amd-strix-halo-vllm-toolboxes/benchmarks/benchmark_results_rocm/btbtyler09_Qwen3-Coder-30B-A3B-Instruct-gptq-8bit_cluster_tp2_throughput.json

{
    "elapsed_time": 497.59323585999664,
    "num_requests": 200,
    "total_num_tokens": 146805,
    "requests_per_second": 0.4019347241614679,
    "tokens_per_second": 295.0301359026215
}