Help Take the Fight to COVID-19 with BOINC and Folding@home


Recommended Posts

11 minutes ago, PilotReelMedia said:

I want to do this so bad but I'm having trouble with the basic settup. The RTP Boinc docker Space Invader One used in his example is not in plugins and I cant figure out how to configure the one that is and its very different and has no GUI :(. Can anyone help me get this up and running?

 

Go to the Apps tab download the linuxserver version of BOINC.  It has the client and manager built into it.  It replaced the RDP-BOINC and is easier to use than the boinc version.

  • Like 2
Link to comment

Been trying to get RDP-BOINC running with Rosetta@home since Sunday and every time I try to connect I receive a "this project is temporarily unavailable" message. I can connect to other projects fine, but I want to add the Rosetta one. Anyone seeing the same issue or have any suggestions?

Link to comment
1 hour ago, testdasi said:

iGPU? It's not supported across the board due to lack of processing power.

Aptalca mentioned he had setup the container to allow the use of Intel integrated GPU and Nvidia GPU (with the Unraid Nvidia build).  Of course, if the project does not support the iGPU, it does not matter if the container does.

Link to comment
22 hours ago, sjaak said:

found i bug in the docker system.

i have 2 F@H docker images setup. the first one is setup to use the gt710, the second one is setup to use the 1050TI

but after create the second images it only sees the 710 (which is not assign to it)

then a change the config.xml:


  <slot id='0' type='CPU'/>
  <slot id='1' type='GPU'/>
  <slot id='2' type='GPU'/>

now it using the 1050ti and de 2nd image has NO work assign to the gt710, only the 1st one...

pretty weird that the 2nd image is seeing the GPU which is NOT setup to using it...

first image:

642117940_Schermafdrukvan2020-03-1911-37-07.png.d026a62e07970a4cca753d1f0fdf302e.png

 

second image:

573685818_Schermafdrukvan2020-03-1911-29-35.png.636024907847c13849a66872f460f3fe.png

 

at least 2 GPU's (of the 3 gpu installed) are working :D

What did you set in the NVIDIA_VISIBLE_DEVICES variable for the containers?

Link to comment
23 hours ago, sjaak said:

i have 2 F@H docker images setup. the first one is setup to use the gt710, the second one is setup to use the 1050TI

642117940_Schermafdrukvan2020-03-1911-37-07.png.d026a62e07970a4cca753d1f0fdf302e.png

Is your GT 710 the passive cooling or active one? What's the temp like when it's running?

 

I have a spare passive GT 710 but it would be sandwiched between my 2 AIC NVMe and I really don't want to melt my data literally.

Edited by testdasi
Link to comment
10 minutes ago, testdasi said:

Is your GT 710 the passive cooling or active one? What's the temp like when it's running?

 

I have a spare passive GT 710 but it would be sandwiched between my 2 AIC NVMe and I really don't want to melt my data literally.

its passive cooled, and running most of the time around 86c (95c is max) but it's on the bottom of the case, almost no airflow there...

if you put a fan near it, it will be much lower.

Link to comment
1 hour ago, sjaak said:

just the UUID of the GPU:

36219566_Schermafdrukvan2020-03-2010-08-57.png.284f6a6eede5e90871c0ea2d727ce0af.png

So just for me to understand what you want, you want the two containers to see just one GPU each?

As far as I know, when setting only one specific device in the variable, the container should only see that one.

 

I saw that you have 2 GPU devices in the xml you posted. Remove one and it will only use one of the gpus. You might have to experiment with the ID to get the correct card.

 

Link to comment
2 minutes ago, saarg said:

So just for me to understand what you want, you want the two containers to see just one GPU each?

As far as I know, when setting only one specific device in the variable, the container should only see that one.

 

I saw that you have 2 GPU devices in the xml you posted. Remove one and it will only use one of the gpus. You might have to experiment with the ID to get the correct card.

 

yes, 2 docker images so each use have its own GPU (i have no idea to assign 2 gpu's to one docker image, if that is possible i going to use 1 docker image)

i changed to config.xml to 2 gpu's because with only one its seeing only the gt710, which is NOT assign to it... that's the weird thing about it...

i did experiment with the ID setting, no luck...

also, f@h2 sees the 710 but does not assign work to it...

 

i would love to know how to assign 2 specific gpu's to one docker image, especially for Plex and f@h.

(i have 1 pcie slot free, so an another 1050ti can fill that up :P )

Link to comment
50 minutes ago, sjaak said:

yes, 2 docker images so each use have its own GPU (i have no idea to assign 2 gpu's to one docker image, if that is possible i going to use 1 docker image)

i changed to config.xml to 2 gpu's because with only one its seeing only the gt710, which is NOT assign to it... that's the weird thing about it...

i did experiment with the ID setting, no luck...

also, f@h2 sees the 710 but does not assign work to it...

 

i would love to know how to assign 2 specific gpu's to one docker image, especially for Plex and f@h.

(i have 1 pcie slot free, so an another 1050ti can fill that up :P )

To make all GPU's available to one container you just set the NVIDIA_VISIBLE_DEVICES to all. You also then need to adjust the xml to reflect all GPU's.

 

image.png.7aafc03cba1ed85c29b2db818b2cc878.png

 

And the slots id part in config.xml:

 

 <!-- Folding Slots -->
  <slot id='0' type='GPU'/>
  <slot id='1' type='GPU'/>

In the above config I removed the CPU as I use it for boinc. If you want all 3, add one more line.

Link to comment
3 hours ago, LeoFender said:

Wondering if anyone is using the Disable Security Mitigations plugin and seeing improved times for CPU based work units?

Trouble with that is that to be able to calculate the speed difference, you would have to do the exact same work unit from start to finish twice (which can't happen), and measure the time it takes.  Because of their nature, disabling the mitigations will always improve the speed. 

Link to comment
12 minutes ago, saarg said:

To make all GPU's available to one container you just set the NVIDIA_VISIBLE_DEVICES to all. You also then need to adjust the xml to reflect all GPU's.

 

image.png.7aafc03cba1ed85c29b2db818b2cc878.png

 

And the slots id part in config.xml:

 


 <!-- Folding Slots -->
  <slot id='0' type='GPU'/>
  <slot id='1' type='GPU'/>

In the above config I removed the CPU as I use it for boinc. If you want all 3, add one more line.

thnx, going change it as soon the current jobs are finished.

i have at this moment the luck that the VM gpu (my desktop that i use to write this message) is an AMD vega64 so it will not be used by the containers (so no crashing)

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.