Francesco Nigro

Adding `time` prefix for a `wrk -t 1 -c 10 -d 10` etc etc run reports: ``` Thread Stats Avg Stdev Max +/- Stdev Latency 102.97μs 40.82μs 2.23ms 79.04% Req/Sec...

@jesperpedersen lovely :D thanks!

> . We need to consider when a large number of connections and users (threads) have been defined, and the impact the config has on warming up the load driver...

> we need to JIT to compile the driver code and allocate the memory so that JIT and GC do not impact the results yeah, but that's an implementation detail,...

> We can (probably should) cover in documentation. That's what this issue can be turned into (I can do it) > I'm not sure that waiting a fixed length of...

Nice one @trustin ☺️ Eheh I see that someone is going to give it a shot for Netty :P

I am just sad that if we just want to pick connection from the same context probably we don't need a combiner executor but we could just use per-context (and...

I think https://github.com/openjdk/jdk/blob/701bc3bbbe49a46aea7efc195463cc2efd64a785/src/java.base/share/classes/jdk/internal/vm/SharedThreadContainer.java#L46 shows that you could use byte-level var handle, but if you use Java Object Layout to print the layout of AtomicBoolean, having less than int doesn't improve...

@vietj This change here is not in conflict with https://github.com/eclipse-vertx/vert.x/pull/5143 given that it works on different synchronization(s) happening in a inefficient way, so, if you agree on the changes, could...

@vietj ping