| H100 SXM | H100 PCIe | |
|---|---|---|
| FP64 | 34 teraFLOPS | 26 teraFLOPS |
| FP64 Tensor Core | 67 teraFLOPS | 51 teraFLOPS |
| FP32 | 67 teraFLOPS | 51 teraFLOPS |
| TF32 Tensor Core | 989 teraFLOPS* | 756teraFLOPS* |
| BFLOAT16 Tensor Core | 1979 teraFLOPS* | 1,513 teraFLOPS* |
| FP16 Tensor Core | 1979 teraFLOPS* | 1,513 teraFLOPS* |
| FP8 Tensor Core | 3958 teraFLOPS* | 3026 teraFLOPS* |
| INT8 Tensor Core | 3958 TOPS* | 3026 TOPS* |
| GPU | 80GB | 80GB |
| GPU | 3.35TB/s | 2TB/s |
| 7 NVDEC 7 JPEG |
7 NVDEC 7 JPEG |
|
| (TDP) | 700W | 300-350 W |
| GPU | 最多 7 个 MIG @每个 10GB | |
| SXM | PCIe |
|
| 互连技术 | NVLink:900GB/s PCIe 5.0:128GB/s | NVLink:600GB/s PCIe 5.0:128GB/s |
| 服务器选项 | NVIDIA HGX™ H100 合作伙伴和配备 4 或 8 个 GPU 的 NVIDIA 认证系统™ ,配备 8 个 GPU 的 NVIDIA DGX™ H100 | 搭载 1 至 8 个 GPU 的合作伙伴系统及 NVIDIA 认证系统 |
| NVIDIA AI Enterprise |