amd-strix-halo-vllm-toolboxes/benchmarks/benchmark_results_rocm/openai_gpt-oss-20b_tp1_throughput.json

{
    "elapsed_time": 362.9645123449736,
    "num_requests": 200,
    "total_num_tokens": 147036,
    "requests_per_second": 0.5510180560294371,
    "tokens_per_second": 405.0974544317216
}