An Analysis of System Balance and Architectural Trends Based on Top500 Supercomputers

Supercomputer design is a complex, multi-dimensional optimization process, wherein several subsystems need to be reconciled to meet a desired figure of merit performance for a portfolio of applications and a budget constraint. However, overall, the HPC community has been gravitating towards ever more Flops, at the expense of many other subsystems. To draw attention to overall system balance, in this paper, we analyze balance ratios and architectural trends in the world’s most powerful supercomputers. Specifically, we have collected the performance characteristics of systems between 1993 and 2019 based on the Top500 lists and then analyzed their architectures from diverse system design perspectives. Notably, our analysis studies the performance balance of the machines, across a variety of subsystems such as compute, memory, I/O, interconnect, intra-node connectivity and power. Our analysis reveals that balance ratios of the various subsystems need to be considered carefully alongside the application workload portfolio to provision the subsystem capacity and bandwidth specifications, which can help achieve optimal performance.

[1]  Xiaoming Zhang,et al.  Hybrid hierarchy storage system in MilkyWay-2 supercomputer , 2014, Frontiers of Computer Science.

[2]  Gilad Shainer,et al.  The high-speed networks of the Summit and Sierra supercomputers , 2020, IBM J. Res. Dev..

[3]  Oscar R. Hernandez,et al.  Titan : Early experience with the Cray XK 6 at Oak Ridge National Laboratory , 2012 .

[4]  Courtenay T. Vaughan,et al.  A Comparison of the Performance Characteristics of Capability and Capacity Class HPC Systems. , 2011 .

[5]  Bronis R. de Supinski,et al.  The Design, Deployment, and Evaluation of the CORAL Pre-Exascale Systems , 2018, SC18: International Conference for High Performance Computing, Networking, Storage and Analysis.

[6]  Valerio Pascucci,et al.  Analyzing Network Health and Congestion in Dragonfly-Based Supercomputers , 2016, 2016 IEEE International Parallel and Distributed Processing Symposium (IPDPS).

[7]  Christopher S. Daley,et al.  A Metric for Evaluating Supercomputer Performance in the Era of Extreme Heterogeneity , 2018, 2018 IEEE/ACM Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems (PMBS).

[8]  Hans Werner Meuer.  The TOP500 Project: Looking Back Over 15 Years of Supercomputing Experience , 2008, Informatik-Spektrum.

[9]  Marc Snir,et al.  The Future of Supercomputing , 2014, 2018 IEEE 25th International Conference on High Performance Computing (HiPC).

[10]  Philip Heidelberger,et al.  An Evaluation of Network Architectures for Next Generation Supercomputers , 2016, 2016 7th International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems (PMBS).

[11]  Erich Strohmaier,et al.  The TOP500 List and Progress in High-Performance Computing , 2015, Computer.

[12]  Jack J. Dongarra,et al.  The LINPACK Benchmark: past, present and future , 2003, Concurr. Comput. Pract. Exp..

[13]  Peter M. Kogge,et al.  Using the TOP500 to trace and project technology and architecture trends , 2011, 2011 International Conference for High Performance Computing, Networking, Storage and Analysis (SC).

[14]  Fei Meng,et al.  Functional Partitioning to Optimize End-to-End Performance on Many-core Architectures , 2010, 2010 ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis.

[15]  Amar Shan,et al.  Heterogeneous processing: a strategy for augmenting moore's law , 2006 .

[16]  Jesse Hanley,et al.  End-to-end I/O portfolio for the summit supercomputing ecosystem , 2019, SC.

[17]  R. Schaller,et al.  Moore's law: past, present and future , 1997 .

[18]  Robert B. Ross,et al.  On the role of burst buffers in leadership-class storage systems , 2012, 012 IEEE 28th Symposium on Mass Storage Systems and Technologies (MSST).

[19]  Neena Imam,et al.  Learning from Five-year Resource-Utilization Data of Titan System , 2019, 2019 IEEE International Conference on Cluster Computing (CLUSTER).

[20]  Robert W. Robey,et al.  Paradigmatic shifts for exascale supercomputing , 2012, The Journal of Supercomputing.

[21]  Denis Foley,et al.  Ultra-Performance Pascal GPU and NVLink Interconnect , 2017, IEEE Micro.

[22]  Sadaf R. Alam,et al.  Cray XT4: an early evaluation for petascale scientific simulation , 2007, Proceedings of the 2007 ACM/IEEE Conference on Supercomputing (SC '07).