Is there any optimized, reliable C++ Shunting-Yard algorithm implementation?

For commercial purpose, my team decided to use a Shunting-Yard algorithm in the calculating engine, and created an sample implementation. Although we have the implementation, if there is other reliable, optimized implementation for doing it, we would seriously consider using it. Is there any? The language is C++, and the usage is for time-critical application.

Since I have at least one implementation, I am looking for open-sourced library which has the algorithm implementation. BSD-like license would be preferable.