Supercomputers/Systems/INL Falcon
USA · decommissioned · hpc
INL Falcon
Also known as Falcon
Operated by INL .
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
- 1.09 PFlop/s
- Rpeak
- 1.17 PFlop/s
- Rmax ÷ Rpeak
- 92.8%
- Cores
- 34,884
- Power
- 354 kW
- Per watt
- 3.1 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 INL Falcon, and those in USA. Computed from our tables; a metric is ranked only when at least eight peers report it.
| Metric | Value | Against its era | Against its country |
|---|---|---|---|
| Rmax | 1.09 PFlop/s | #66 of 245 · top 27% | #174 of 332 · bottom 49% |
| Energy efficiency | 3.1 GF/W | #15 of 198 · top 7% | #61 of 162 · top 37% |
| Rmax ÷ Rpeak | 92.8 % | #12 of 243 · top 5% | #9 of 331 · top 2% |
| Power draw | 354 kW | #122 of 199 · bottom 40% | #111 of 171 · bottom 36% |
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 | - SGI ICE X | SGI HPENYSE
rolls up to Hewlett Packard Enterprise | - | Reported | top500.orgdocs.c3plus3.org |
| CPU | - Intel Xeon E5-2680v3 12C 2.5GHz | Intel | - | Reported | docs.c3plus3.orgtop500.org |
| CPU | Intel Xeon E5-2695 v4 CPUs x86-64 · 18C · 2.1 GHz · 14 nm · 120 W Intel Xeon E5-2695v4 18C 2.1GHz | Intel | - | Reported | top500.orgdocs.c3plus3.org |
| Interconnect | Mellanox InfiniBand FDR Interconnect Fat tree · 56 Gb/s per 4x port · About 1 microsecond Infiniband FDR | Mellanox NVDANASDAQ
rolls up to NVIDIA | - | Reported | top500.org |
| Operating system | SUSE Linux Enterprise Server Operating systems Linux distribution · Open source, commercial support subscription SUSE Linux Enterprise Server 11 | SUSE | - | Reported | top500.org |
Engineering, funding and access
What INL Falcon runs on, who paid for it and how people get time on it. Each line carries who said it and when.
Hardware
| Item | Detail | As of | Basis | Source |
|---|---|---|---|---|
| Compute nodes | 684 nodesOriginal 10-rack configuration with dual 12-core Xeon E5-2680v3. A 2017 expansion added 4 racks (19,200 cores) but no new node count is stated. | 2014-11 | Operator stated | docs.c3plus3.org |
| Interconnect and topology | InfiniBand | 2022 | Operator stated | docs.c3plus3.org |
Software
| Item | Detail | As of | Basis | Source |
|---|---|---|---|---|
| Parallel filesystem | Lustre parallel file system, 1.3 PB | 2022 | Operator stated | docs.c3plus3.org |
| Programming and software stack | Rocky Linux 8.5 on compute nodes (rebuilt 2022) | 2022-04 | Operator stated | docs.c3plus3.org |
Funding
| Item | Detail | As of | Basis | Source |
|---|---|---|---|---|
| Procurement | Bought by Idaho National Laboratory in 2014 as an SGI ICE X, about five times faster than predecessor Fission | 2014 | Operator stated | docs.c3plus3.org |
Access
| Item | Detail | As of | Basis | Source |
|---|---|---|---|---|
| How researchers get time | Idaho C3+3 collaboration of University of Idaho, Boise State University and Idaho State University, from April 2022 | 2022-04 | Operator stated | docs.c3plus3.org |
Results
| Item | Detail | As of | Basis | Source |
|---|---|---|---|---|
| Notable result | Ranked 97th on the TOP500 and 31st on the Green500 when it came online in November 2014 (2014) | 2014-11 | Operator stated | docs.c3plus3.org |
The read
Idaho National Laboratory's SGI ICE X cluster, bought in 2014 as the successor to a machine called Fission and succeeded in turn by Lemhi (lemhi-inl). The C3+3 documentation describes the original build as 10 racks and 684 nodes of dual 12-core Xeon E5-2680v3 (16,416 cores, about 657 TFlop/s peak), later extended by four racks and re-equipped with 18-core Xeon E5-2695v4 chips for a stated 1.17 PFlop/s peak. TOP500 follows the growth: 16,416 cores at 511.24 TFlop/s Rmax in 2015, 19,200 and 25,056 cores in 2016, and 34,884 cores at 1.09 PFlop/s Rmax, 1.17 PFlop/s Rpeak and 354 kW on the November 2017 list (rank 144), which is the configuration recorded here. The documentation says 19,200 cores after the upgrade, which conflicts with TOP500's 34,884, so the tier stays reported. In April 2022 INL handed the hardware to the University of Idaho, Boise State University and Idaho State University to run collaboratively; INL's own service is recorded as ending then, the same convention used for Lemhi. The INL site is not named in a fetched page.
Timeline
- 2014 First operational source
- 2014-11 Came online ranked 97th on the TOP500 and 31st on the Green500 source
- 2017 Four racks were added, bringing the core count to 19,200, and processors were replaced with 18-core Xeon E5-2695v4, lifting peak to 1.17 petaflops source
- 2018-06 List debut at #219 (June 2018) source
- 2022-04 INL turned Falcon over to the Universities of Idaho and Idaho State and Boise State to run collaboratively, rebuilt on Rocky Linux 8.5 source
- 2022-04 Decommissioned source
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Change history
- 2026-10-17 INL Falcon added from its source pages
Nearby in scale
The machines closest to INL Falcon by Rmax, smaller to larger.
- DiRAC BlueGene/Q · United Kingdom, decommissioned · 1.07 PFlop/s
- Turing (IDRIS) · France, decommissioned · 1.07 PFlop/s
- BeBop · USA, operational · 1.07 PFlop/s
- Cartesius · Netherlands, decommissioned · 1.09 PFlop/s
- Magnus · Australia, decommissioned · 1.1 PFlop/s
- Cielo · USA, decommissioned · 1.11 PFlop/s
Related systems
Same operator
- Fission (INL) · USA, operational, 2011
- Lemhi · USA, decommissioned, 2018
- Sawtooth · USA, operational, 2019
- Ice Storm · USA, operational, 2007
- Teton · USA, operational, 2026
- Wind River · USA, operational
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 |
|---|---|---|
| docs.c3plus3.org | names system and part | Intel Xeon E5-2695 v4, Mellanox InfiniBand FDR |
| top500.org | names system and part | Intel Xeon E5-2695 v4, Mellanox InfiniBand FDR, SUSE Linux Enterprise Server |
Checked 2026-10-18 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.
Inside a node
Latency-optimised versus throughput-optimised processors, NUMA, and the end of the host-to-device copy.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.