An order-of-magnitude leap for accelerated computing.
The NVIDIA H100 Tensor Core GPU delivers unprecedented acceleration at every scale for AI, data analytics, and high-performance computing (HPC) to tackle the world’s toughest computing challenges. With NVIDIA® NVLink® Switch System, up to 256 H100s can be connected to accelerate exascale workloads, along with a dedicated Transformer Engine to solve trillion-parameter language models. H100’s combined technology innovations can speed up large language models by an incredible 30X over the previous generation to deliver industry-leading conversational AI.
GET YOUR FREE TEST DRIVE
Technical Data
| FP64 | 26 TF |
| FP64 Tensor Core | 51 TF |
| FP32 | 51 TF |
| TF32 Tensor Core | 756 TF |
| BFLOAT16 Tensor Core | 1,513 TF |
| FP16 Tensor Core | 1,513 TF |
| INT8 Tensor Core | 3,026 TOPS |
| GPU memory | 80 GB |
| GPU memory bandwidth | 2 TB/s |
| Max thermal design power (TDP) | 300 bis 350 W (configurable) |
| Multi-Instance-GPUs | Up to 7 MIGs @ 10 GB each |
NVIDIA H100
PCIe
| Form factor | PCIe Dual-slot air-cooled |
| Interconnect | NVLINK: 600GB/s PCIe Gen5: 128 GB/s |
Technical Data
| FP64 | 34 TF |
| FP64 Tensor Core | 67 TF |
| FP32 | 67 TF |
| TF32 Tensor Core | 989 TF |
| BFLOAT16 Tensor Core | 1,979 TF |
| FP16 Tensor Core | 1,979 TF |
| INT8 Tensor Core | 3,958 TOPS |
| GPU memotry | 80 GB |
| GPU memory bandwidth | 3.35 TB/s |
| Max thermal design power (TDP) | Up to 700 W (configurable) |
| Multi Instance GPUs | Up to 7 MIGs @ 10 GB each |
NVIDIA H100
SXM
| Form factor | SXM |
| Interconnect | NVLINK: 900GB/s PCIe Gen5: 128 GB/s |