TY - GEN
T1 - Vertical, Temporal, and Horizontal Scaling of Hierarchical Hypersparse GraphBLAS Matrices
AU - Kepner, Jeremy
AU - Davis, Tim
AU - Byun, Chansup
AU - Arcand, William
AU - Bestor, David
AU - Bergeron, William
AU - Gadepally, Vijay
AU - Houle, Michael
AU - Hubbell, Matthew
AU - Jones, Michael
AU - Klein, Anna
AU - Milechin, Lauren
AU - Mullen, Julie
AU - Prout, Andrew
AU - Reuther, Albert
AU - Rosa, Antonio
AU - Samsi, Siddharth
AU - Yee, Charles
AU - Michaleas, Peter
N1 - Funding Information:
This material is based upon work supported by the Assistant Secretary of Defense for Research and Engineering under Air Force Contract No. FA8702-15-D-0001, National Science Foundation CCF-1533644, and United States Air Force Research Laboratory Cooperative Agreement Number FA8750-19-2-1000. Any opinions, findings, conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Assistant Secretary of Defense for Research and Engineering, the National Science Foundation, or the United States Air Force. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.
Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - Hypersparse matrices are a powerful enabler for a variety of network, health, finance, and social applications. Hierarchical hypersparse GraphBLAS matrices enable rapid streaming updates while preserving algebraic analytic power and convenience. In many contexts, the rate of these updates sets the bounds on performance. This paper explores hierarchical hypersparse update performance on a variety of hardware with identical software configurations. The high-level language bindings of the GraphBLAS readily enable performance experiments on simultaneous diverse hardware. The best single process performance measured was 4,000,000 updates per second. The best single node performance measured was 170,000,000 updates per second. The hardware used spans nearly a decade and allows a direct comparison of hardware improvements for this computation over this time range; showing a 2x increase in single-core performance, a 3x increase in single process performance, and a 5x increase in single node performance. Running on nearly 2,000 MIT SuperCloud nodes simultaneously achieved a sustained update rate of over 200,000,000,000 updates per second. Hierarchical hypersparse GraphBLAS allows the MIT SuperCloud to analyze extremely large streaming network data sets.
AB - Hypersparse matrices are a powerful enabler for a variety of network, health, finance, and social applications. Hierarchical hypersparse GraphBLAS matrices enable rapid streaming updates while preserving algebraic analytic power and convenience. In many contexts, the rate of these updates sets the bounds on performance. This paper explores hierarchical hypersparse update performance on a variety of hardware with identical software configurations. The high-level language bindings of the GraphBLAS readily enable performance experiments on simultaneous diverse hardware. The best single process performance measured was 4,000,000 updates per second. The best single node performance measured was 170,000,000 updates per second. The hardware used spans nearly a decade and allows a direct comparison of hardware improvements for this computation over this time range; showing a 2x increase in single-core performance, a 3x increase in single process performance, and a 5x increase in single node performance. Running on nearly 2,000 MIT SuperCloud nodes simultaneously achieved a sustained update rate of over 200,000,000,000 updates per second. Hierarchical hypersparse GraphBLAS allows the MIT SuperCloud to analyze extremely large streaming network data sets.
KW - GraphBLAS
KW - horizontal scaling
KW - hypersparse matrices
KW - streaming graphs
KW - vertical scaling
UR - http://www.scopus.com/inward/record.url?scp=85123503151&partnerID=8YFLogxK
U2 - 10.1109/HPEC49654.2021.9622802
DO - 10.1109/HPEC49654.2021.9622802
M3 - Conference contribution
AN - SCOPUS:85123503151
T3 - 2021 IEEE High Performance Extreme Computing Conference, HPEC 2021
BT - 2021 IEEE High Performance Extreme Computing Conference, HPEC 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2021 IEEE High Performance Extreme Computing Conference, HPEC 2021
Y2 - 20 September 2021 through 24 September 2021
ER -