Performance analysis

[vorner]

This is somewhat open-ended task. There was nearly no performance analysis done and certainly no optimisation. We want to do some research in that area:

• [ ] Write some more benchmarks (we have very few single-threaded, we want to have other scenarios too)
• [ ] Try to use it under some realistic workload
• [ ] Trying to run it in perf (or something similar) to see if there are obvious bottlenecks that would be easy to fix

It would also be worth running the benchmarks or perf on some other architecture than just x86-64, like Arm.

(Marked as easy ‒ while there are no specific instructions, I believe trying to stress-test the data structure in some way should not be hard to do; reading the outputs of perf might be a bit harder though).