And the benchmark game is using Ruby 3.1, whereas 3.2 is significantly faster. YMMV though, it is going to depend on your use case, but we are always working on making Ruby faster, and if you run into a use case where Python is a lot faster, you can ping me on twitter @Love2Code and tell me about it. We'll take a look.
Python 3.11 also claims [1] to be 10-60% faster than 3.10, which is the version that the benchmarks game is using at this time. The difference in used RSS is also quite interesting in this comparison.
> The difference in used RSS is also quite interesting in this comparison.
YJIT 3.1 was eagerly allocating it's executable memory region regardless of whether it was needed or not. With the default settings that meant an extra 256M of RSS right at boot.
As mentioned in TFA that's no longer the case in 3.2, so expect this RSS number to go down by at least a good 200MB once they upgrade to 3.2.
And the benchmark game is using Ruby 3.1, whereas 3.2 is significantly faster. YMMV though, it is going to depend on your use case, but we are always working on making Ruby faster, and if you run into a use case where Python is a lot faster, you can ping me on twitter @Love2Code and tell me about it. We'll take a look.