Those are different implementations; I know those exist, and which is why I had specifically said there were multiple implementations and pointed at the "faster one", which is obviously the one most people are going to be paying attention to, and which is the one that is also most relevant as we could and probably should expect the implementations for other languages to use some crazy one-off library: it demonstrates and undermines the limitations placed on the various languages seem arbitrary enough that you can't read into these results as being indicative of the languages themselves, but simultaneously that we shouldn't even care as the fast Java entry effectively "cheated".

I am going to take a step back: you seem irked by my comments, due to the knee-jerk link to have me contribute a different implementation for C++ as well as arguing with the short reply what I really maintain is a nearly-offtopic point about the Java library in question here (both of which I am reading as slightly aggressive towards me or my comment).

Do you really like this benchmark? Are you super excited that Rust is slightly faster right now than C/C++/Java? What is going through your head when you read my comments? Do you disagree that we should expect all native-ish languages to do equally-enough well for most people given these constraints? Is it that my comment is coming off to you as "no fun allowed", as I am trying to point out that this is a meaningless competition that mostly teaches us about different things than "who is winning"?

In fact, looking at your recent comment history, I am realizing you are being super aggressive about this with everyone, and are nigh unto spamming the play.html link to everyone who tries to comment about the high-level idea of "what are we doing here". What's up?