No description available.

Today's high-performance computing (HPC) clusters are seeing an increase in the adoption of accelerators like GPUs, FPGAs and co-processors, leading to heterogeneity in the computation and memory subsystems. To program such systems, application developers typically employ a hybrid programming model of MPI across the compute nodes in the cluster and an accelerator-specific library (e.g.; CUDA, OpenCL, OpenMP, OpenACC) across the accelerator devices within each compute node. Such explicit management of disjointed computation and memory resources leads to reduced productivity and performance. This dissertation focuses on designing, implementing and evaluating a runtime system for HPC clusters with heterogeneous computing devices. This work also explores extending existing programming models to make use of our runtime system for easier code modernization of existing applications. Specifically, we present MPI-ACC, an extension to the popular MPI programming model and runtime system for efficient data movement and automatic task mapping across the CPUs and accelerators within a cluster, and discuss the lessons learned. MPI-ACC's task-mapping runtime subsystem performs fast and automatic device selection for a given task. MPI-ACC's data-movement subsystem includes careful optimizations for end-to-end communication among CPUs and accelerators, which are seamlessly leveraged by the application developers. MPI-ACC provides a familiar, flexible and natural interface for programmers to choose the right computation or communication targets, while its runtime system achieves efficient cluster utilization. / Ph. D.

There are many challenges that need to be met before efficient and reliable computation at the petascale is possible. Many scientific and engineering codes running at the petascale are likely to be memory intensive, which makes thrashing a serious problem for many petascale applications. One way to overcome this challenge is to use a dynamic number of processes, so that the total amount of memory available for the computation can be increased on demand. This thesis describes modifications made to the massively parallel global optimization code pVTdirect in order to allow for a dynamic number of processes. In particular, the modified version of the code monitors memory use and spawns new processes if the amount of available memory is determined to be insufficient. The primary design challenges are discussed, and performance results are presented and analyzed. / Master of Science

Training a Support Vector Machine (SVM) requires the solution of a very large quadratic programming (QP) optimization problem. Sequential Minimal Optimization (SMO) is a decomposition-based algorithm which breaks this large QP problem into a series of smallest possible QP problems. However, it still costs O(n2) computation time. In our SVM implementation, we can do training with huge data sets in a distributed manner (by breaking the dataset into chunks, then using Message Passing Interface (MPI) to distribute each chunk to a different machine and processing SVM training within each chunk). In addition, we moved the kernel calculation part in SVM classification to a graphics processing unit (GPU) which has zero scheduling overhead to create concurrent threads. In this thesis, we will take advantage of this GPU architecture to improve the classification performance of SVM.