Key Architecture Highlights
- Specifications Version: The H200 SXM Pre-Specification marked on the official website, which may be adjusted
- FP64 Compute performance: FP64: 34 TFLOPS; FP64 Tensor Core: 67 TFLOPS
- FP32 Compute performance: 67 TFLOPS
Applications
- AI training and inference
- High-performance computing
- Data center
Engineering Parameters
| Specifications Version | The H200 SXM Pre-Specification marked on the official website, which may be adjusted |
|---|---|
| FP64 Compute performance | FP64: 34 TFLOPS FP64 Tensor Core: 67 TFLOPS |
| FP32 Compute performance | 67 TFLOPS |
| Tensor Compute performance | TF32 Tensor Core: 989 TFLOPS (Sparse) BFLOAT16 Tensor Core: 1, 979 TFLOPS (Sparse) FP16 Tensor Core: 1, 979 TFLOPS (Sparse) FP8 Tensor Core: 3, 958 TFLOPS (Sparse) |
| GPU memory ／ Memory | 141GB |
| GPU memory bandwidth | 4.8TB/s |
| Video decoding | 7 NVDEC 7 JPEG |
| Hardware security | Supported |
| Power consumption | Up to 700W (configurable) |
| MIG Division | Up to 7 MIGs @18GB each |
| Form factor and installation | SXM |
| Interconnect Interface | NVIDIA NVLink™: 900GB/s PCIe Gen5: 128GB/s |
Warranty
Public manufacturer information does not specify the warranty term applicable to our supply channel. The supply region, channel and actual warranty service are confirmed in the supply contract.
Manufacturer information
Manufacturer informationVerified Documentation & Datasheets
Manufacturer product information
Download Document