A couple weeks ago, well 11 days, I set up the Boinc Rosetta@Home project on my big server and desktop. They've been churning away trying to help design proteins, or whatnot.

The big server is an 8x4-core AMD server. The desktop is a single 6-core AMD. Neither has a GPU, which might help even more. I figured the desktop could be used when I'm not using it, and the server is largely idle anyway, so it should certainly be used.

After 11 days, using only 75% of the CPUs (I wanted to see if there was any compute impact), the big server (named Big128, because it has 128GB of RAM, as well as all of those CPUs), and it's dealt 92,356 units of work. This is implemented via the boinc-client Docker container, with my credentials passed as environment variables. I also ensured its port was shared on my LAN, so I could use the Boinc Manager from my desktop. Pretty slick.

The desktop, not really a slouch, but getting a little aged, pokes in smack at 4000 units of work. It's called hexus, because it's got six cores, of course. While Docker also runs on this system for development and practice, I chose to install the native client, since I occasionally mess with Docker.

Neither machine is particularly stressed by all of this. The server's still responsive, and the desktop doesn't seem impacted either, despite both showing near 100% CPU utilization. The RAM on both is also available, with little impact coming from all of the instances of Rosetta running. Negligible disk space is used. I've got a fairly large file server available, which is running Solaris on Sparc or I'd have it also running this, but I don't think I need it with the small demands.

I did the same thing for SETI@Home way back when. That was implemented as a screensaver on my workstations and PlayStation, too. They never found anything that I learned about because of the efforts from all of the home computers. I haven't heard of any help coming from this either. Maybe I should mine cryptocurrency instead.

For now I'll give what I can to try to give whatever use this might be for science. For science, I say!