Build for different platform

I’m building a CUDA app in C/C++ on a WinXP 32 bit machine using VS2005 and am having problems with running that app on a Vista 32 bit machine. The app runs fine on other WinXP 32 bit machines, but I’ve tried it on 2 different Vista machines and have problems on both. Am I missing some obvious incompatibility between the two platforms? I’m using the 2.0 CUDA Toolkit and drivers on the Vista machine and the 2.0 CUDA Toolkit, drivers, and SDK on my development machine.



“having problems running the app” is awfully vague.


Okay… Thought there was enough information in the post to allow someone to call me out as a bonehead for trying to build on a WinXP system and run on a Vista system.

In as few words a possible: I seem to be receiving an “invalid device pointer” several times before the wheels fall off and I can no longer call cudaMallocHost successfully when the function call returns “Unknown error”.

I understand this may not be enough information at this point, so I’ll post when I have more information and it’s organized. Thanks for the reply.