Fix benchmarks with regards to jsonwebtoken performance

[Gobd]

Closes #536

The benchmark results in the readme will also need updating but not sure if you wanted me to do that or run them yourselves. The results are exactly like the linked issue suggests, 50x better performance after this change.

I also added HS256 to the benchmarks because why not, and also that's what I use.

[ilteoood]

I think you could also update the benchmarks, so we could also close #507

[Gobd]

I had to regenerated the PS512 token to get those verification benchmarks to work again. These were run on my MacBook Pro with an M2 Pro and 16GB of RAM.

[simoneb]

@Gobd we've done a bit of overhaul on the way we run and capture the benchmark results. When you get a chance, would you mind doing this work on top of what's in master now? I chose to replace cronometro with mitata due to issues with running is cross-platform.

[Gobd]

Sure let me work on that

[Gobd]

Done! Now you can see jsonwebtoken's performance is much more reasonable so no one panics and thinks they need to switch ASAP. I mean they do need to switch to using the keys correctly ASAP to fix their performance. I have switched to fast-jwt anyways because it was almost no work and is faster.

[simoneb]

Awesome, thanks for this!