The programs we write in our group are usually meant for execution in a parallel environment. We use MPI libraries to distribute the compute load across processors. Some of this compute load can be offloaded to a GPU as well.
References
MPI + OpenCL implementation of a phase-field method incorporating CALPHAD description of Gibbs energies on heterogeneous computing platforms
P. Gerald Tennyson, G. M. Karthik, G. Phanikumar
Computer Physics Communications, pp. 48-64, vol. 186 (2015) DOI
Supercomputing Applications in Materials Engineering
G. Phanikumar, Pradip Dutta and K. Chattopadhyay
Current Science, Vol. 78, No. 7(April 10), pp. 847-849 (2000) LINK