> Are you serious? Would you base your research career on software that you don't trust? I wouldn't.

I guess you haven't seen the quality of academic code recently. The choices are basing your credibility on slow badly written code or fast badly written code.