Cosmos (SDSC)

Also known as Cosmos; SDSC Cosmos

Operated by SDSC at San Diego Supercomputer Center .

Phase 1 seed dataset, compiled by hand. These rows were built from public operator, laboratory and vendor sources. A mechanical second-reader pass has since fetched every cited source: 1092 of 1153 systems have a readable citation that names them, and 40 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: 5.18 PFlop/s
Rpeak: 10.41 PFlop/s
Rmax ÷ Rpeak: 49.8%
Cores: 42,336
Power: 99 kW
Per watt: 52.3 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.

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	- HPE Cray Supercomputing EX2500, 42 nodes	Hewlett Packard Enterprise HPENYSE	-	Reported	sdsc.edu top500.org
Accelerator	AMD Instinct MI300A Accelerators ROCm (CDNA 3) · TSMC N5/N6 · 550 W 168 AMD Instinct MI300A APUs, 4 per node	AMD AMDNASDAQ	168 reported	Reported	sdsc.edu
CPU	AMD 4th Gen EPYC 24C (in MI300A) CPUs x86-64 · 24C · 1.8 GHz · TSMC N5 24 Zen4 cores per APU (integrated in MI300A)	AMD AMDNASDAQ	168 reported	Reported	sdsc.edu
Interconnect	HPE Slingshot-11 Interconnect Dragonfly · 200 Gb/s per port · Around 2 microseconds HPE Slingshot dragonfly, 200 GB/s links (TOP500: Slingshot-11)	Hewlett Packard Enterprise HPENYSE	-	Reported	sdsc.edu top500.org
Storage	- 200 TB usable VAST flash performance tier	VAST Data	-	Reported	sdsc.edu
Operating system	SUSE Linux Enterprise Server Operating systems Linux distribution · Open source, commercial support subscription SLES 15	SUSE	-	Reported	top500.org

The read

NSF-funded (award 2404323) HPE Cray Supercomputing EX2500 system at SDSC with 42 nodes of four AMD Instinct MI300A APUs each, 168 APUs in total, on a Slingshot dragonfly fabric with 200 TB of VAST flash and 5 PB of Ceph. SDSC counts 24 Zen4 cores and 228 GPU compute units per APU. TOP500 lists 42,336 cores, 5.18 PFlop/s Rmax, 10.41 PFlop/s Rpeak and about 99 kW, 272nd in November 2025 and 297th in June 2026. SDSC and TOP500 agree on the platform, but the HPL figures come from TOP500 alone, so the tier is reported.

Timeline

2024 Announced source
2025 First operational source
2025-11 List debut at #272 (November 2025) source

Public list appearances

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

#272: TOP500, November 2025 →

Site & facility

San Diego Supercomputer Center

Location: La Jolla, CA, USA
Commissioned: 2020

sdsc.edu

Change history

2026-10-02 Added Cosmos, SDSC's NSF-funded MI300A system, 272nd on the November 2025 TOP500

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 source	Result	Found on the page
top500.org	names system and part	AMD Instinct MI300A, HPE Slingshot-11
sdsc.edu	names system, part and count	AMD Instinct MI300A, SUSE Linux Enterprise Server · counts: 168, 168

Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.