Hacker News Reader

While I agree about million times, [micro]benchmarks that run seconds (or milliseconds) have extremely limited use, esp. when combined with non-compiled languages, that's effectively comparing compilation, guidance, and run time.

I'd consider the entire site/process flawed.

It's useful in demonstrating there isn't a factor of a million difference, which was the discussion.

Do the java benchmarks include JVM startup?

Yeah, it appears they just run a script, I don't think the code itself is allowed to say how long the test took. That's why you have to subtract like 80ms (estimating from the hello world comparison) for most of them.

That link does not work for me?

Not sure why they posted an AMP link, here's the correct one: https://programming-language-benchmarks.vercel.app/rust-vs-j...

I fixed it. Didn't trim enough of the URL on mobile..