Supercomputers/Systems/Skipjack (JHU)
USA · operational · ai training
Skipjack (JHU)
Also known as JHU ARCH DSAI cluster; DSAI Cluster (JHU)
Operated by Johns Hopkins University .
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
- -
- Rmax ÷ Rpeak
- -
- Cores
- 18,836
- 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 |
|---|---|---|---|---|---|
| Accelerator | NVIDIA L40S Accelerators CUDA (Ada Lovelace) · 350 W 8 nodes with eight NVIDIA L40S 48GB GPUs each | NVIDIA | 64 reported | Reported | docs.arch.jhu.edu |
| Accelerator | - 15 nodes with eight NVIDIA A100 80GB GPUs each | NVIDIA | 120 reported | Reported | docs.arch.jhu.edu |
| Accelerator | - 16 nodes with four NVIDIA H100 80GB GPUs each | NVIDIA | 64 reported | Reported | docs.arch.jhu.edu |
| Accelerator | - 16 nodes with four NVIDIA H100-NVL 96GB GPUs each | NVIDIA | 64 reported | Reported | docs.arch.jhu.edu |
| Accelerator | NVIDIA H200 SXM Accelerators CUDA (Hopper GH100) · 16,896C · 1.98 GHz · TSMC 4N · 700 W 6 nodes with four NVIDIA H200 GPUs each | NVIDIA | 24 reported | Reported | docs.arch.jhu.edu |
| Accelerator | NVIDIA B200 SXM Accelerators CUDA (Blackwell) · TSMC 4NP · 1,000 W 16 nodes with eight NVIDIA B200 GPUs each | NVIDIA | 128 reported | Reported | docs.arch.jhu.edu |
| Accelerator | NVIDIA B300 SXM Accelerators CUDA (Blackwell Ultra) · TSMC 4NP · 1,100 W 1 node with eight NVIDIA B300 GPUs | NVIDIA | 8 reported | Reported | docs.arch.jhu.edu |
| Accelerator | - 3 nodes with 20 NVIDIA RTX PRO 6000 GPUs in total | NVIDIA | 20 reported | Reported | docs.arch.jhu.edu |
| CPU | - Intel Xeon Platinum 8480+ 56-core, 80 CPU nodes | Intel | - | Reported | docs.arch.jhu.edu |
| CPU | - AMD EPYC 9534 64-core and AMD EPYC 7443 24-core hosts on GPU nodes | AMD | - | Reported | docs.arch.jhu.edu |
| Storage | WEKA Data Platform Storage Parallel filesystem, software-defined 5 PB of WEKA storage backed by a parallel file system | WEKA | - | Reported | docs.arch.jhu.edu |
AI datacenter metrics
Figures beyond an accelerator count, each with who stated it. What can and cannot be compared across AI datacenters.
| Metric | Value | As of | Basis | Note | Source |
|---|---|---|---|---|---|
| Accelerators | 492 accelerators | 2026-10 | Operator stated | Page total: 64 L40S, 120 A100, 64 H100, 64 H100-NVL, 24 H200, 128 B200, 8 B300 and 20 RTX PRO 6000 GPUs. | docs.arch.jhu.edu |
The read
Johns Hopkins University's ARCH cluster for DSAI members, documented first as the DSAI cluster and now as Skipjack, to which all DSAI members are invited to migrate. The hardware summary lists 80 CPU nodes with Intel Xeon Platinum 8480+ and GPU nodes of 8 L40S, 15 A100, 16 H100, 16 H100-NVL, 6 H200, 3 RTX PRO 6000, 16 B200 and 1 B300 machines, for 18,836 cores across 161 nodes and 492 GPUs (64 L40S, 120 A100, 64 H100, 64 H100-NVL, 24 H200, 128 B200, 8 B300 and 20 RTX PRO 6000). An earlier summary listed 20,676 cores and 312 GPUs. No FLOPS figure is published.
Timeline
No dated events recorded.
Change history
- 2026-10-07 Added DSAI Cluster (JHU)
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.arch.jhu.edu | names system, part and count | NVIDIA L40S, NVIDIA H200 SXM, NVIDIA B200 SXM, NVIDIA B300 SXM · counts: 64, 24, 128, 8 |
Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.