Monte Carlo option pricing: other implementations

I am looking for implementations of Monte Carlo option pricing model on the GPU.

I see an example in the CUDA SDK, which uses Monte Carlo methods for option pricing.
But I am wondering if people have worked towards optimizing that implementation or writing new implementations of the same.

Thanks.
Mayank

I also want to know if there is a stochastic volatility model implementation using Brownian Bridge to speed it up.