World Community Grid: OpenPandemics - COVID-19 Now Running on Machines with Graphics

[RSS]

The software that powers OpenPandemics - COVID-19 has been adapted to use GPU power. This will not only help the researchers to screen more molecules, but also could help them study molecules that are more complex.

More... (https://www.worldcommunitygrid.org/about_us/viewNewsArticle.do?articleId=693)

[MindCrime]

OpenPandemics - COVID-19 - GPU is showing up as non-beta now.

[zombie67]

Work availability will be limited

[MindCrime]

Forum mod said
"Good afternoon,

The new GPU application for OpenPandemics will be sending out 1700 work units on average every 30 minutes.

Thanks,
-Uplinger"

[MindCrime]

"4/6/2021 9:42:19 PM | World Community Grid | Tasks for Intel GPU are available, but your preferences are set to not accept them"

[Al]

Forum mod said
"Good afternoon,

The new GPU application for OpenPandemics will be sending out 1700 work units on average every 30 minutes.

Thanks,
-Uplinger"
With the pent up demand for another gpu app, especially one that pays as well as this one does, those 3400/hour don't last long. I rarely have more than 2 or 3 of my gpus working. Presently I have only 50 wus across 2-NV and one Intel gpu.

Edit: And it comes with a new wuprop app.

[zombie67]

Note: They added a project preference setting to allow GPU crunching. You have to go into your account and turn that on.

[MindCrime]

Note: They added a project preference setting to allow GPU crunching. You have to go into your account and turn that on.

the option is in your device settings, and remember you can have more than one profile so make sure they all have it, if you have more than one. this link should take you there if you're logged in.

https://www.worldcommunitygrid.org/ms/device/viewBoincProfileConfiguration.do?name=Default scroll to the GPU section.

question is can i set up an app_info for only gpu? I'm nearing 100k hours on openpandemics (cpu) and i'd like to work on other cpu stuff while getting the gpu OPNG wus.

[Al]

the option is in your device settings, and remember you can have more than one profile so make sure they all have it, if you have more than one. this link should take you there if you're logged in.

https://www.worldcommunitygrid.org/ms/device/viewBoincProfileConfiguration.do?name=Default scroll to the GPU section.

question is can i set up an app_info for only gpu? I'm nearing 100k hours on openpandemics (cpu) and i'd like to work on other cpu stuff while getting the gpu OPNG wus.

You don't need to, just scroll up to the Processor Usage section and say NO to "Allow research to run on my CPU?"

[zombie67]

question is can i set up an app_info for only gpu? I'm nearing 100k hours on openpandemics (cpu) and i'd like to work on other cpu stuff while getting the gpu OPNG wus.

So you want only the GPU tasks for OpenPandemics, yet still get CPU for all the rest of the projects. An obvious answer is to just run two instances of BOINC, one with a GPU-only profile and the other with the CPU-only profile (with OpenPandemics un-selected). But running more than one instance always seemed hard for me to wrap my head around. This was posted on their forum, which sounds like it should work (I have not tried it myself):


In website device profile,
- Turn on "If there is no work available for the project(s) I have selected above, please send me work from another project."
- Turn off OpenPandemics and turn on other projects.
- Turn on some or all of: Use AMD / NVidia / intel graphics card.
- Make sure project limits is very high or unlimited. (scroll down to the bottom)
- If you want to, BOINC project preference, store 0.5 days or less.

CPU can always do the selected projects as long as selected projects have some CPU work available.

GPU will check if other projects have GPU work, nothing there, then with "please send me work from another project." on, can send GPU tasks from OpenPandemics. There appear to be very little amount of OpenPandemic GPU work so it may take a while to grab any GPU tasks.

3rd option: Use a VM, and in the VM do all of your CPU crunching. Run the GPU tasks in the native instance of BOINC (which can run tasks on the GPU).

[MindCrime]

Thanks zombie this is good info but Al pretty much got it for what i was meaning to ask. but now that you mention it i might run some other wcg projects on CPU and only open pandemics on gpu.

[MindCrime]

After seeing the "Tasks for Intel GPU are available, but your preferences are set to not accept them" over and over trying to get gpu tasks I finally got around to turning on and getting the igpu working in boinc again. I was able to turn it on in bios and see all three display devices in windows device manager but boinc only saw the 2x 7970s. Had to do a little cable rodeo and plug in another monitor to the mobo hdmi and now I got all three gpus showing in boinc.

But now that I've finally got boinc asking for amd/ati and intel WUs it says no tasks available... of course it does.

edit: haven't seen an igpu WU yet

Also I got tired of BOINC holding off on GPU WUs because of the higher claimed CPU usage so I figured i'd make an App_config, i guessed what the app would be called because it'll tell you the known ones when you get it wrong. I guessed right the first time..

<app_config>
<app>
<name>opng</name>
<gpu_versions>
<gpu_usage>1.0</gpu_usage>
<cpu_usage>0.20</cpu_usage>
</gpu_versions>
</app>
</app_config>

[MindCrime]

30,000 gpu batches released to stress the system. I haven't been paying attention enough to know how big that is. but i have multiple machines with 100wus, might be a limit per host.
https://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,43386_offset,0

edit: limit maybe is 50wu's per gpu.

[Al]

Uploads/downloads are slow.
Scheduler request failed: HTTP service unavailable

[scole of TSBT]

If you want to earn those hours, load up your GPUs with as many WUs that will run without erroring out. The longer they run the better....


<app_config>
<project_max_concurrent>0</project_max_concurrent>
<app>
<name>opng</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.5</cpu_usage>
</gpu_versions>
</app>
</app_config>

[zombie67]

Put something like this in your updater script for each machine. It will prevent the crazy long backoffs for up/downloads.


"C:\Program Files\BOINC\boinccmd" --host 192.168.7.248 --passwd password --network_available

[MindCrime]

upload/downloads seem to have cleared up, not sure if they made some fixes or if everyone is now loaded up.

I'm now getting WUs that are taking 3-5x longer than the previous ones that were just a few minutes. They're paying better but not scaling 3-5x, more like 2-3x. Still great credit relative to WCG. I switched this machine to 2x concurrency for the hours, maybe there will be a little net credit improvement too since these WUs are up and down on GPU loads.

[Dave gpu]

Just looked at my GPU with MSI afterburner, and it shows o% utilization .

[Dave gpu]

Just looked at my GPU with MSI afterburner, and it shows o% utilization .

Well now it seems like they are working again.@-)

[Al]

Just looked at my GPU with MSI afterburner, and it shows o% utilization .

I'm running 8 wus on a GTX 1650 with all 8 cpu threads free. It took that many to get gpu utilization up to 99% at least part of the time. CPU is maxed out at 100%. These gpu wus are, in reality, cpu wus that use a little gpu. Can't bitch about the credits though.

[trigggl]

If you want to earn those hours, load up your GPUs with as many WUs that will run without erroring out. The longer they run the better....


<app_config>
<project_max_concurrent>0</project_max_concurrent>
<app>
<name>opng</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>0.5</cpu_usage>
</gpu_versions>
</app>
</app_config>


Do I just place this in the project folder as app_config.xml and CPU will be unaffected? Do I need to shut boinc down first before doing it? This seems like a good one to run two at a time of.

Thanks,

[scole of TSBT]

Do I just place this in the project folder as app_config.xml and CPU will be unaffected? Do I need to shut boinc down first before doing it? This seems like a good one to run two at a time of.

Thanks,
The Open Pandemic CPU app is opn1 so it won't be affected by it. You can put the file in place any time but restarting the client will make sure it uses that config.