Something to share - I had difficulties compiling a .cu file with #include with cuda 3.0 and msvc 9.0 The nvcc compiler first complained about not having stdint.h (true, it's not part of cuda or msvc). I #undef'd BOOST_HAS_STDINT_H in the config/compiler/nvcc.hpp header but then it complained about wrong int limits in base_traits. I figured out that setting #define BOOST_MSVC 1500 makes that error go away. I assume this breaks portability since nvcc.hpp is now locked to msvc but there doesn't seem to be a mechanism for telling boost about the nvcc host compiler. So ugly for now, but it seems to work at least for boost/random There's another problem with my code but it's probably not boost related, that is the mvsc debugger pointer jumps around wildly at some point in the code and then it crashes. In release mode, the program runs fine until the end when it too crashes. If you think you have an idea, I'll post the code. Sven ________________________________ The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message.