These types of comparisons aren't "fair". As in most fields of engineering, programming languages/databases/etc. are about tradeoffs, and benchmarks aren't even more tricky.
From the blog:
> If you don’t have time: Rust can run much faster, but so can the other languages, and it turns out that the Java implementation might run faster than the Rust fastest implementation, according to the new benchmarks I’ve run after many Rust developers came to assist in making Rust faster. Common Lisp may have fallen behind, but that’s likely just because it was not nearly as optimized as the Java and Rust implementations were.
The reason PHP, Ruby, Python, bash, etc. are popular are because of developer productivity, not performance.
Also, it's a false dichotomy. Java uses `zlib` for its GZIP*Streams for performance reasons, and if there was a nice Rust library I wanted to use, I would just build a shared library and call it from Java, just like Java internally uses for `zlib`.
Additionally, there's a fairly large difference in performance between C compilers. See clang, gcc, icc, etc., and the 100+ different options you can use when compiling something.
It's not even close to "apples vs oranges", it's like more "apples vs fire trucks". :-P