NVIDIA Hopper architecture
Transformer Engine with FP8.
- Introduced
- 2022
- Process
- TSMC 4N
- Named after
- Grace Hopper
- Products in database
- 3
Data center
| Model | Launch | Shader cores | Memory | Bandwidth | Power | US launch |
|---|---|---|---|---|---|---|
| H200 NVIDIA Hopper | 2024 | — | 141 GB HBM3e | 4,800 GB/s | up to 700 W | — |
| GH200 Grace Hopper NVIDIA Hopper | 2023 | — | — | 4,900 GB/s | 450–1000 W configurable (CPU + GPU + memory) | — |
| H100 SXM NVIDIA Hopper | 2022 | 16,896 CUDA | 80 GB HBM3 | 3,350 GB/s | up to 700 W | — |