@inproceedings{10711053bfd241a2a932780a4ed9dcbf,
title = "Dataflow Execution of Hierarchically Tiled Arrays",
abstract = "As the parallelism in high-performance supercomputers continues to grow, new programming models become necessary to maintain programmer productivity at today{\textquoteright}s levels. Dataflow is a promising execution model because it can represent parallelism at different granularity levels and to dynamically adapt for efficient execution. The downside is the low-level programming interface inherent to dataflow. We present a strategy to translate programs written in Hierarchically Tiled Arrays (HTA) to the dataflow API of Open Community Runtime (OCR) system. The goal is to enable program development in a convenient notation and at the same time take advantage of the benefits of a dataflow runtime system. Using HTA produces more comprehensive codes than those written using the dataflow runtime programming interface. Moreover, the experiments show that, for applications with high asynchrony and sparse data dependences, our implementation delivers superior performance than OpenMP using parallel for loops.",
keywords = "Dataflow, High-level programming abstraction, Parallel algorithm, Parallel programming",
author = "Yang, {Chih Chieh} and Pichel, {Juan C.} and Padua, {David A.}",
note = "Publisher Copyright: {\textcopyright} 2019, Springer Nature Switzerland AG.; 25th International European Conference on Parallel and Distributed Computing, Euro-Par 2019 ; Conference date: 26-08-2019 Through 30-08-2019",
year = "2019",
doi = "10.1007/978-3-030-29400-7_22",
language = "English (US)",
isbn = "9783030293994",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "304--316",
editor = "Ramin Yahyapour",
booktitle = "Euro-Par 2019",
address = "Germany",
}