The NVIDIA H200 Tensor Core GPU in its PCIe form factor offers groundbreaking performance for AI workloads, featuring 141GB of memory and a staggering 4.8TB/s bandwidth. This configuration is optimized for large-scale deployments, supporting up to 8 GPUs per server and utilizing NVLink bridges for high-speed data transfer at 900GB/s. With advanced tensor cores delivering nearly 4,000 TFLOPS in FP8 and INT8 operations, the H200 PCIe is designed for demanding data center environments, scalable AI, and multi-tenant workloads through MIG partitioning for maximum efficiency.
| Specification | H200 NVL (PCIe) |
|---|---|
| FP64 | 34 TFLOPS |
| FP64 Tensor Core | 67 TFLOPS |
| FP32 | 67 TFLOPS |
| TF32 Tensor Core² | 989 TFLOPS |
| BFLOAT16 Tensor Core² | 1,979 TFLOPS |
| FP16 Tensor Core² | 1,979 TFLOPS |
| FP8 Tensor Core² | 3,958 TFLOPS |
| INT8 Tensor Core² | 3,958 TFLOPS |
| GPU Memory | 141GB |
| GPU Memory Bandwidth | 4.8TB/s |
| Decoders | 7 NVDEC, 7 JPEG |
| Confidential Computing | Supported |
| Max Thermal Design Power (TDP) | Up to 600W (configurable) |
| Multi-Instance GPUs | Up to 7 MIGs @16.5GB each |
| Form Factor | PCIe |
| Interconnect | 2- or 4-way NVIDIA NVLink bridge: 900GB/s, PCIe Gen5: 128GB/s |
| Server Options | NVIDIA MGX™ H200 NVL partner and NVIDIA-Certified Systems with up to 8 GPUs |
| NVIDIA AI Enterprise | Add-on |
Reviews
There are no reviews yet.