WASM AES impl

WASM targets do not have access to AES hardware acceleration. Many protocols in `mpz` use AES heavily.

Investigate whether an AES implementation which utilizes SIMD for WASM targets would provide significant performance benefits, keeping in mind that we want to stick with the constant time bitslicing technique.

No files, tests, or entry points are named. Start by locating the existing constant-time bitsliced AES implementation and the WASM target configuration, then compare a SIMD-based approach against the current implementation. Done requires a measured performance assessment and a clear decision about whether the change is worthwhile without weakening the constant-time technique.

Written by the indexing model from the issue text.