Supercomputers/Systems/DelftBlue
Netherlands · operational · mixed
DelftBlue
Operated by TU Delft .
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
- -
- Rpeak
- 2.06 PFlop/s
- Rmax ÷ Rpeak
- -
- Cores
- 17,842
- Power
- -
- Per watt
- - 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 | - Fujitsu hardware with Intel processors | Fujitsu | - | Reported | tudelft.nl |
| Accelerator | - 10 Phase 1 GPU nodes with NVIDIA Tesla V100S | NVIDIA | - | Reported | doc.dhpc.tudelft.nl |
| Accelerator | - 10 Phase 2 GPU nodes with NVIDIA A100 80 GB (80 GPUs in total across both generations per TU Delft) | NVIDIA | - | Reported | doc.dhpc.tudelft.nltudelft.nl |
| CPU | - 218 compute-p1 nodes with Intel Xeon Gold 6248R (48 cores per node; quantity is nodes) | Intel | 218 reported | Reported | doc.dhpc.tudelft.nl |
| CPU | - 90 compute-p2 nodes with Intel Xeon Gold 6448Y (64 cores per node; quantity is nodes) | Intel | 90 reported | Reported | doc.dhpc.tudelft.nl |
| Interconnect | NVIDIA InfiniBand HDR Interconnect Fat tree or dragonfly+ · 200 Gb/s per port · About 0.6 microseconds Mellanox InfiniBand HDR100/HDR non-blocking network | NVIDIA | - | Reported | doc.dhpc.tudelft.nltudelft.nl |
| Operating system | - Red Hat Enterprise Linux 8 with Slurm | Red Hat | - | Reported | tudelft.nl |
The read
TU Delft's central research supercomputer, built on Fujitsu hardware with Intel processors. Phase 1 went into service in June 2022 with about 11,000 cores; Phase 2 came online in January 2024. The DHPC hardware page lists 338 compute nodes (218 with Intel Xeon Gold 6248R, 90 with Xeon Gold 6448Y, and high-memory nodes) with 17,842 compute cores and 1.45 PFlop/s theoretical CPU peak, plus 20 GPU nodes with 80 GPUs in total (NVIDIA Tesla V100S nodes and Phase 2 A100 80 GB nodes) adding 0.61 PFlop/s of double-precision peak. The Rpeak in this row is the sum of those two stated figures (over 2 PFlop/s per TU Delft). Interconnect is Mellanox InfiniBand HDR100/HDR, storage is 696 TB of BeeGFS, and the OS is Red Hat Enterprise Linux 8. Both sources are TU Delft pages, so the tier is reported.
Timeline
- 2022-06 First operational source
Change history
- 2026-10-02 Added DelftBlue (TU Delft)
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 |
|---|---|---|
| doc.dhpc.tudelft.nl | names system and part | NVIDIA InfiniBand HDR |
| tudelft.nl | names system and part | NVIDIA InfiniBand HDR |
Checked 2026-10-09 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.