This program was written as an introduction to parallel computing and uses CUDA to calculate the Fibonacci sequence on a GPU. To compile:
nvcc fib_cuda.cu -o fib_cuda
To run:
./fib_cuda x
where x is the desired numbers of the sequence you wish to calculate.
Note: this program uses unsigned long integers, which overflow after the 93rd number in the sequence. Kind of a crippling restriction; but again, this program was written mainly to familiarize with CUDA/parallel computing.