NVIDIA H200 SXM 141GB vs NVIDIA H200 NVL 141GB

Both GPUs launched in 2024: the H200 SXM 141GB on Hopper and the H200 NVL 141GB on Hopper.

Dense throughput for the H200 SXM 141GB against the H200 NVL 141GB: FP64 34 vs 30 TFLOPS, FP32 67 vs 60 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 141 GB against 141 GB, fed at 4.8 TB/s versus 4.8 TB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

Power budgets are 700 W for the H200 SXM 141GB and 600 W for the H200 NVL 141GB. At FP32 that works out to 0.10 against 0.10 TFLOPS per watt.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.