, and they can use OpenCL, but Folding is not at the moment implementing this use of GROMACs). https://foldingforum.org/viewtopic.php?f=72&t=33496&p=337353#p337353 https://foldingforum.org/viewtopic.php?f=74&t=35624 In theory, they can do more or less the same type of work on either the CPU or GPU, but due to differences in the size of the molecules, one set of work runs best on the CPU and another best on the GPU. They are not interchangeable in practice. [Edit 2 times, last edit by Jim1348 at Aug 15, 2020 6:25:35 PM] |
||
|
|
robertmiles
Senior Cruncher US Joined: Apr 16, 2008 Post Count: 445 Status: Offline Project Badges:
|
The GPUs that can do CUDA 11 are all from Nvidia, and do not include graphical interfaces. In other words, you cannot find them on any graphics cards, only on cards that can do CUDA or OpenCL, but not graphics. And yet all my 1080 Ti & 2080 Ti GPUs have:NVIDIA-SMI 450.57 Driver Version: 450.57 CUDA Version: 11.0 including the one I'm reading this web page on. I'm confused, maybe it's the subtle distinction between being CUDA 11 capable and actually being used. I know I have numerous projects that run CUDA 10.x WUs. Maybe what I found at an Nvidia site for CUDA 11 wasn't complete enough. Or possibly BOINC is saying that the driver can handle CUDA 11 and does not say whether it can do so with this GPU. It reports CUDA 11.0 with my GTX 1080 and a 451.67 driver also. [Edit 1 times, last edit by robertmiles at Aug 16, 2020 12:26:31 AM] |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges:
|
Or possibly BOINC is saying that the driver can handle CUDA 11 and does not say whether it can do so with this GPU. It reports CUDA 11.0 with my GTX 1080 and a 451.67 driver also. All of my GPUs from the GTX 1060 on up have CUDA 11 with the latest drivers, which are 450.57 on my Ubuntu 18.04.5 machines. I would guess the same for Windows. I don't know about the older cards. |
||
|
|
adrianxw
Senior Cruncher Denmark Joined: Apr 13, 2008 Post Count: 196 Status: Offline Project Badges: 
|
If the project introduces a GPU version without the ability to refuse GPU work on a per machine basis, I'll be leaving. I can run GPU work on this system, but the GPU in my other machine is starting to die, I have CPU only projects in its portfolio.
|
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges: 
|
If the project introduces a GPU version without the ability to refuse GPU work on a per machine basis, I'll be leaving. I can run GPU work on this system, but the GPU in my other machine is starting to die, I have CPU only projects in its portfolio. What you want already exists on the device profiles.  - 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 |
||
|
|
supdood
Senior Cruncher USA Joined: Aug 6, 2015 Post Count: 333 Status: Offline Project Badges:
|
Just use your cc_config file on the machine you don't want using the GPU. From the BOINC Client Configuration page :
----------------------------------------Option:<exclude_gpu> Don't use the given GPU for the given project. If <device_num> is not specified, exclude all GPUs of the given type. <type> is required if your computer has more than one type of GPU; otherwise it can be omitted. <app> specifies the short name of an application (i.e. the <name> element within the <app> element in client_state.xml). If specified, only tasks for that app are excluded. You may include multiple <exclude_gpu> elements. If you change GPU exclusions, you must restart the BOINC client for these changes to take effect. If you want to exclude the GPU use for all projects, look at the <ignore_ati_dev>, <ignore_nvidia_dev> and <ignore_intel_dev> options further down. Requires a client restart. <exclude_gpu> <url>project_URL</url> [<device_num>N</device_num>] [<type>NVIDIA|ATI|intel_gpu</type>] [<app>appname</app>] </exclude_gpu> |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes, adding to that
<no_gpus>0|1</no_gpus> If 1, don't use GPUs even if they're present. Requires a client restart. There's further up mention of combining this with anonymous platform app_info.xml but remember WCG specifically disabled app_info.xml functionality for which app_config.xml was created instead. What's unclear is, if this is communicated to the server i.e. if work is requested, does the server know not to send GPU work? |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
Come on team, I know that most of us are cooped up in our houses and the zombie invasion has severely interrupted our accustomed activities (I know it has affected me). Lets turn our frustration on figuring out how to get more OPN1 WUs crunched. I think we are all here for the science, donating CPU cycles for good causes.
----------------------------------------Thanks, CJSL Crunching as fast as my four CPU threads will let me crunch |
||
|
|
|