In the file that I've told you, there is allready an wiki2html parser in C source( the file wiki2html.tar.gz). Just compile it and run it! This is a standalone program, which is only takeing one file at the time. So my effort is to make this to an php module to run this parseing much much faster.
How accurately does it replicate the standard parser? The parser has some very odd behaviour in corner cases, much of which we want to keep for backwards compatibility. If it replicates it almost completely, then it would be really useful if you could remember where you found it...