alt Hacker Newspersonally, I disregard and ignore cross-language benchmarks. Benchmarks are hard, benchmarks in managed languages are extra hard.
And then the person writing the code has to be an expert in all languages involved to make it even half fair, instead of just a sound-bite/click-bait.
I think the benchmark game was really useful. It wasn't a which is better, it was what are the tradeoffs. Some of that was provided subjectively by curating it into safe and portable vs hyperoptimized categories. But there is also memory footprint etc. It was also a gauge of which languages people who care about performance preferred, as those got rapid updates.