This blog post will show you how to use Amazon EC2 GPU instances with Folding@home. This is a great way to help researchers, so please consider donating some GPU time. If you have AWS credits about to expire, why not burn them for a good cause?
First, I fire up an Amazon EC2 P3 instance, which hosts an NVIDIA V100 GPU. I use the NVIDIA Deep Learning AMI 19.11.3 in order to make sure that I have the latest NVIDIA drivers. This should also work on other AMIs, but your mileage may vary.
Then, I simply follow these instructions to manually install the Folding@home client. Here are my exact steps:
wget [https://download.foldingathome.org/releases/public/release/fahclient/debian-testing-64bit/v7.4/fahclient\_7.4.4\_amd64.deb](https://download.foldingathome.org/releases/public/release/fahclient/debian-testing-64bit/v7.4/fahclient_7.4.4_amd64.deb) wget [https://download.foldingathome.org/releases/public/release/fahcontrol/debian-testing-64bit/v7.4/fahcontrol\_7.4.4-1\_all.deb](https://download.foldingathome.org/releases/public/release/fahcontrol/debian-testing-64bit/v7.4/fahcontrol_7.4.4-1_all.deb) sudo dpkg -i --force-depends fahclient\_7.4.4\_amd64.deb sudo dpkg -i --force-depends fahcontrol\_7.4.4-1\_all.deb
Once I’ve completed the wizard setup, the client starts automatically. ‘htop’ confirms that the ‘a7’ Folding@Home core is crunching data.
By default, training is only running on the CPU. Let’s put that GPU to work!
I need to edit /etc/fahclient/config.xml (sudo required):
<config> <! — Client Control → <fold-anon v=’true’/> <! — Folding Slot Configuration → **<gpu v=’true’/>** <! — Slot Control → <power v=’full’/> <! — User Information → <user v=’JulienS’/> <! — Folding Slots → <slot id=’0' type=’CPU’/> **<slot id=’1' type=’GPU’/>** </config>
Then, I just stop and start the client:
sudo /etc/init.d/FAHClient stop sudo /etc/init.d/FAHClient start
This fires up a GPU-optimized Folding@home core (core ‘22’).
Pretty soon, nvidia-smi tells me that the GPU is now crunching as well.
Take that, COVID-19. Your days are counted.
Again, please consider donating some GPU time if you can. Thank you.