My experience using loom for a load generator

[Eric Ackermann]

Hello,


> If you have time you might explore event loop per core approach.
I did some testing with different numbers of threads. As long as I used one thread 
per CPU or more the results were within the margin of error from each other.

> But with your code you got up to ~10K requests per second?
In one of the scenarios I posted earlier my code reached throughputs within 
the margin of error of wrk2 (~13k requests per second, highest 3-run average around
100 rps slower than wrk2).  In the past, I have even seen 14k requests per second,
but I could not reproduce these numbers. Windows updates I did in the meantime 
might explain the performance loss.

Kind regards,

Eric Ackermann