I never really finished the project, thus only the rough qualitative benchmarks you get at the bottom (measured mostly by profiling and size(1)); I saw that it wasn't enough of a win in the larger context where I needed it, thus it made sense to stop early instead of making exhaustive benchmarks.

The blog post was mainly for curious readers, I'm surprised it hit HN at all. :-)

gperf is very limited in the number of keys it can handle as opposed to, say, https://burtleburtle.net/bob/hash/perfect.html

Well, again, different problem constraints, different solutions. Seemingly that tool can handle larger sets than gperf (although it claims gperf stops at a couple hundred, which is an exaggeration; try it with the first 1000 lines of /usr/dict/words and it's nearly instant, and with the first 10k it needs 35 seconds or so), but it also says the runtime is even slower. My goal was to have faster runtime, not handle more keys. YMMV.

Not an exaggeration, just written when machines were a lot slower. Anyway, more work in this space is always welcome, so thanks.