And keep in mind, OpenGL and CUDA are two very different things. OpenGL is a graphic language (thus, the GL)… good for 3D graphic rendering and other graphics operations. Since that’s much of what BCC plug-ins do, it’s pretty well suited to this task. OpenGL has been around a very long time… it originated at Silicon Graphics in 1992, but has long since come under the management of an industry SIG (see https://www.opengl.org).
CUDA is a different thing, an API for GPGPU Computing (General Purpose GPU). This means accelerating pretty much anything, just as coding for FPU or MMX/SSE has. Rather than work in terms of grapics primitives as OpenGL does, it works in actual mathematics.
As John pointed out, OpenGL is, well, open, while CUDA is a proprietary nVidia thing. The similarly named OpenCL is an open standard for GPGPU computing, and the only GPGPU platform supported by both nVidia and AMD/ATi. Boris is not using any GPGPU computing, at least not yet. But since graphics cards were actually meant to, well, render graphics, OpenGL may be all they really need, at least for most effects.
-Dave