Infrastructure atlas
NVIDIA/Hopper
H200 Tensor Core
Hopper’s memory-heavy sibling, built around 141 GB of HBM3e for larger models and KV caches.
GPUtraininginferencelong contextCUDA

Precision & peak compute
| Precision | Published peak | Basis |
|---|---|---|
| FP8 Tensor | 3,958 TFLOPS | SXM, with sparsity |
| FP16 / BF16 Tensor | 1,979 TFLOPS | SXM, with sparsity |
| INT8 Tensor | 3,958 TOPS | SXM, with sparsity |
Peak FLOPS and TOPS are theoretical hardware figures. Sparse and dense modes, chip and system totals, and form factors are not interchangeable.
NVIDIA H200 specifications