I ran into a similar problem with docker on the same problem in zig; podman and on-metal had no performance regression but docker did. I kind of gave the fuck u

I ran into a similar problem with docker on the same problem in zig; podman and on-metal had no performance regression but docker did. I kind of gave the fuck up on the plummer's sieve shootout at that point because (of this and other performance regressions I was finding, like CPU throttling) I felt like I was fighting the stupid rules of the contest more than I was discovering things about performance. Anyways for the authors: try running it in podman and see if it eliminates the perf regression