Parallel Computing Research and Development is taking place on two main nodes. The first is an Intel Xeon server with a Xeon Phi coprocessor model 3100-P with 57 physical cores. With hyperthreading the available x86-64 cores is 227. OpenMP is the main software parallelization technology. Operating System: CentOS 7.0.
For GPU-based system, we have an Intel Haswell system with a GTX-750ti GPU as the main development server. It has CUDA and OpenACC development on this node. Operating System: Ubuntu 15.10.
All the network configuration and software development of these servers is done by me.
I am currently considering – what would be required to make a compute cluster – but this has not been done yet.