The demo loads a volume using VolumieViz and computes the marching cubes for this volume. It is possible to change the ISO value.
The shape is computed using the InventorCuda API. The CUDA code is located in the file MarchingCube.cu. The cubin file is loaded using the InventorCuda API and during the compute process the CUDA code is called in orde rto speed up the process.
The CUDA code is based on the NVidia CUDA demo provided in the CUDA SDK.
Info: To use correctly this demo you need at least 512Mb of graphic memory. If you don't have enough memory you can try to comment the USE_GL_BUFFERS define (MarchingCubes.cxx line 17) to use the central memory.
SoBufferedShape, SoGLBufferObject, SoCudaBufferObject, SoCpuBufferObject, SoCudaContext, SoCudaDevice, SoCuda, SoCudaApi