[
  {
    "50.00 percentile latency (ns)": 33389487468,
    "90.00 percentile latency (ns)": 60156965845,
    "95.00 percentile latency (ns)": 63522258392,
    "97.00 percentile latency (ns)": 64838146823,
    "99.00 percentile latency (ns)": 66131890688,
    "99.90 percentile latency (ns)": 66750697007,
    "Max latency (ns)": 66767479050,
    "Mean latency (ns)": 33413101415,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 159300756,
    "Min queries satisfied": "Yes",
    "Mode": "Performance",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 1680.46,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "40 GB",
    "accelerator_memory_configuration": "HBM2",
    "accelerator_model_name": "A100-PCIE-40GB",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 2,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.mAP": 20.111,
    "characteristics.samples_per_second": 1680.46,
    "characteristics.samples_per_second.normalized_per_core": 840.23,
    "characteristics.samples_per_second.normalized_per_processor": 840.23,
    "ck_system": "QCT_D43KQ_A100-PCIex2_TRT",
    "ck_used": true,
    "cooling": "",
    "dataset": "COCO 2017 (300x300)",
    "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/coco2017.md",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.mAP",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "ssd-mobilenet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package",
    "framework": "TensorRT 7.2, CUDA 11.0, cuDNN 8.0.2, cuBLAS 11.2.0, libjemalloc2, cub 1.8.0, tensorrt-laboratory mlperf branch",
    "host_memory_capacity": "512 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 120,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "AMD EPYC 7V13 64-Core Processor",
    "host_processors_per_node": 2,
    "host_storage_capacity": "NVMe SSD",
    "host_storage_type": "2.0 TB",
    "hw_notes": "",
    "informal_model": "ssd-resnet34",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mAP",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 60000,
    "min_query_count": 1,
    "mlperf_version": 0.7,
    "normalize_cores": 2,
    "normalize_processors": 2,
    "note_code": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/QCT/code",
    "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/QCT/results/QCT_D43KQ_A100-PCIex2_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 18.04.5 LTS (Linux-5.4.0-1055-azure-x86_64-with-Ubuntu-18.04-bionic)",
    "other_software_stack": "docker 19.03.12, python 3.6.8, gcc 5.5.0, onnx 1.3.0, tensorflow 1.13.1, pytorch 1.1.0, torchvision 0.3.0, pycuda 2019.1, sacrebleu 1.3.3, SimpleJSON, OpenCV 4.1.1; GCC 7.5.0; Python 3.7.10",
    "performance_issue_same": true,
    "performance_issue_same_index": 0,
    "performance_issue_unique": true,
    "performance_sample_count": 64,
    "print_timestamps": true,
    "problem": false,
    "qsl_rng_seed": 12786827339337101903,
    "retraining": "N",
    "sample_index_rng_seed": 12640797754436136668,
    "samples_per_query": 112200,
    "schedule_rng_seed": 3135815929913719677,
    "starting_weights_filename": "resnet34-ssd1200.pytorch",
    "status": "available",
    "submitter": "QCT",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/QCT",
    "sw_notes": "Powered by CK v2.5.8 (https://github.com/ctuning/ck)",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/QCT_D43KQ_A100-PCIex2_TRT",
    "system_name": "Microsoft Corporation 7.0 (Virtual Machine)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 1700,
    "task": "object detection",
    "task2": "object detection",
    "total_cores": 240,
    "uid": "fa8ba3c36b0ef968",
    "use_accelerator": true,
    "weight_data_types": "int8,fp16",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 33144903060,
    "90.00 percentile latency (ns)": 59820589280,
    "95.00 percentile latency (ns)": 63159455506,
    "97.00 percentile latency (ns)": 64492184462,
    "99.00 percentile latency (ns)": 65859618329,
    "99.90 percentile latency (ns)": 66429387180,
    "Max latency (ns)": 66503724152,
    "Mean latency (ns)": 33186120217,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 78011207,
    "Min queries satisfied": "Yes",
    "Mode": "Performance",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 8256.98,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "40 GB",
    "accelerator_memory_configuration": "HBM2",
    "accelerator_model_name": "A100-PCIE-40GB",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 10,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.mAP": 20.111,
    "characteristics.samples_per_second": 8256.98,
    "characteristics.samples_per_second.normalized_per_core": 825.698,
    "characteristics.samples_per_second.normalized_per_processor": 825.698,
    "ck_system": "QCT_D52G_A100-PCIex10_TRT",
    "ck_used": true,
    "cooling": "",
    "dataset": "COCO 2017 (300x300)",
    "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/coco2017.md",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.mAP",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "ssd-mobilenet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package",
    "framework": "TensorRT 7.2, CUDA 11.0, cuDNN 8.0.2, cuBLAS 11.2.0, libjemalloc2, cub 1.8.0, tensorrt-laboratory mlperf branch",
    "host_memory_capacity": "768 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 120,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "AMD EPYC 7V13 64-Core Processor",
    "host_processors_per_node": 2,
    "host_storage_capacity": "NVMe SSD",
    "host_storage_type": "1.0 TB",
    "hw_notes": "",
    "informal_model": "ssd-resnet34",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mAP",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 60000,
    "min_query_count": 1,
    "mlperf_version": 0.7,
    "normalize_cores": 10,
    "normalize_processors": 10,
    "note_code": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/QCT/code",
    "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/QCT/results/QCT_D52G_A100-PCIex10_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 18.04.5 LTS (Linux-5.4.0-1055-azure-x86_64-with-Ubuntu-18.04-bionic)",
    "other_software_stack": "docker 19.03.12, python 3.6.8, gcc 5.5.0, onnx 1.3.0, tensorflow 1.13.1, pytorch 1.1.0, torchvision 0.3.0, pycuda 2019.1, sacrebleu 1.3.3, SimpleJSON, OpenCV 4.1.1; GCC 7.5.0; Python 3.7.10",
    "performance_issue_same": true,
    "performance_issue_same_index": 0,
    "performance_issue_unique": true,
    "performance_sample_count": 64,
    "print_timestamps": true,
    "problem": false,
    "qsl_rng_seed": 12786827339337101903,
    "retraining": "N",
    "sample_index_rng_seed": 12640797754436136668,
    "samples_per_query": 549120,
    "schedule_rng_seed": 3135815929913719677,
    "starting_weights_filename": "resnet34-ssd1200.pytorch",
    "status": "available",
    "submitter": "QCT",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/QCT",
    "sw_notes": "Powered by CK v2.5.8 (https://github.com/ctuning/ck)",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/QCT_D52G_A100-PCIex10_TRT",
    "system_name": "Microsoft Corporation 7.0 (Virtual Machine)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 8320,
    "task": "object detection",
    "task2": "object detection",
    "total_cores": 240,
    "uid": "b05480d8b24bc3d2",
    "use_accelerator": true,
    "weight_data_types": "int8,fp16",
    "weight_transformations": "quantization, affine fusion"
  }
]