Using SIMD for `string.h`

I wrote a few SIMD optimized versions of functions from [`string.h`](https://github.com/ncruces/go-sqlite3/blob/main/sqlite3/libc/string.h).

I'm using them in a project of mine that uses `wasi-sdk`, and as such, I'm working with them to upstream the work to `wasi-libc`; [`strlen`](https://github.com/WebAssembly/wasi-libc/blob/main/libc-top-half/musl/src/string/strlen.c) was the first one merged, after several rounds of reviews.

I've got some decent [results](https://github.com/WebAssembly/wasi-libc/issues/580#issuecomment-2864764025) on `wazero` (the runtime I've used) but the tests I did using `wasmtime` weren't too different. I assume browsers would similarly benefit.

The improvements are significant enough to be noticeable (by a few percent) in macro benchmarks of my project which is SQLite. I can only assume the browser version of SQLite could similarly benefit.

This issue is to figure out if it makes sense to bring these to Emscripten, and how.