The Open Graphics Language, or OpenGL, is an industry open standard that has existed since the early 90's. While in some ways it is showing its age, it is a stable API that enjoys widespread support, and if you write a program that makes use of this, it is pretty much guaranteed to work on any relatively modern GPU in existence. The CUDA samples folder actually contains many examples of how OpenGL can interface with CUDA (particularly in the 2_Graphics subdirectory), so interested readers may consider going over these examples. (The default location is C:ProgramDataNVIDIA CorporationCUDA Samples in Windows, and /usr/local/cuda/samples in Linux.)