Whenever I research the performance problem of a JavaScript library, I stumble over many deep clones even when it is not necessary most of the time. People are just not aware how CPU expensive a deep clone in JavaScript is, even more on big objects.

For example when you run a Node.JS profiling on the PouchDB test suite, most CPU cycles are spend in running the clone() function.

Having an updated performance comparison of structuredClone() would be great.