Abstract
Two-tiered direct network topologies such as Dragonflies have been proposed for future post-petascale and exascale machines, since they provide a high-radix, low-diameter, fast interconnection network. Such topologies call for redesigning MPI collective communication algorithms in order to attain the best performance. Yet as increasingly more applications share a machine, it is not clear how these topology-aware algorithms will react to interference with concurrent jobs accessing the same network. In this paper, we study three topology-aware broadcast algorithms, including one designed by ourselves. We evaluate their performance through event-driven simulation for small- and large-sized broadcasts (in terms of both data size and number of processes). We study the effect of different routing mechanisms on the topology-aware collective algorithms, as well as their sensitivity to network contention with other jobs. Our results show that while topology-aware algorithms dramatically reduce link utilization, their advantage in terms of latency is more limited.
Original language | English (US) |
---|---|
Title of host publication | Proceedings - 2016 IEEE International Conference on Cluster Computing, CLUSTER 2016 |
Publisher | Institute of Electrical and Electronics Engineers Inc. |
Pages | 40-49 |
Number of pages | 10 |
ISBN (Electronic) | 9781509036530 |
DOIs | |
State | Published - Dec 6 2016 |
Event | 2016 IEEE International Conference on Cluster Computing, CLUSTER 2016 - Taipei, Taiwan, Province of China Duration: Sep 13 2016 → Sep 15 2016 |
Other
Other | 2016 IEEE International Conference on Cluster Computing, CLUSTER 2016 |
---|---|
Country/Territory | Taiwan, Province of China |
City | Taipei |
Period | 9/13/16 → 9/15/16 |
ASJC Scopus subject areas
- Software
- Hardware and Architecture
- Signal Processing