For that, it looks like unless you're running the same js on a really huge dataset webassembly will win (going from the second speed test). Even when you're compiling 50MB of JS with that thing, Wasm is 5% slower than JS, and when you're compiling 500KB (more typical) it's 300% faster.
