TensorRT-LLM: Llama 3.3 70B FP8 on 2x H200 SXM 141GB (TP=2), ISL/OSL 1000/2000

Identical workflow to the paired H100 TP=2 point: TensorRT-LLM PyTorch backend, trtllm-bench throughput, 2x NVIDIA H200 SXM 141GB, TP=2, PP=1, ISL/OSL 1000/2000 uniform, 1500 requests, offline infinite-rate feed (concurrency not fixed, achieved in-flight count omitted). Aggregate output tokens/s across both GPUs (about 2895 tok/s per GPU). KV cache free-memory fraction 0.95; CUDA graphs 1-8192.

Measured by NVIDIA, published in the TensorRT-LLM performance overview and reproducible with the documented `prepare_dataset.py` + `trtllm-bench throughput --backend pytorch` workflow (llm_options.yml with CUDA-graph batch sizes 1..8192, kv_cache_free_gpu_mem_fraction 0.95). Metric is 'Total Output Throughput (tokens/sec)' as trtllm-bench reports it: AGGREGATE output tokens/s across all in-flight requests at the stated TP size, under an offline infinite-rate request feed, NOT per-request and NOT per-GPU. Model weights are NVIDIA-published ModelOpt-quantized checkpoints. The page explicitly states these are reference points, not peak performance, and that batch size / parallelism were left mostly at defaults. LOAD-CONDITION AUDIT (bq- 2026-10-04). SCENARIO: MLPerf-style offline scenario. ABSENT CONDITIONS (1 of 7 checked: model, hardware, batch-or-concurrency, input length, output length, engine version, warmup): - warmup: unrecoverable from bench-nv-trtllm-perf-overview - would need the raw harness log's warmup phase or the benchmark command line (loadgen min_duration/min_query_count, or trtllm-bench --warm_up N / --num_runs N) None of these were filled in. A plausible value is worse than an absent one: it would be trusted.