[
  {
    "50.00 percentile latency (ns)": 355303925633,
    "90.00 percentile latency (ns)": 639515276752,
    "95.00 percentile latency (ns)": 675035348974,
    "97.00 percentile latency (ns)": 689239171839,
    "99.00 percentile latency (ns)": 703457364862,
    "99.90 percentile latency (ns)": 709847481107,
    "Max latency (ns)": 710559582430,
    "Mean latency (ns)": 355306775343,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 88894179,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 306.519,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "24 GB",
    "accelerator_memory_configuration": "HBM2",
    "accelerator_model_name": "NVIDIA A30",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 10,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "boot_firmware_version": "",
    "characteristics.enhancing tumor": 0.7789,
    "characteristics.mean": 0.85407,
    "characteristics.samples_per_second": 306.519,
    "characteristics.samples_per_second.normalized_per_core": 30.6519,
    "characteristics.samples_per_second.normalized_per_processor": 30.6519,
    "characteristics.tumor core": 0.8691,
    "characteristics.whole tumor": 0.9142,
    "ck_system": "A30x10_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "disk_controllers": "",
    "disk_drives": "",
    "division": "closed",
    "filesystem": "",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.1, CUDA 11.3",
    "host_memory_capacity": "2 TB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 38,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Platinum 8368",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "",
    "informal_model": "3d-unet-99",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "management_firmware_version": "",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "network_speed_mbit": "",
    "nics_enabled_connected": "",
    "nics_enabled_firmware": "",
    "nics_enabled_os": "",
    "normalize_cores": 10,
    "normalize_processors": 10,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Nettrix/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Nettrix/results/A30x10_TRT",
    "number_of_nodes": 1,
    "number_of_type_nics_installed": "",
    "operating_system": "Ubuntu 20.04.1",
    "other_hardware": "",
    "other_software_stack": "TensorRT 8.0.1, CUDA 11.3, cuDNN 8.2.1, Driver 470.42.01, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_management": "",
    "power_supply_details": "",
    "power_supply_quantity_and_rating_watts": "",
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 217800,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Nettrix",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Nettrix",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/A30x10_TRT",
    "system_name": "Nettrix X640 G40 (10x A30, TensorRT)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 330,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 76,
    "uid": "490bb4f3bbde5a4d",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 355303925633,
    "90.00 percentile latency (ns)": 639515276752,
    "95.00 percentile latency (ns)": 675035348974,
    "97.00 percentile latency (ns)": 689239171839,
    "99.00 percentile latency (ns)": 703457364862,
    "99.90 percentile latency (ns)": 709847481107,
    "Max latency (ns)": 710559582430,
    "Mean latency (ns)": 355306775343,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 88894179,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 306.519,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "24 GB",
    "accelerator_memory_configuration": "HBM2",
    "accelerator_model_name": "NVIDIA A30",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 10,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "boot_firmware_version": "",
    "characteristics.enhancing tumor": 0.7789,
    "characteristics.mean": 0.85407,
    "characteristics.samples_per_second": 306.519,
    "characteristics.samples_per_second.normalized_per_core": 30.6519,
    "characteristics.samples_per_second.normalized_per_processor": 30.6519,
    "characteristics.tumor core": 0.8691,
    "characteristics.whole tumor": 0.9142,
    "ck_system": "A30x10_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "disk_controllers": "",
    "disk_drives": "",
    "division": "closed",
    "filesystem": "",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.1, CUDA 11.3",
    "host_memory_capacity": "2 TB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 38,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Platinum 8368",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "",
    "informal_model": "3d-unet-99.9",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "management_firmware_version": "",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "network_speed_mbit": "",
    "nics_enabled_connected": "",
    "nics_enabled_firmware": "",
    "nics_enabled_os": "",
    "normalize_cores": 10,
    "normalize_processors": 10,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Nettrix/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Nettrix/results/A30x10_TRT",
    "number_of_nodes": 1,
    "number_of_type_nics_installed": "",
    "operating_system": "Ubuntu 20.04.1",
    "other_hardware": "",
    "other_software_stack": "TensorRT 8.0.1, CUDA 11.3, cuDNN 8.2.1, Driver 470.42.01, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_management": "",
    "power_supply_details": "",
    "power_supply_quantity_and_rating_watts": "",
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 217800,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Nettrix",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Nettrix",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/A30x10_TRT",
    "system_name": "Nettrix X640 G40 (10x A30, TensorRT)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 330,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 76,
    "uid": "28301ec461a822e3",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 322865654987,
    "90.00 percentile latency (ns)": 581839160694,
    "95.00 percentile latency (ns)": 614194446874,
    "97.00 percentile latency (ns)": 627142923432,
    "99.00 percentile latency (ns)": 640091828680,
    "99.90 percentile latency (ns)": 645912448765,
    "Max latency (ns)": 646560876083,
    "Mean latency (ns)": 322920651217,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 96787263,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 178.637,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "16 GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA A10",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 8,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "boot_firmware_version": "",
    "characteristics.enhancing tumor": 0.7789,
    "characteristics.mean": 0.854,
    "characteristics.samples_per_second": 178.637,
    "characteristics.samples_per_second.normalized_per_core": 22.329625,
    "characteristics.samples_per_second.normalized_per_processor": 22.329625,
    "characteristics.tumor core": 0.869,
    "characteristics.whole tumor": 0.9141,
    "ck_system": "A10x8_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "disk_controllers": "",
    "disk_drives": "",
    "division": "closed",
    "filesystem": "",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.1, CUDA 11.3",
    "host_memory_capacity": "2048 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 40,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Platinum 8380 CPU @ 2.30GHz",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "",
    "informal_model": "3d-unet-99",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "management_firmware_version": "",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "network_speed_mbit": "",
    "nics_enabled_connected": "",
    "nics_enabled_firmware": "",
    "nics_enabled_os": "",
    "normalize_cores": 8,
    "normalize_processors": 8,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Nettrix/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Nettrix/results/A10x8_TRT",
    "number_of_nodes": 1,
    "number_of_type_nics_installed": "",
    "operating_system": "Ubuntu 20.04.1",
    "other_hardware": "",
    "other_software_stack": "TensorRT 8.0.1, CUDA 11.3, cuDNN 8.2.1, Driver 470.42.01, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_management": "",
    "power_supply_details": "",
    "power_supply_quantity_and_rating_watts": "",
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 115500,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Nettrix",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Nettrix",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/A10x8_TRT",
    "system_name": "Nettrix X620 G40 (8x A10, TensorRT)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 175,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 80,
    "uid": "1b9e600df12da8ec",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 322865654987,
    "90.00 percentile latency (ns)": 581839160694,
    "95.00 percentile latency (ns)": 614194446874,
    "97.00 percentile latency (ns)": 627142923432,
    "99.00 percentile latency (ns)": 640091828680,
    "99.90 percentile latency (ns)": 645912448765,
    "Max latency (ns)": 646560876083,
    "Mean latency (ns)": 322920651217,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 96787263,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 178.637,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "16 GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA A10",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 8,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "boot_firmware_version": "",
    "characteristics.enhancing tumor": 0.7789,
    "characteristics.mean": 0.854,
    "characteristics.samples_per_second": 178.637,
    "characteristics.samples_per_second.normalized_per_core": 22.329625,
    "characteristics.samples_per_second.normalized_per_processor": 22.329625,
    "characteristics.tumor core": 0.869,
    "characteristics.whole tumor": 0.9141,
    "ck_system": "A10x8_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "disk_controllers": "",
    "disk_drives": "",
    "division": "closed",
    "filesystem": "",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.1, CUDA 11.3",
    "host_memory_capacity": "2048 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 40,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Platinum 8380 CPU @ 2.30GHz",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "",
    "informal_model": "3d-unet-99.9",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "management_firmware_version": "",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "network_speed_mbit": "",
    "nics_enabled_connected": "",
    "nics_enabled_firmware": "",
    "nics_enabled_os": "",
    "normalize_cores": 8,
    "normalize_processors": 8,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Nettrix/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Nettrix/results/A10x8_TRT",
    "number_of_nodes": 1,
    "number_of_type_nics_installed": "",
    "operating_system": "Ubuntu 20.04.1",
    "other_hardware": "",
    "other_software_stack": "TensorRT 8.0.1, CUDA 11.3, cuDNN 8.2.1, Driver 470.42.01, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_management": "",
    "power_supply_details": "",
    "power_supply_quantity_and_rating_watts": "",
    "print_timestamps": 0,
    "problem": false,
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 115500,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Nettrix",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Nettrix",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/A10x8_TRT",
    "system_name": "Nettrix X620 G40 (8x A10, TensorRT)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 175,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 80,
    "uid": "43d90308685e4c34",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 80833891651,
    "90.00 percentile latency (ns)": 145704200620,
    "95.00 percentile latency (ns)": 153827861406,
    "97.00 percentile latency (ns)": 157075266685,
    "99.00 percentile latency (ns)": 160322857690,
    "99.90 percentile latency (ns)": 161782537120,
    "Max latency (ns)": 161941564555,
    "Mean latency (ns)": 80878448931,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 50908804,
    "Min queries satisfied": "Yes",
    "Mode": "Performance",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 151.758,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "40GB",
    "accelerator_memory_configuration": "HBM2",
    "accelerator_model_name": "NVIDIA A100-PCIe",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 4,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.enhancing tumor": 0.7786,
    "characteristics.mean": 0.85337,
    "characteristics.samples_per_second": 151.758,
    "characteristics.samples_per_second.normalized_per_core": 37.9395,
    "characteristics.samples_per_second.normalized_per_processor": 37.9395,
    "characteristics.tumor core": 0.8692,
    "characteristics.whole tumor": 0.9123,
    "ck_system": "A100-PCIex4_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 7.2, CUDA 11.0 Update 1",
    "host_memory_capacity": "192 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 56,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Platinum 8280M CPU @ 2.70GHz",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "",
    "informal_model": "3d-unet-99",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "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": 4,
    "normalize_processors": 4,
    "note_code": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Nettrix/code",
    "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Nettrix/results/A100-PCIex4_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 18.04.3",
    "other_software_stack": "TensorRT 7.2, CUDA 11.0 Update 1, cuDNN 8.0.2, DALI 0.25.0",
    "performance_issue_same": true,
    "performance_issue_same_index": 0,
    "performance_issue_unique": true,
    "performance_sample_count": 16,
    "print_timestamps": true,
    "problem": false,
    "qsl_rng_seed": 12786827339337101903,
    "retraining": "N",
    "sample_index_rng_seed": 12640797754436136668,
    "samples_per_query": 24576,
    "schedule_rng_seed": 3135815929913719677,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Nettrix",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Nettrix",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/A100-PCIex4_TRT",
    "system_name": "Nettrix X640 G30 (4x A100-PCIe)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 170,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 112,
    "uid": "a816d92fd44f189d",
    "use_accelerator": true,
    "weight_data_types": "int8,fp16",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 80833891651,
    "90.00 percentile latency (ns)": 145704200620,
    "95.00 percentile latency (ns)": 153827861406,
    "97.00 percentile latency (ns)": 157075266685,
    "99.00 percentile latency (ns)": 160322857690,
    "99.90 percentile latency (ns)": 161782537120,
    "Max latency (ns)": 161941564555,
    "Mean latency (ns)": 80878448931,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 50908804,
    "Min queries satisfied": "Yes",
    "Mode": "Performance",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 151.758,
    "Scenario": "offline",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "40GB",
    "accelerator_memory_configuration": "HBM2",
    "accelerator_model_name": "NVIDIA A100-PCIe",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 4,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "characteristics.enhancing tumor": 0.7786,
    "characteristics.mean": 0.85337,
    "characteristics.samples_per_second": 151.758,
    "characteristics.samples_per_second.normalized_per_core": 37.9395,
    "characteristics.samples_per_second.normalized_per_processor": 37.9395,
    "characteristics.tumor core": 0.8692,
    "characteristics.whole tumor": 0.9123,
    "ck_system": "A100-PCIex4_TRT",
    "ck_used": false,
    "cooling": "",
    "dataset": "BraTS 2019",
    "dataset_link": "https://www.med.upenn.edu/cbica/brats2019/data.html",
    "dim_x_default": "characteristics.samples_per_second",
    "dim_x_maximize": true,
    "dim_y_default": "characteristics.mean",
    "dim_y_maximize": true,
    "division": "closed",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 7.2, CUDA 11.0 Update 1",
    "host_memory_capacity": "192 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 56,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Platinum 8280M CPU @ 2.70GHz",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "",
    "informal_model": "3d-unet-99.9",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "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": 4,
    "normalize_processors": 4,
    "note_code": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Nettrix/code",
    "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Nettrix/results/A100-PCIex4_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 18.04.3",
    "other_software_stack": "TensorRT 7.2, CUDA 11.0 Update 1, cuDNN 8.0.2, DALI 0.25.0",
    "performance_issue_same": true,
    "performance_issue_same_index": 0,
    "performance_issue_unique": true,
    "performance_sample_count": 16,
    "print_timestamps": true,
    "problem": false,
    "qsl_rng_seed": 12786827339337101903,
    "retraining": "N",
    "sample_index_rng_seed": 12640797754436136668,
    "samples_per_query": 24576,
    "schedule_rng_seed": 3135815929913719677,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Nettrix",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Nettrix",
    "sw_notes": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/A100-PCIex4_TRT",
    "system_name": "Nettrix X640 G30 (4x A100-PCIe)",
    "system_type": "datacenter",
    "target_latency (ns)": 0,
    "target_qps": 170,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 112,
    "uid": "f310406eb43e1957",
    "use_accelerator": true,
    "weight_data_types": "int8,fp16",
    "weight_transformations": "quantization, affine fusion"
  }
]