CUDA 5.5 wglCreateContextAttribsARB​ and Wrangler Library (glew)

Accelerated Computing CUDA CUDA Programming and Performance
1

I was trying to point my CUDA 5.5 (yes I am a crusty old version CUDA developer) to the SDK version of glew (wglew.h) in the SDK located C:\ProgramData\NVIDIA Corporation\CUDA Samples\v5.5\common\inc\GL

It appears that the version of glew supplied with the SDK does not support this. My download for CUDA 7.0 is going to take about 5 hours. When/If it completes will I find wglCreateContextAttribsARB​ in that version? What is status of Wrangler/glew library in SDK?

The question comes as I am trying to compile code which uses the latest wrangler api off the web. I decided to try and compile in an example from sdk into this project… then the fighting for which version of glew. SDK uses glew and freeglut from SDK, but app I am working with uses latest glew but not glut (freeglut).

What is best way to use glut, glew? From SDK or random bits of the internet? Clearly my 5.5 glew is a bit stale, but are newer SDKs including latest wrangler?

2

Sorry to not answer your question but if you have the time then taking a look at GLFW + GLAD might be worth it.

I’ve been banging away on GLFW lately and CUDA interop seems to work fine and involves zero hassle.

I retch whenever I bring up the CUDA Samples. They’re splattered with #define’s and code sedimentation. :)

3

What is “code sedimentation”? Never heard that term before. Accumulated cruft? Removing cruft (beyond what is needed to keep the code functional) is probably low on the list of priorities as long as new features need to be developed and new GPUs need to be supported, while engineering resources remain limited.

4

Exactly! If you squint, you can almost see early CUDA engineers (and their decisions) trapped in amber.

5

LOL! Maybe one day you can contribute an article on that to the “Journal of Software Archaeology” :-)

6

OpenGL/CUDA interop (and a buffer swap):