Well I made an attempt at rewriting the legacy parser, and my findings revealed that the parser I had written (the one in use) is actually much more efficient that MediaWiki's so there is really no reason to switch. The difference is that my parser finds all the open, close, and delimiter items at once, whereas MediaWiki does it on a tag-by-tag basis in almost the same fashion. This creates another nested loop as far as MW is concerned.
Marking this as Complete since we got the converter finished.