Supercomputers/Systems/PARAM Siddhi-AI
India · operational · ai training
PARAM Siddhi-AI
Also known as PARAM Siddhi
Operated by C-DAC at C-DAC (Pashan, Pune) .
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
- 4.62 PFlop/s
- Rpeak
- 5.27 PFlop/s
- Rmax ÷ Rpeak
- 87.7%
- Cores
- 41,664
- 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 A100 SXM4 40GB Accelerators CUDA (Ampere GA100) · 6,912C · 1.41 GHz · 7 nm · 400 W 42 NVIDIA DGX A100 nodes, eight A100 SXM4 40GB GPUs per node | NVIDIA | 336 estimated | Verified | devdiscourse.comtop500.org |
| CPU | AMD EPYC 7742 (Rome) CPUs x86-64 · 64C · 2.25 GHz · 7 nm · 225 W AMD EPYC 7742 host CPUs across the 42 DGX A100 nodes; no source states the exact chip count | AMD | - | Verified | top500.orgdevdiscourse.com |
| Interconnect | NVIDIA InfiniBand HDR Interconnect Fat tree or dragonfly+ · 200 Gb/s per port · About 0.6 microseconds Mellanox HDR InfiniBand fabric | NVIDIA | - | Verified | top500.orgdevdiscourse.com |
| Operating system | Ubuntu Server Operating systems Linux distribution · Open source, free with paid support tiers Ubuntu 20.04.1 LTS | Canonical | - | Reported | top500.org |
The read
C-DAC's HPC-AI system at the National PARAM Supercomputing Facility in Pune, built on NVIDIA's DGX SuperPOD reference architecture: 42 DGX A100 nodes, each with two AMD EPYC 7742 hosts and eight A100 GPUs (the only A100 SXM4 variant that existed in mid-2020, before the 80GB part shipped), linked by Mellanox HDR InfiniBand and layered with C-DAC's own HPC-AI engine, software frameworks and cloud platform. NVIDIA's own commissioning release and TOP500 independently agree on the vendor, node count, host CPU and fabric, so the tier is verified; TOP500 alone gives the exact Rmax, Rpeak and core count. It debuted 62nd on the November 2020 TOP500, India's best-ever ranking at the time, though some Indian government retrospectives round that to 63rd. From 2023 it has shared its Pune facility with the newer AIRAWAT system; C-DAC markets a combined '410 AI petaflops mixed precision' across the two machines, a figure this dataset does not reproduce since TOP500 measures each system separately.
Timeline
- 2020-06 Announced source
- 2020-11 List debut at #62 (November 2020) source
- 2020-11 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
Where sources disagree
We record conflicts instead of picking a winner quietly. All open questions.
- Peak double-precision performance: 5.27 PFlop/s (TOP500 Rpeak) against 6.5 PFlop/s Peak DP (NVIDIA commissioning release). NVIDIA's own release and TOP500's measured Rpeak disagree on the machine's double-precision peak; TOP500's figure is recorded as Rpeak.
Change history
- 2026-09-22 Added with a rank pointer; NVIDIA's DP peak figure conflicts with TOP500's Rpeak
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 |
|---|---|---|
| devdiscourse.com | names system and part | NVIDIA A100 SXM4 40GB, NVIDIA InfiniBand HDR |
| top500.org | names system and part | NVIDIA A100 SXM4 40GB, AMD EPYC 7742 (Rome), NVIDIA InfiniBand HDR, Ubuntu Server |
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.
Programming a supercomputer
The software stack beneath the application, and why the OS is now a sovereignty signal as much as a technical choice.