Graph500


The Graph500 is a rating of supercomputer systems, focused on data-intensive loads. The project was announced on International Supercomputing Conference in June 2010. The first list was published at the ACM/IEEE Supercomputing Conference in November 2010. New versions of the list are published twice a year. The main performance metric used to rank the supercomputers is GTEPS.
Richard Murphy from Sandia, says that "The Graph500's goal is to promote awareness of complex data problems", instead of focusing on computer benchmarks like HPL, which TOP500 is based on.
Despite its name, there were several hundreds of systems in the rating, growing up to 174 in June 2014.
The algorithm and implementation that won the championship is published in the paper titled "Extreme scale breadth-first search on supercomputers".
There is also list Green Graph 500, which uses same performance metric, but sorts list according to performance per Watt, like Green 500 works with TOP500.

Benchmark

The benchmark used in Graph500 stresses the communication subsystem of the system, instead of counting double precision floating-point. It is based on a breadth-first search in a large undirected graph. There are three computation kernels in the benchmark: the first kernel is to generate the graph and compress it into sparse structures CSR or CSC ; the second kernel does a parallel BFS search of some random vertices ; the third kernel runs a single-source shortest paths computation. Six possible sizes of graph are defined: toy, mini, small, medium, large, and huge.
The reference implementation of the benchmark contains several versions:
The implementation strategy that have won the championship on the Japanese K computer is described in.

Top 10 ranking

2020

Arm-based Fugaku took the op spot of the list.

2016

According to June 2016 release of the list:
RankSiteMachine Number of nodesNumber of coresProblem scaleGTEPS
1Riken Advanced Institute for Computational ScienceK computer 829446635524038621.4
2National Supercomputing Center in WuxiSunway TaihuLight 40768105996804023755.7
3Lawrence Livermore National LaboratoryIBM Sequoia 9830415728644123751
4Argonne National LaboratoryIBM Mira 491527864324014982
5Forschungszentrum JülichJUQUEEN 16384262144385848
6CINECAFermi 8192131072372567
7Changsha, ChinaTianhe-2 8192196608362061.48
8CNRS/IDRIS-GENCITuring 409665536361427
8Science and Technology Facilities Council – Daresbury LaboratoryBlue Joule 409665536361427
8University of EdinburghDIRAC 409665536361427
8EDF R&DZumbrota 409665536361427
8Victorian Life Sciences Computation InitiativeAvoca 409665536361427

2014

According to June 2014 release of the list:
RankSiteMachine Number of nodesNumber of coresProblem scaleGTEPS
1RIKEN Advanced Institute for Computational ScienceK computer 655365242884017977.1
2Lawrence Livermore National LaboratoryIBM Sequoia 6553610485764016599
3Argonne National LaboratoryIBM Mira 491527864324014328
4Forschungszentrum JülichJUQUEEN 16384262144385848
5CINECAFermi 8192131072372567
6Changsha, ChinaTianhe-2 8192196608362061.48
7CNRS/IDRIS-GENCITuring 409665536361427
7Science and Technology Facilities Council - Daresbury LaboratoryBlue Joule 409665536361427
7University of EdinburghDIRAC 409665536361427
7EDF R&DZumbrota 409665536361427
7Victorian Life Sciences Computation InitiativeAvoca 409665536361427

2013

According to June 2013 release of the list:
RankSiteMachine Number of nodesNumber of coresProblem scaleGTEPS
1Lawrence Livermore National LaboratoryIBM Sequoia 6553610485764015363
2Argonne National LaboratoryIBM Mira 491527864324014328
3Forschungszentrum JülichJUQUEEN 16384262144385848
4RIKEN Advanced Institute for Computational ScienceK computer 65536524288405524.12
5CINECAFermi 8192131072372567
6Changsha, ChinaTianhe-2 8192196608362061.48
7CNRS/IDRIS-GENCITuring 409665536361427
7Science and Technology Facilities Council - Daresbury LaboratoryBlue Joule 409665536361427
7University of EdinburghDIRAC 409665536361427
7EDF R&DZumbrota 409665536361427
7Victorian Life Sciences Computation InitiativeAvoca 409665536361427