Hardware for Folding@Home

Introduction

Update: After writing this, I have done a little hardware rationalization, by moving some discrete video cards around to different systems. I’ve updated the information in the main post to reflect this.

The main issue right now is getting work units from Folding@Home. It appears that there are more GPU work units than CPU work units.

As I have previously discussed, one useful thing you may be able to do to help fight COVID-19 is to volunteer some of your idle CPU and GPU resources to run Folding@Home. If you do this, you can also join the SQLFamily team (which is Team 236388), so your statistics show up here.

I’ve gotten several questions on Twitter about what hardware I am using for Folding@Home. I have four systems running this full time right now. These systems are not ideal for folding, since they are (by design) CPU-heavy instead of GPU-heavy.

From what I can tell, the Folding software really runs well on modern, discrete NVidia GPUs. That is not what I have in my four full-time Folding machines. Unfortunately, they all have very inexpensive and older AMD discrete graphics cards.

System 1

This system has a 6C/12T 7nm AMD Ryzen 5 3600X in an MSI B450 Tomahawk Max motherboard with 32GB of DDR4-3200 RAM. It has an NVidia GeForce GTX 1060 video card with 6GB of RAM. This system uses 45 watts at idle and 134 watts when running Folding@Home on both the CPU and GPU.

System 2

This system has an 8C/16T 12nm AMD Ryzen 7 2700X in an MSI B450 Tomahawk Max motherboard with 32GB of DDR4-3200 RAM. It has an NVidia GeForce GTX 960 video card with 2GB of RAM. This system uses 63 watts at idle and 245 watts when running Folding@Home on both the CPU and GPU.

System 3

This system has a 6C/12T 12nm AMD Ryzen 5 2600X in an MSI B450 Tomahawk motherboard with 32GB of DDR4-3000 RAM. It has an MSI Radeon R7 240 video card with 2GB of RAM. This system uses 51 watts at idle and 187 watts when running Folding@Home on both the CPU and GPU.

System 4

This system has a 6C/12T 12nm AMD Ryzen 5 2600 in an MSI B450 Tomahawk motherboard with 32GB of DDR4-3000 RAM. It has an MSI Radeon R7 240 video card with 2GB of RAM. This system uses 55 watts at idle and 142 watts when running Folding@Home on both the CPU and GPU.

Part Time Systems

I also have some other machines running Folding@Home at least part of the time.

Big Beast System

This is my main workstation (for the moment). This system has a 16C/32T 12nm AMD Ryzen Threadripper 2950X in an ASRock Fatal1ty X399 Professional Gaming motherboard with 96GB of DDR4-3200 RAM. It has an MSI NVidia GeForce GTX 1070 Ti video card with 8GB of RAM.

Gaming System

This system has an 8C/16T 7nm AMD Ryzen 7 3700X in an ASRock X570 Phantom Gaming 4 motherboard with 16GB of DDR4-3600 RAM. It has an XFX Radeon 5700 video card with 8GB of RAM.

Old Beast System

This system has a 16C/32T 14nm AMD Ryzen Threadripper 1950X in an ASRock X399 Phantom Gaming 6 motherboard with 64GB of DDR4-3200 RAM. It has an ASUS NVidia GeForce GT 1030 OC video card with 2GB of RAM.

Final Words

I have a few other systems down in the basement that I could be using, but I have to consider my aggregate electrical usage. I also have to be careful how many systems are running on a single 20 amp 120V circuit.

One thing I may do is to cannibalize some discrete video cards that I have in a couple of other unused systems, and put them in my full time machines. These include:

• NVidia GeForce GTX 1060 6GB
• NVidia GeForce GTX 960
• Radeon R9 Fury Nano

In order to use those faster discrete video cards, I’ll have to find the PCIe power cables for the power supplies that I am using in the full time systems.