CUDA runtime header error what does it meant?

i am working on cuda compilers. nvcc. i am using both windows 7 and YDEL Cuda linux in my both of them i have insatlled nvidia cuda 3.0 toolkit and their respective drivers. but i need to execute a default cuda program a matrix transpose, but when i give command " nvcc -o " , i found the error showing no directory . what does it mean , plz tell me how cuda runtime header file is given?

you should try posting this in the CUDA GPU computing sections, or an admin should move it

You can use [font=“Courier New”]-I [/font] to tell nvcc about additional directories to search for include files.