The benchmark game is, as the name suggest, a game. A game that is won by tuning a specific program until it performs good on the specific input used in the benchmark and/or, in this case, using highly-optimized regexp libraries.

The page you linked, for example, shows that C++ is slower than PHP, Hack, Node.js, and Typescript.

Is it, though?

This issue is, IMHO, a little too broad to be actionable. "Make the regexp engine faster" well, yes, everyone likes fast regexp engines. We have a few (a little more specific) issues about that.

@ALTree > The benchmark game is, as the name suggest, a game.
The name "benchmarks game" signifies nothing more than the fact that programmers contribute programs that compete (but try to remain comparable) for fun not money.

shows that C++ is slower than PHP, Hack, Node.js, and Typescript.

No. Shows a C++ program slower than PHP, Hack, Node.js and Typescript programs.

Is it, though?

Yes.

@dongweigogo

regex-redux is a new task, based on regexdna.

Try re-writing some of the old Go regex-dna programs to perform the new substitutions.

@igouy ,
yes, the original code seems written a long time ago, and I've made some improvements.
After I tried various regular expressions and self-made input data to avoid biases, the conclusion is still the same that in all cases, the go regexp is beaten by many other languages (I tried python and rust. For python, I'm guessing it is because most of the part in python is implemented in C).
This bothers me! -_-!

As you say, in a language like Python, the regexp engine is implemented in C. So you are comparing a somewhat tuned Go implementation with a highly tuned C implementation. You also need to consider the characteristics of the engine. Go has chosen to follow the re2 path (not surprising, since Russ Cox is a major author of both Go and re2). re2 has much better performance characteristics than some other regexp engines, in that it never has an exponential slowdown, but that comes at a cost for other regexps (https://swtch.com/~rsc/regexp/). Even then, the C++ re2 implementation has many optimizations that are not in the Go implementation. See, for example, #11646.

So: can regexp be sped up? Yes. But it will take work. I'm not sure whether to bother leaving this issue open or not.

I'm going to close this in favor of concrete optimization bugs like #11646.