Exploring Parallel Bitonic Sort on a Migratory Thread Architecture

摘要: Large scale, data-intensive applications pose challenges to systems with a traditional memory hierarchy due their unstructured data sources and irregular access patterns. In response, that employ migratory threads have been proposed mitigate bottlenecks as well reduce energy consumption. One such system is the Emu Chick, which migrates small program context being referenced in access. Sorting an unordered list of elements critical kernel for countless applications, graph processing tensor decomposition. As can be considered highly suitable thread architecture, it imperative understand performance sorting algorithms on these systems. this paper, we implement parallel bitonic sort target Chick system. We investigate explicit comparison-based approach network implementation. Furthermore, explore two different layouts network, namely cyclic blocked. From results our study, find while migrations dictate overall application, cost creation management out-grow migration.