[
  {
    "50.00 percentile latency (ns)": 521903295412,
    "90.00 percentile latency (ns)": 939461221987,
    "95.00 percentile latency (ns)": 991622041089,
    "97.00 percentile latency (ns)": 1012524225979,
    "99.00 percentile latency (ns)": 1033424404541,
    "99.90 percentile latency (ns)": 1042771335258,
    "Max latency (ns)": 1043789499283,
    "Mean latency (ns)": 521867541246,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 127633099,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 23.545,
    "Scenario": "offline",
    "accelerator_cooling_type": "passive",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "24GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA A10",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "boot_firmware_version": "2.12.3",
    "characteristics.enhancing tumor": 0.7789,
    "characteristics.mean": 0.854,
    "characteristics.samples_per_second": 23.545,
    "characteristics.samples_per_second.normalized_per_core": 23.545,
    "characteristics.samples_per_second.normalized_per_processor": 23.545,
    "characteristics.tumor core": 0.869,
    "characteristics.whole tumor": 0.9141,
    "ck_system": "XE2420_A10x1_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",
    "filesystem": "ext3/ext4",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.2, CUDA 11.3",
    "host_cooling_type": "air",
    "host_memory_capacity": "512 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 24,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Gold 6252N CPU @ 2.30GHz",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "ECC on",
    "informal_model": "3d-unet-99",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "management_firmware_version": "5.00.00.00",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/results/XE2420_A10x1_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2",
    "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_settings": "closed/Dell/power/XE2420_A10x1_power_settings.md",
    "print_timestamps": 0,
    "problem": false,
    "psu_details": "2x2000W",
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 24576,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Dell",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Dell",
    "sw_notes": "",
    "system_cooling_type": "air",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/XE2420_A10x1_TRT",
    "system_name": "Dell EMC PowerEdge XE2420 (1x A10, TensorRT)",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 48,
    "uid": "f4e98b056052779f",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 521903295412,
    "90.00 percentile latency (ns)": 939461221987,
    "95.00 percentile latency (ns)": 991622041089,
    "97.00 percentile latency (ns)": 1012524225979,
    "99.00 percentile latency (ns)": 1033424404541,
    "99.90 percentile latency (ns)": 1042771335258,
    "Max latency (ns)": 1043789499283,
    "Mean latency (ns)": 521867541246,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 127633099,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 23.545,
    "Scenario": "offline",
    "accelerator_cooling_type": "passive",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "24GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA A10",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "boot_firmware_version": "2.12.3",
    "characteristics.enhancing tumor": 0.7789,
    "characteristics.mean": 0.854,
    "characteristics.samples_per_second": 23.545,
    "characteristics.samples_per_second.normalized_per_core": 23.545,
    "characteristics.samples_per_second.normalized_per_processor": 23.545,
    "characteristics.tumor core": 0.869,
    "characteristics.whole tumor": 0.9141,
    "ck_system": "XE2420_A10x1_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",
    "filesystem": "ext3/ext4",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.2, CUDA 11.3",
    "host_cooling_type": "air",
    "host_memory_capacity": "512 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 24,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Gold 6252N CPU @ 2.30GHz",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "ECC on",
    "informal_model": "3d-unet-99.9",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "management_firmware_version": "5.00.00.00",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/results/XE2420_A10x1_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2",
    "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_settings": "closed/Dell/power/XE2420_A10x1_power_settings.md",
    "print_timestamps": 0,
    "problem": false,
    "psu_details": "2x2000W",
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 24576,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Dell",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Dell",
    "sw_notes": "",
    "system_cooling_type": "air",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/XE2420_A10x1_TRT",
    "system_name": "Dell EMC PowerEdge XE2420 (1x A10, TensorRT)",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 48,
    "uid": "99c622510404d95a",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 531422695378,
    "90.00 percentile latency (ns)": 956305098833,
    "95.00 percentile latency (ns)": 1009391081728,
    "97.00 percentile latency (ns)": 1030660328257,
    "99.00 percentile latency (ns)": 1051932002991,
    "99.90 percentile latency (ns)": 1061442601613,
    "Max latency (ns)": 1062480395066,
    "Mean latency (ns)": 531294677299,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 129202752,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 23.1308,
    "Scenario": "offline",
    "accelerator_cooling_type": "passive",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "24GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA A10",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "boot_firmware_version": "2.12.3",
    "characteristics.enhancing tumor": 0.7789,
    "characteristics.mean": 0.854,
    "characteristics.power": 399.7663653483991,
    "characteristics.power.normalized_per_core": 399.7663653483991,
    "characteristics.power.normalized_per_processor": 399.7663653483991,
    "characteristics.samples_per_second": 23.1308,
    "characteristics.samples_per_second.normalized_per_core": 23.1308,
    "characteristics.samples_per_second.normalized_per_processor": 23.1308,
    "characteristics.tumor core": 0.869,
    "characteristics.whole tumor": 0.9141,
    "ck_system": "XE2420_A10x1_TRT_MaxQ",
    "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",
    "filesystem": "ext3/ext4",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.2, CUDA 11.3",
    "host_cooling_type": "air",
    "host_memory_capacity": "512 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 24,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Gold 6252N CPU @ 2.30GHz",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "ECC on",
    "informal_model": "3d-unet-99",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "management_firmware_version": "5.00.00.00",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/results/XE2420_A10x1_TRT_MaxQ",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2",
    "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_settings": "closed/Dell/power/XE2420_A10x1_power_settings.md",
    "print_timestamps": 0,
    "problem": false,
    "psu_details": "2x2000W",
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 24576,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Dell",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Dell",
    "sw_notes": "",
    "system_cooling_type": "air",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/XE2420_A10x1_TRT_MaxQ",
    "system_name": "Dell EMC PowerEdge XE2420 (1x A10, MaxQ, TensorRT)",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 48,
    "uid": "bf4d399d38464e2e",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 531422695378,
    "90.00 percentile latency (ns)": 956305098833,
    "95.00 percentile latency (ns)": 1009391081728,
    "97.00 percentile latency (ns)": 1030660328257,
    "99.00 percentile latency (ns)": 1051932002991,
    "99.90 percentile latency (ns)": 1061442601613,
    "Max latency (ns)": 1062480395066,
    "Mean latency (ns)": 531294677299,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 129202752,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 23.1308,
    "Scenario": "offline",
    "accelerator_cooling_type": "passive",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "24GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA A10",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 1,
    "accuracy_log_probability": 0,
    "accuracy_log_rng_seed": 0,
    "accuracy_log_sampling_target": 0,
    "boot_firmware_version": "2.12.3",
    "characteristics.enhancing tumor": 0.7789,
    "characteristics.mean": 0.854,
    "characteristics.power": 399.7663653483991,
    "characteristics.power.normalized_per_core": 399.7663653483991,
    "characteristics.power.normalized_per_processor": 399.7663653483991,
    "characteristics.samples_per_second": 23.1308,
    "characteristics.samples_per_second.normalized_per_core": 23.1308,
    "characteristics.samples_per_second.normalized_per_processor": 23.1308,
    "characteristics.tumor core": 0.869,
    "characteristics.whole tumor": 0.9141,
    "ck_system": "XE2420_A10x1_TRT_MaxQ",
    "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",
    "filesystem": "ext3/ext4",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.2, CUDA 11.3",
    "host_cooling_type": "air",
    "host_memory_capacity": "512 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 24,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Gold 6252N CPU @ 2.30GHz",
    "host_processors_per_node": 2,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "ECC on",
    "informal_model": "3d-unet-99.9",
    "input_data_types": "int8",
    "key.accuracy": "characteristics.mean",
    "management_firmware_version": "5.00.00.00",
    "max_async_queries": 1,
    "max_duration (ms)": 0,
    "max_query_count": 0,
    "min_duration (ms)": 600000,
    "min_query_count": 1,
    "mlperf_version": 1.1,
    "normalize_cores": 1,
    "normalize_processors": 1,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/results/XE2420_A10x1_TRT_MaxQ",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2",
    "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_settings": "closed/Dell/power/XE2420_A10x1_power_settings.md",
    "print_timestamps": 0,
    "problem": false,
    "psu_details": "2x2000W",
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 24576,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Dell",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Dell",
    "sw_notes": "",
    "system_cooling_type": "air",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/XE2420_A10x1_TRT_MaxQ",
    "system_name": "Dell EMC PowerEdge XE2420 (1x A10, MaxQ, TensorRT)",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 25,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 48,
    "uid": "d4f60b2faf244acf",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 375524168356,
    "90.00 percentile latency (ns)": 676893296917,
    "95.00 percentile latency (ns)": 714532217065,
    "97.00 percentile latency (ns)": 729600847079,
    "99.00 percentile latency (ns)": 744673452051,
    "99.90 percentile latency (ns)": 751433281638,
    "Max latency (ns)": 752164220800,
    "Mean latency (ns)": 375590889645,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 107402652,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 43.8734,
    "Scenario": "offline",
    "accelerator_cooling_type": "",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "24GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA A10",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 2,
    "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": 43.8734,
    "characteristics.samples_per_second.normalized_per_core": 21.9367,
    "characteristics.samples_per_second.normalized_per_processor": 21.9367,
    "characteristics.tumor core": 0.869,
    "characteristics.whole tumor": 0.9141,
    "ck_system": "XR12_edge_A10x2_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",
    "filesystem": "",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.0,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.2, CUDA 11.3",
    "host_cooling_type": "",
    "host_memory_capacity": "512 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 28,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Gold 6330 CPU @ 2.00GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "ECC on",
    "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,
    "normalize_cores": 2,
    "normalize_processors": 2,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/results/XR12_edge_A10x2_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2",
    "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_settings": "",
    "print_timestamps": 0,
    "problem": false,
    "psu_details": "",
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 33000,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Dell",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Dell",
    "sw_notes": "",
    "system_cooling_type": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/XR12_edge_A10x2_TRT",
    "system_name": "Dell EMC PowerEdge XR12 (2x A10, TensorRT)",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 50,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 28,
    "uid": "6b62548990a4e3d8",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  },
  {
    "50.00 percentile latency (ns)": 375524168356,
    "90.00 percentile latency (ns)": 676893296917,
    "95.00 percentile latency (ns)": 714532217065,
    "97.00 percentile latency (ns)": 729600847079,
    "99.00 percentile latency (ns)": 744673452051,
    "99.90 percentile latency (ns)": 751433281638,
    "Max latency (ns)": 752164220800,
    "Mean latency (ns)": 375590889645,
    "Min duration satisfied": "Yes",
    "Min latency (ns)": 107402652,
    "Min queries satisfied": "Yes",
    "Mode": "PerformanceOnly",
    "Result is": "VALID",
    "SUT name": "LWIS_Server",
    "Samples per second": 43.8734,
    "Scenario": "offline",
    "accelerator_cooling_type": "",
    "accelerator_frequency": "",
    "accelerator_host_interconnect": "",
    "accelerator_interconnect": "",
    "accelerator_interconnect_topology": "",
    "accelerator_memory_capacity": "24GB",
    "accelerator_memory_configuration": "GDDR6",
    "accelerator_model_name": "NVIDIA A10",
    "accelerator_on-chip_memories": "",
    "accelerators_per_node": 2,
    "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": 43.8734,
    "characteristics.samples_per_second.normalized_per_core": 21.9367,
    "characteristics.samples_per_second.normalized_per_processor": 21.9367,
    "characteristics.tumor core": 0.869,
    "characteristics.whole tumor": 0.9141,
    "ck_system": "XR12_edge_A10x2_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",
    "filesystem": "",
    "formal_model": "3d-unet",
    "formal_model_accuracy": 99.9,
    "formal_model_link": "",
    "framework": "TensorRT 8.0.2, CUDA 11.3",
    "host_cooling_type": "",
    "host_memory_capacity": "512 GB",
    "host_memory_configuration": "",
    "host_networking": "",
    "host_networking_topology": "",
    "host_processor_caches": "",
    "host_processor_core_count": 28,
    "host_processor_frequency": "",
    "host_processor_interconnect": "",
    "host_processor_model_name": "Intel(R) Xeon(R) Gold 6330 CPU @ 2.00GHz",
    "host_processors_per_node": 1,
    "host_storage_capacity": "4 TB",
    "host_storage_type": "NVMe SSD",
    "hw_notes": "ECC on",
    "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,
    "normalize_cores": 2,
    "normalize_processors": 2,
    "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/code",
    "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Dell/results/XR12_edge_A10x2_TRT",
    "number_of_nodes": 1,
    "operating_system": "Ubuntu 20.04.2",
    "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02, DALI 0.31.0",
    "performance_issue_same": 0,
    "performance_issue_same_index": 0,
    "performance_issue_unique": 0,
    "performance_sample_count": 16,
    "power_settings": "",
    "print_timestamps": 0,
    "problem": false,
    "psu_details": "",
    "qsl_rng_seed": 1624344308455410291,
    "retraining": "N",
    "sample_index_rng_seed": 517984244576520566,
    "samples_per_query": 33000,
    "schedule_rng_seed": 10051496985653635065,
    "starting_weights_filename": "224_224_160_dyanmic_bs.onnx",
    "status": "available",
    "submitter": "Dell",
    "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Dell",
    "sw_notes": "",
    "system_cooling_type": "",
    "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/XR12_edge_A10x2_TRT",
    "system_name": "Dell EMC PowerEdge XR12 (2x A10, TensorRT)",
    "system_type": "edge",
    "target_latency (ns)": 0,
    "target_qps": 50,
    "task": "image segmentation",
    "task2": "image segmentation",
    "total_cores": 28,
    "uid": "42ae13b89ca4c4f8",
    "use_accelerator": true,
    "weight_data_types": "int8",
    "weight_transformations": "quantization, affine fusion"
  }
]