In mitkGPGPU.h, there's this:
#ifdef _WIN32
#include <windows.h>
#else
#include <GL/glx.h>
#include <X11/Xlib.h>
#endif
Which means you either have to be building on Windows, or building using X11. This means the build will fail for MAC OS X unless you build VTK, Qt, and MITK as an X11 application.
Alan Antonuk (on mitk-users-list):
Taking a look at the problem of compiling the mitkGPGPU.cpp on Mac OS X it looks like all that is missing is a native way to create a OpenGL context to run the GPGPU stuff.
Why not use Qt to manage the creation of the context (using the QGLContext and QGLFormat classes)?
I'm willing to create a patch for this.
-Alan
MacOS support for GPGPU is not 1.0 relevant,
cause this feature is barely used on any platform,
(except for an internal filter)
and we have a new OpenCL framework for advanced GPGPU programming.
however the fix proposed by Alan Antonuk should work.
Resetting all bugs without active assignee flag to "CONFIRMED". Change status to IN_PROGRESS if you are working on it.
Needle segmentation depends on this.
We should discuss whether to keep this feature or not. Yet, it is not compiling on Mac. Furthermore, OpenCL became very popular, so we have an alternative for this.