HPCG Performance List was announced at a BOF at SC18
The HPCG results continue to be integrated into the TOP500 list. The 10th edition
list of results contains the complete list of submissions some of which
might have not made the TOP500 list.
The 10th edition of the bi-annual High Performance Conjugate Gradients (HPCG) Benchmark list
was announced on November 13, 2018 at the SC18 conference in Dallas, TX, USA.
The session entitled HPCG Benchmark Update
took place on Tuesday, November 13, 2018 in Room D167 at 12:15pm in Kay Bailey
Hutchison Convention Center Dallas.
This is the tenth list produced for the benchmark designed to complement the
traditional High Performance LINPACK (HPL) benchmark used as the official
metric for ranking the TOP500 systems. The first HPCG list was announced at ISC
2014 over four years ago, containing only 15 entries. Since then, HPCG list
was released bi-annually and the number of entries has steadily increased as
the new results became available. The current list contains 140 entries as HPCG
continues to gain traction in the HPC community.
It is now customary that the new list contains entries from many of the
TOP500's highest ranked systems, mainly the top-50. The current list is no
exception. However, the HPCG ranking continues to provide a significantly different
ordering of the machines when compared with the TOP500 ranking.
A major change in TOP500 ranking is now reflected in the HPCG results.
Summit and Sierra supercomputers lead the HPCG list with their score breaking
the 1 Pflop/s barrier and almost approaching 3 Pflop/s for the number 1 system.
Summary highlights from the 10th HPCG list include:
- Only US and Japanese systems in top 5.
- A few large machines received upgrades and higher HPCG scores including
Piz Daint, Electra, SuperMUC, Endeavor.
- The first ARM system made appearance on the list. It is called Astra and is hosted at Sandia National Laboratories.
- More than 140 systems are now recorded on the list despite retirement of some machines from the previous list.
- Mike Heroux, SNL, Introduction
- Daniel Ruiz Munoz, ARM, HPCG @ ARM
- Alexander Kalinkin, Intel, The Intel Optimized High Performance Conjugate Gradients
- Massimiliano Fatica, Everett Phillips, NVIDIA, HPCG Benchmark on Coral Systems
- Si Hammond, SNL (for HPE)
- Christoph Hagleitner, IBM Research, Zurich Research Laboratory, Porting Optimized HPCG 3.1 for IBM BG/Q to IBM POWER9