Volume convolution on the GPU using OpenCL.
For 27M voxels using 100 iterations, OpenCL is 650 times faster than C++ and 12525 times faster than VEX.

Volume convolution on the GPU using OpenCL.
For 27M voxels using 100 iterations, OpenCL is 650 times faster than C++ and 12525 times faster than VEX.