Abstract
Stream processing is currently central to handle large volumes of data generated at high rates. However, the efficient processing of such quantity of data demands massively parallel hardware. The usual approach is to rely on clusters of multi-processors, where network communication may become a bottleneck. Some work has also been done in the GPU computing field. Yet, the GPUs’ programming complexity and the existence of synchronization-related overheads, when the streaming graph scales, have hampered the integration of GPUs in the Big Data streaming frameworks. In this paper we explore the unique characteristics of the Intel Xeon Phi processor to develop a stream processing framework for hybrid CPU/Intel Xeon Phi systems. We built atop the Intel Threading Building Blocks library and the Marrow algorithmic skeleton framework to offer an easily programmable high performance system. Our experimental results show that offloading the computationally heavy nodes of a streaming graph to the Xeon Phi may earn considerable speed-ups. Furthermore, additional gains may be obtained by sharing the processing load between the CPU(s) and the Xeon Phi processor(s).
| Original language | English |
|---|---|
| Title of host publication | Euro-Par 2018: Parallel Processing - 24th International Conference on Parallel and Distributed Computing, Proceedings |
| Editors | M. Torquati, M. Aldinucci, L. Padovani |
| Publisher | Springer Verlag |
| Pages | 796-810 |
| Number of pages | 15 |
| ISBN (Print) | 9783319969824 |
| DOIs | |
| Publication status | Published - 2018 |
| Event | 24th International European Conference on Parallel and Distributed Computing, Euro-Par 2018 - Turin, Italy Duration: 27 Aug 2018 → 31 Aug 2018 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Publisher | Springer Verlag |
| Volume | 11014 LNCS |
| ISSN (Print) | 0302-9743 |
| ISSN (Electronic) | 1611-3349 |
Conference
| Conference | 24th International European Conference on Parallel and Distributed Computing, Euro-Par 2018 |
|---|---|
| Country | Italy |
| City | Turin |
| Period | 27/08/18 → 31/08/18 |
Fingerprint
Keywords
- Algorithmic skeletons
- Intel Xeon Phi
- Parallel computing
- Stream processing
Cite this
}
Stream Processing on Hybrid CPU/Intel® Xeon Phi™ Systems. / Ferrão, Paulo; Marques, Hélder; Paulino, Hervé.
Euro-Par 2018: Parallel Processing - 24th International Conference on Parallel and Distributed Computing, Proceedings. ed. / M. Torquati; M. Aldinucci; L. Padovani. Springer Verlag, 2018. p. 796-810 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11014 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Stream Processing on Hybrid CPU/Intel® Xeon Phi™ Systems
AU - Ferrão, Paulo
AU - Marques, Hélder
AU - Paulino, Hervé
N1 - info:eu-repo/grantAgreement/FCT/5876/147279/PT#
PY - 2018
Y1 - 2018
N2 - Stream processing is currently central to handle large volumes of data generated at high rates. However, the efficient processing of such quantity of data demands massively parallel hardware. The usual approach is to rely on clusters of multi-processors, where network communication may become a bottleneck. Some work has also been done in the GPU computing field. Yet, the GPUs’ programming complexity and the existence of synchronization-related overheads, when the streaming graph scales, have hampered the integration of GPUs in the Big Data streaming frameworks. In this paper we explore the unique characteristics of the Intel Xeon Phi processor to develop a stream processing framework for hybrid CPU/Intel Xeon Phi systems. We built atop the Intel Threading Building Blocks library and the Marrow algorithmic skeleton framework to offer an easily programmable high performance system. Our experimental results show that offloading the computationally heavy nodes of a streaming graph to the Xeon Phi may earn considerable speed-ups. Furthermore, additional gains may be obtained by sharing the processing load between the CPU(s) and the Xeon Phi processor(s).
AB - Stream processing is currently central to handle large volumes of data generated at high rates. However, the efficient processing of such quantity of data demands massively parallel hardware. The usual approach is to rely on clusters of multi-processors, where network communication may become a bottleneck. Some work has also been done in the GPU computing field. Yet, the GPUs’ programming complexity and the existence of synchronization-related overheads, when the streaming graph scales, have hampered the integration of GPUs in the Big Data streaming frameworks. In this paper we explore the unique characteristics of the Intel Xeon Phi processor to develop a stream processing framework for hybrid CPU/Intel Xeon Phi systems. We built atop the Intel Threading Building Blocks library and the Marrow algorithmic skeleton framework to offer an easily programmable high performance system. Our experimental results show that offloading the computationally heavy nodes of a streaming graph to the Xeon Phi may earn considerable speed-ups. Furthermore, additional gains may be obtained by sharing the processing load between the CPU(s) and the Xeon Phi processor(s).
KW - Algorithmic skeletons
KW - Intel Xeon Phi
KW - Parallel computing
KW - Stream processing
UR - http://www.scopus.com/inward/record.url?scp=85052930688&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-96983-1_56
DO - 10.1007/978-3-319-96983-1_56
M3 - Conference contribution
SN - 9783319969824
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 796
EP - 810
BT - Euro-Par 2018: Parallel Processing - 24th International Conference on Parallel and Distributed Computing, Proceedings
A2 - Torquati, M.
A2 - Aldinucci, M.
A2 - Padovani, L.
PB - Springer Verlag
ER -