Compare GPUs
| Spec | NVIDIA H100 PCIe | A100 |
|---|---|---|
| Provider | RunPod | AWS |
| VRAM | 80 GB | 80 GB |
| Architecture | — | — |
| Compute cap. | — | — |
| Mem. bandwidth | — | — |
| GPUs / instance | 1 | 1 |
| vCPU | — | — |
| System RAM | — | — |
| Network | — | — |
| CUDA cores | 14,592 | — |
| Tensor cores | 456 | — |
| TDP | 350 W | — |
| Interconnect | PCIe | — |
| Per hour | $2.49/hr | $0.64/hr |
| Per month | $1817.70 | — |
| Available | Yes | Yes |
| Regions | — | — |
Observed · read from provider listings, not ranked