Unfortuatly rayon has a bug that if you Threadpool.spawn from inside a parallel iterator from inside a Threadpool.install, that the parallel iterator will BLOCK till the Threadpool.spawn finished, which causes many many lag spikes.
I assume this might be the case for the pictures in the gantt chart where a system took unusual long or had a long pause that was unexplained.
I also raise the number of threads by 1, as this rayon thread will prob be useless in all cases and have no real work to do.
EDIT: it turns out the tests are sporadicly failing and this soluction doesnt work
See that we spawn 2 jobs in the first loop, the loop seems to NOT complete until those jobs are executed
Next step is to do everything with plain rayon coding
