Examples
Worked examples
- Is an instance
A national HPC service publishes annual PUE and grid carbon intensity alongside its allocations.
- Is an instance
A research consortium prefers a warm-water-cooled cluster with heat re-use over a slightly faster air-cooled alternative.
Counter-examples
Looks similar, but isn't
- Not an instance
Marketing a cluster as green based solely on its placement next to a wind farm without efficiency commitments is not sustainable HPC.
- Not an instance
A traditional air-cooled machine room with no measurement is the opposite.
Editorial commentary
Sustainable HPC covers the operational and procurement decisions that determine a high-performance computing facility’s environmental footprint across energy, carbon, water and hardware lifecycle, while still delivering the scientific throughput researchers need. It spans data-centre design (free-air and warm-water cooling loops, and heat re-use into district or campus heating systems), workload scheduling (carbon-aware queues, shutting down or consolidating under-used nodes), procurement (weighing accelerators partly on FLOPs-per-watt rather than raw performance alone, and hardware refresh cycles that extend useful life rather than replacing on a fixed schedule regardless of condition), and increasingly the facility’s scope-3 embodied emissions — the carbon cost of manufacturing the hardware itself, which for a modern HPC cluster is frequently the largest single component of its lifecycle footprint, larger than its operational electricity use over a typical service life.
Real examples
The EuroHPC-funded LUMI supercomputer (Finland) and Adastra (France) are commonly cited flagship examples specifically because they route waste heat from cooling into district heating networks serving surrounding buildings, turning a cooling cost into a community energy input rather than pure waste. The Green500 list ranks the world’s most efficient supercomputers by performance-per-watt, distinct from the raw-performance Top500 ranking, and is a reference point institutions use when evaluating a procurement.
How it connects to other computing-sustainability concepts
Sustainable HPC draws on energy proportionality (right-sizing and consolidating allocations) and carbon-aware computing (scheduling around grid intensity) as operational tools, and its output feeds directly into a project’s compute carbon footprint and, at proposal stage, a climate impact statement.
What it isn’t
Marketing a cluster as sustainable purely on the basis of its physical location, such as adjacency to a wind farm, without measured efficiency commitments, published power usage effectiveness, or a carbon-aware scheduling policy does not meet the bar this term describes.
References
- Green500 list methodology; EuroHPC Joint Undertaking sustainability materials (LUMI, Adastra).
Also known as
Green HPC · Sustainable supercomputing
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="Sustainable HPC"
vocab-term-identifier="https://casrai.org/dictionary/term/sustainable-hpc" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/sustainable-hpc",
"name": "Sustainable HPC",
"identifier": "https://casrai.org/dictionary/term/sustainable-hpc",
"description": "High-performance computing operations and procurement choices that minimise energy use, carbon emissions, water use, and hardware waste while delivering scientific compute capacity.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/sustainable-research#set",
"url": "https://casrai.org/dictionary/term/sustainable-hpc",
"alternateName": [
"Green HPC",
"Sustainable supercomputing"
],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"author": {
"@id": "https://casrai.org/#editorial-team"
},
"datePublished": "2026-05-21T02:23:02",
"dateModified": "2026-09-05T14:27:19",
"inLanguage": "en-GB",
"isAccessibleForFree": true
}
