The code compilation benchmarks on page 3 of this article [1] make me think about building a 7950X box as a CI runner. Just a massively brute force way to reduce continuous integration build and test times. :)

Has anyone benchmarked these against typical CI cloud providers, or deployed self-hosted runners like this? I’m curious if it’s truly possible to reduce CI times -50% to -75%. If so, huge win for developer happiness… and hardware and power costs would be pretty insignificant compared to the happiness and productivity gains.

[1] https://www.phoronix.com/review/ryzen-7600-7700-7900-linux/3