We present a new compiler framework for truly heterogeneous 3D stencil computation on GPU clusters. Our framework consists of a simple directive-based programming model and a tightly integrated source-to-source compiler. Annotated with a small number of directives, sequential stencil C codes can be automatically parallelized for large-scale GPU clusters. The most distinctive feature of the compiler is its capability to generate hybrid MPI+ CUDA+ OpenMP code that uses concurrent CPU+ GPU computing to unleash the full potential of powerful GPU clusters. The auto-generated hybrid codes hide the overhead of various data motion by overlapping them with computation. Test results on the Titan supercomputer and the Wilkes cluster show that auto-translated codes can achieve about 90 % of the performance of highly optimized handwritten codes, for both a simple stencil benchmark and a real-world application in cardiac modeling. The user-friendliness and performance of our domain-specific compiler framework allow harnessing the full power of GPU-accelerated supercomputing without painstaking coding effort.

A recent trend in modern high-performance computing environments is the introduction of powerful, energy-efficient hardware accelerators such as GPUs and Xeon Phi coprocessors. These specialized computing devices coexist with CPUs and are optimized for highly parallel applications. In regular computing-intensive applications with predictable data access patterns, these devices often far outperform CPUs and thus relegate the latter to pure control functions instead of computations. For irregular applications, however, the performance gap can be much smaller and is sometimes even reversed. Thus, maximizing the overall performance on heterogeneous systems requires making full use of all available computational resources, including both accelerators and CPUs.