Parts/Accelerators/NVIDIA A100 SXM4 80GB
Accelerators · Ampere · 2020
NVIDIA A100 SXM4 80GB
A100 SXM module with 80 GB of HBM, eight per node in HGX/DGX-class training servers.
Why this part
Added for the Seren and Kalos clusters, whose paper states eight A100-SXM 80GB GPUs per node joined by NVLink and NVSwitch. No release year is recorded because the cited paper does not give one.
Systems using this part
- Systems
- 11
- Rmax underneath
- 54.16 PFlop/s
- Units recorded
- 6,192
8 of 11 edges disclose a quantity
| System | Role | Country | Units | System Rmax | Supplied by | Rank |
|---|---|---|---|---|---|---|
| SK Telecom Titan 2022-05 | Accelerator | South Korea | 1,040 | 19.53 PFlop/s | NVIDIA | #92 → |
| Champollion 2022-05 | Accelerator | France | 160 | 2.32 PFlop/s | NVIDIA | #10 → |
| Tethys 2021-11 | Accelerator | USA | - | 2.26 PFlop/s | NVIDIA | #3 → |
| Hortense 2021 | Accelerator | Belgium | 80 | - | NVIDIA | - |
| Voyager-EUS2 2021 | Accelerator | USA | - | 30.05 PFlop/s | NVIDIA | #10 → |
| Mahuika 2018 | Accelerator | New Zealand | - | - | NVIDIA | - |
| Seren (Shanghai AI Lab) | Accelerator | China | 2,288 | - | NVIDIA | - |
| Kalos (Shanghai AI Lab) | Accelerator | China | 2,416 | - | NVIDIA | - |
| Imperial HX1 | Accelerator | United Kingdom | 60 | - | NVIDIA | - |
| Quest (Northwestern) | Accelerator | USA | 72 | - | NVIDIA | - |
| CSF3 | Accelerator | United Kingdom | 76 | - | NVIDIA | - |
“Units recorded” sums only the edges where a public source states a quantity. It is a floor, not a total, and should never be read as installed base.