COMPUTE ATLAS Supercomputer supply-chain graph
1581 systems 558 sites Sourced data

Supercomputers/Systems/Summit

USA · decommissioned · mixed

Summit

Also known as OLCF-4

Operated by Oak Ridge National Laboratory at Oak Ridge Leadership Computing Facility .

Researched with AI assistance, checked by scripts. These rows were researched by AI agents reading public operator, laboratory and vendor sources, appended by scripts and checked by automated validation; no subject-matter expert has reviewed them (how this site is made). A mechanical second-reader pass has since fetched every cited source: 1350 of 1581 systems have a readable citation that names them, and 64 are genuinely weakly sourced. Every claim carries its source and a confidence tier. Treat anything below verified as a lead, not a citation.

Measured performance

Rmax
148.6 PFlop/s
HPL measured
Rpeak
200.8 PFlop/s
Rmax ÷ Rpeak
74.0%
Cores
2,414,592
Power
10.10 MW
Per watt
14.7 GF/W

Figures are as last publicly reported for the configuration described below, not a live measurement. Where a system was upgraded in place, the post-upgrade configuration is shown and the earlier one appears in the timeline.

How it compares

Ranked against the other machines in this dataset that report the same figure: those that first came into service within two years of Summit, and those in USA. Computed from our tables; a metric is ranked only when at least eight peers report it.

MetricValueAgainst its eraAgainst its country
Rmax148.6 PFlop/s#2 of 223 · top 1%#8 of 332 · top 2%
Energy efficiency14.7 GF/W#16 of 133 · top 11%#31 of 162 · top 19%
Rmax ÷ Rpeak74.0 %#71 of 216 · top 33%#171 of 331 · bottom 49%
Power draw10,096 kW#3 of 137 · top 1%#4 of 171 · top 2%

Every metric is ranked largest-first, so a high power-draw rank means a big power draw, not a good one. Rmax ÷ Rpeak uses HPL-measured results only. Rankings move as machines are added; they describe this dataset, not the world.

What this machine is made of

One row per supplier relationship. “Supplier at build” is the company that shipped the part at the time; where that company has since been acquired, the parent it rolls up to today is shown beside it. That distinction is what makes ticker-level aggregation possible across a thirty-year dataset.

Role Part Supplier at build Quantity Confidence Source
Integrator -
IBM Power System AC922
IBM - Reported olcf.ornl.govibm.com
Accelerator NVIDIA Tesla V100
Accelerators
CUDA (Volta GV100) · 5,120C · 1.53 GHz · 12 nm · 300 W
NVIDIA Volta GV100, 6 per node
NVIDIA 27,648
reported
Reported olcf.ornl.govibm.com
CPU IBM POWER9 22C
CPUs
POWER · 22C · 3.07 GHz · 14 nm · 190 W
IBM POWER9 22C 3.07GHz
IBM 9,216
reported
Reported olcf.ornl.govibm.com
Interconnect Mellanox InfiniBand EDR
Interconnect
Fat tree · 100 Gb/s per 4x port · About 0.9 microseconds
Dual-rail Mellanox EDR InfiniBand
Mellanox
NVDANASDAQ rolls up to NVIDIA
- Reported olcf.ornl.govibm.com
Storage IBM Spectrum Scale ESS (GPFS)
Storage
Parallel filesystem (GPFS), erasure-coded declustered RAID · 250 PB on Summit (Alpine) · About 2.5 TB/s
Alpine, IBM Spectrum Scale
IBM - Reported olcf.ornl.govibm.com
Cooling Direct-to-chip cold plate
Cooling
Cold plate on the package, air for the remainder · Removes roughly 50-80 percent of node heat to liquid
IBM Power System AC922: water-cooled cold plates on CPU and GPU
IBM - Reported olcf.ornl.govibm.com
Operating system Red Hat Enterprise Linux
Operating systems
Linux distribution · Open source, commercial support subscription
Red Hat Enterprise Linux for POWER
Red Hat
IBMNYSE rolls up to IBM
- Reported olcf.ornl.govibm.com

Engineering, funding and access

What Summit runs on, who paid for it and how people get time on it. Each line carries who said it and when.

Hardware

ItemDetailAs ofBasisSource
Compute nodes4,608 nodesEach node had 2 IBM POWER9 CPUs and 6 NVIDIA V100 GPUs, 27,648 GPUs in total; OLCF text elsewhere says about 4,6002024Operator statedolcf.ornl.gov
Storage capacity250 PBAlpine file system; node-local non-volatile memory acts as a burst buffer2024Operator statedolcf.ornl.gov
Interconnect and topologyDual-rail Mellanox EDR InfiniBandOLCF describes a non-blocking fat tree with 25 GB/s node injection bandwidth; NVLink links CPUs and GPUs inside a node2024Operator statedolcf.ornl.govtop500.org

Software

ItemDetailAs ofBasisSource
Parallel filesystemAlpine, 250 PB IBM Spectrum Scale (GPFS) with about 2.5 TB/s sequential bandwidth; HPSS for archive2024Operator statedolcf.ornl.gov
Programming and software stackRHEL 7.4 with IBM XL compilers, nvcc, ESSL, CUBLAS 9.2 and IBM Spectrum MPIAs reported with the TOP500 submission2019Operator statedtop500.org

Funding

ItemDetailAs ofBasisSource
Funded byU.S. Department of EnergyFunded under the CORAL collaboration of Oak Ridge, Argonne and Livermore2014-11Operator statedscience.osti.gov
ProcurementCORAL award announced 14 Nov 2014: $325 million to build Summit at ORNL and Sierra at LLNL, both on IBM Power, NVIDIA Volta GPUs and Mellanox interconnectThe $325 million is for both machines combined2014-11-14Operator statedscience.osti.gov

Access

ItemDetailAs ofBasisSource
How researchers get timeOLCF Early Science program from 1 Oct 2018, then INCITE allocations from 1 Jan 20192019Operator statedolcf.ornl.gov

Results

ItemDetailAs ofBasisSource
Notable resultLinpack of 148.6 PFlop/s, HPCG of 2.926 PFlop/s and measured power draw of 10.1 MW on the TOP500 submission2018-11Operator statedtop500.org

The read

IBM POWER9 hosts with six NVIDIA V100s per node, coherently attached over NVLink rather than PCIe: the last leadership-class system to use a non-x86, non-Arm host CPU. Its retirement in late 2024 removed the largest remaining block of POWER installed base from the graph in a single step, which is visible as a discontinuity on the CPU-architecture share chart and is worth flagging whenever that chart is quoted.

Timeline

  • 2014 Announced source
  • 2014-11-14 DOE announces the CORAL award of $325 million to build Summit at ORNL and Sierra at LLNL on IBM Power9, NVIDIA Volta and Mellanox technology source
  • 2018-06 List debut at #1 (June 2018) source
  • 2018-06 First operational source
  • 2018-10-01 Early Science projects begin running on Summit ahead of INCITE allocations in 2019 source
  • 2024-11 Retired after six years of operation source
  • 2024-11 Decommissioned source

Public list appearances

Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.

Site & facility

Oak Ridge Leadership Computing Facility

Location
Oak Ridge, TN, USA
Utility
Tennessee Valley Authority
Cooling
Direct liquid cooling (warm water)
Commissioned
1993

Change history

Nearby in scale

The machines closest to Summit by Rmax, smaller to larger.

Same site

Same operator

Same accelerator

  • ABCI · Japan, decommissioned, 2018
  • Ansel (LLNL) · USA, operational, 2018
  • Circe · USA, operational, 2018
  • CLAIX-2018 · Germany, decommissioned, 2018
  • COBRA · Germany, decommissioned, 2018
  • Govorun · Russia, upgraded, 2018

Source check

We fetched this system's own citations and recorded whether each page actually mentions it. This is a corroboration signal, not a fact check, and it is published so you can see how well the sourcing holds up rather than take it on trust.

Cited sourceResultFound on the page
ibm.comdoes not name this system-
olcf.ornl.govnames system, part and countIBM POWER9 22C, Mellanox InfiniBand EDR, IBM Spectrum Scale ESS (GPFS) · counts: 27648

Checked 2026-10-18 by pnpm verify. Re-run it and the table changes with the web.

Related news

  • Summit retired after six years at Oak Ridge · Oak Ridge Leadership Computing Facility
    The retirement removes the largest remaining block of IBM POWER installed base from the graph in a single step, which is visible as a discontinuity on any CPU-architecture share chart.

All news →

Further reading