Supercomputers/Systems/Christofari
Russia · operational · ai training
Christofari
Operated by Sberbank at SberCloud (undisclosed site, Russia) .
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
- 6.67 PFlop/s
- Rpeak
- 8.79 PFlop/s
- Rmax ÷ Rpeak
- 75.9%
- Cores
- 99,600
- 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 | - NVIDIA DGX-2 based system designed by Sberbank and SberCloud together with NVIDIA (TOP500 manufacturer: NVIDIA) | NVIDIA | - | Reported | top500.orgfinancialit.net |
| Accelerator | NVIDIA Tesla V100 Accelerators CUDA (Volta GV100) · 5,120C · 1.53 GHz · 12 nm · 300 W NVIDIA Tesla V100 GPUs in NVIDIA DGX-2 nodes (node and GPU count not published) | NVIDIA | - | Verified | top500.orgfinancialit.net |
| CPU | - Intel Xeon Platinum 8168 24C 2.7GHz (TOP500 listing) | Intel | - | Reported | top500.org |
| Interconnect | Mellanox InfiniBand EDR Interconnect Fat tree · 100 Gb/s per 4x port · About 0.9 microseconds Mellanox InfiniBand EDR (TOP500 listing) | NVIDIA | - | Reported | top500.org |
The read
Christofari is the first of two NVIDIA DGX-based AI supercomputers built by Sberbank and its SberCloud subsidiary together with NVIDIA. It was presented on 8 November 2019 at the AI Journey conference and offered to clients from 12 December 2019 through the SberCloud platform. The TOP500 entry (SberCloud, manufacturer NVIDIA) lists NVIDIA DGX-2 nodes with Xeon Platinum 8168 24C 2.7GHz CPUs, NVIDIA Tesla V100 GPUs and Mellanox InfiniBand EDR, 99,600 cores, Rmax 6.67 PFlop/s and Rpeak 8.79 PFlop/s; it debuted at rank 29 on the November 2019 list, was still ranked 72nd in November 2021, and the TOP500 page reports rank 256 on the June 2026 list, which is the basis for the status date. Sberbank's own announcement, as carried by Financial IT, independently gives the DGX-2 and Tesla V100 basis and a 6.7 PFLOPs LINPACK result, so the tier is verified. Neither source publishes the DGX-2 node or GPU count, so no accelerator quantity is recorded; the exact data center location is not stated in the sources used here, so the site is left unlocated. Power is not listed by TOP500.
Timeline
- 2019-11 List debut at #29 (November 2019) source
- 2019-11-08 Announced source
- 2019-12 First operational source
Public list appearances
Pointers only: rank and edition, linking to the canonical entry. We do not reproduce list tables. See sourcing policy.
Site & facility
Change history
- 2026-10-01 Added Christofari (Sberbank)
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 |
|---|---|---|
| financialit.net | names system and part | NVIDIA Tesla V100 |
| top500.org | names system and part | NVIDIA Tesla V100, Mellanox InfiniBand EDR |
Checked 2026-10-09 by pnpm verify. Re-run it and the table changes with the web.
Further reading
Inside a node
Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.