US 2014/0180989 A1 has been applied by Google. It seems that the patent is about convolutional neural network with parallel architecture. I have searched further and found that several companies including Microsoft have applied for a similar patents. Neurala, LLC already has a patent, which seems to cover a broad range of computer programs utilizing GPU.
On the other hand, nVidia, a major GPU maker, provides CUDA technology as a tool for developing GPU-based programs and seems to be encouraging the development. They also provide a large list of such applications.
It seems that use of GPU as a parallel platform is expanding but nobody is worried about patent infringement. I don’t understand what’s going on. Do we really have to worry about anything about patent at all when we develop a parallel program for GPU with CUDA, which is readily provided for this purpose?