I made a new implementation of the Cuda2GL sample provided in the CUDA Samples by Nvidia. The sample shows how to generate a texture in CUDA and use it in an OpenGL context without copying it to main memory. The original implementation uses Glut and OpenGL immediate mode (Compatibility profile) to draw the resulting texture. The sample is a good basis for writing GPU-accelerated raytracers/raymarchers.
This implementation improves on the original example by using the libraries GLEW, GLFW and strict OpenGL 4.5 core, using buffer objects, allowing your applications to be analyzed by excellent tools such as Renderdoc, which only support core OpenGL Profiles.
Check it out on Github: https://github.com/Forceflow/cuda2GLcore