Chess is a challenging problem, that fascinated many if not any of the best programmers in the world, such as John Von Neumann, Alan Turing, Ken Thompson, David Levy, …
CUDA on it’s side is a fascinating architecture that is a challenge to develop for, in many ways: incredible performance-level, low register count, few local memory for each SM (Stream Processors that regroups 8 SP-Scalar Processor), big global memory but with bad latencies, inter-thread communications & synchronizations are real hard to optimize, threaded-code should follow the same execution path to exploit an SM, …
CUDA Chess is the idea of exploiting the raw power of modern nVidia’s GPU to have a GPU Chess Engine.
I have programmed Chess-Engines before, but I am nor the best chess engine developer nor a strong chess player, so it won’t probably be the best chess engine from far, but still it is interesting to do a Proof-of-Concept using brute-force (or better put “brute-weakness” !). 
I should have forgot to say a word about Open-Source. I like open-source projects, I use many open-source software and I couldn’t be too thankfull for what this movement gives to me and the world. As such I would like to take part, and from start this project will be hosted in a publicly accessible SVN repository, so you could get the code, try it by yourself, modify it, send me patch, makes it better or even fork and do your own on your side!
As open-source is about freedom for each and anyone, and that freedom comes from knowledge, I would try to explain in the code and on my blog any idea that is used, because I would like to share the true basement of the code : algorithms, ideas, thoughts.
Welcome to CUDA Chess, hope to read from you on the comments section!