[ { "50.00 percentile latency (ns)": 328060762265, "90.00 percentile latency (ns)": 590470362975, "95.00 percentile latency (ns)": 623249789926, "97.00 percentile latency (ns)": 636375612783, "99.00 percentile latency (ns)": 649490331393, "99.90 percentile latency (ns)": 655398580902, "Max latency (ns)": 656047662521, "Mean latency (ns)": 328065350821, "Min duration satisfied": "Yes", "Min latency (ns)": 59249891, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 246.476, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "40 GB", "accelerator_memory_configuration": "HBM2", "accelerator_model_name": "NVIDIA A100-SXM-80GB-500W", "accelerator_on-chip_memories": "", "accelerators_per_node": 4, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.7803, "characteristics.mean": 0.85443, "characteristics.samples_per_second": 246.476, "characteristics.samples_per_second.normalized_per_core": 61.619, "characteristics.samples_per_second.normalized_per_processor": 61.619, "characteristics.tumor core": 0.8697, "characteristics.whole tumor": 0.9133, "ck_system": "Lenovo-SD650v2-N-A100_A100-SXM-80GBx4_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 8.0.2, CUDA 11.3", "host_memory_capacity": "768 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) Platinum 8360Y CPU @ 2.60GHz", "host_processors_per_node": 2, "host_storage_capacity": "2 TB", "host_storage_type": "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": 4, "normalize_processors": 4, "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/results/Lenovo-SD650v2-N-A100_A100-SXM-80GBx4_TRT", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Centos 8.4", "other_hardware": "", "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02", "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": 161700, "schedule_rng_seed": 10051496985653635065, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "sw_notes": "", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/Lenovo-SD650v2-N-A100_A100-SXM-80GBx4_TRT", "system_name": "Lenovo ThinkSystem SD650N-v2 Server", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 245, "task": "image segmentation", "task2": "image segmentation", "total_cores": 56, "uid": "e81d684379b00df5", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 328060762265, "90.00 percentile latency (ns)": 590470362975, "95.00 percentile latency (ns)": 623249789926, "97.00 percentile latency (ns)": 636375612783, "99.00 percentile latency (ns)": 649490331393, "99.90 percentile latency (ns)": 655398580902, "Max latency (ns)": 656047662521, "Mean latency (ns)": 328065350821, "Min duration satisfied": "Yes", "Min latency (ns)": 59249891, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 246.476, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "40 GB", "accelerator_memory_configuration": "HBM2", "accelerator_model_name": "NVIDIA A100-SXM-80GB-500W", "accelerator_on-chip_memories": "", "accelerators_per_node": 4, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.7803, "characteristics.mean": 0.85443, "characteristics.samples_per_second": 246.476, "characteristics.samples_per_second.normalized_per_core": 61.619, "characteristics.samples_per_second.normalized_per_processor": 61.619, "characteristics.tumor core": 0.8697, "characteristics.whole tumor": 0.9133, "ck_system": "Lenovo-SD650v2-N-A100_A100-SXM-80GBx4_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 8.0.2, CUDA 11.3", "host_memory_capacity": "768 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) Platinum 8360Y CPU @ 2.60GHz", "host_processors_per_node": 2, "host_storage_capacity": "2 TB", "host_storage_type": "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": 4, "normalize_processors": 4, "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/results/Lenovo-SD650v2-N-A100_A100-SXM-80GBx4_TRT", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Centos 8.4", "other_hardware": "", "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02", "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": 161700, "schedule_rng_seed": 10051496985653635065, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "sw_notes": "", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/Lenovo-SD650v2-N-A100_A100-SXM-80GBx4_TRT", "system_name": "Lenovo ThinkSystem SD650N-v2 Server", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 245, "task": "image segmentation", "task2": "image segmentation", "total_cores": 56, "uid": "38dbc360ed5df84b", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 315155231623, "90.00 percentile latency (ns)": 567847093947, "95.00 percentile latency (ns)": 599401199468, "97.00 percentile latency (ns)": 612035542134, "99.00 percentile latency (ns)": 624677753107, "99.90 percentile latency (ns)": 630376683138, "Max latency (ns)": 630992127947, "Mean latency (ns)": 315219106609, "Min duration satisfied": "Yes", "Min latency (ns)": 142303911, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 120.287, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "48 GB", "accelerator_memory_configuration": "GDDR6", "accelerator_model_name": "NVIDIA A30", "accelerator_on-chip_memories": "", "accelerators_per_node": 4, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.7803, "characteristics.mean": 0.85443, "characteristics.samples_per_second": 120.287, "characteristics.samples_per_second.normalized_per_core": 30.07175, "characteristics.samples_per_second.normalized_per_processor": 30.07175, "characteristics.tumor core": 0.8697, "characteristics.whole tumor": 0.9133, "ck_system": "A30x4_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 8.0.2, CUDA 11.3", "host_memory_capacity": "704 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) Platinum 8280 CPU @ 2.60GHz", "host_processors_per_node": 2, "host_storage_capacity": "2 TB", "host_storage_type": "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": 4, "normalize_processors": 4, "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/results/A30x4_TRT", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Ubuntu 20.04.4", "other_hardware": "", "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02", "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": 75900, "schedule_rng_seed": 10051496985653635065, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "sw_notes": "", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/A30x4_TRT", "system_name": "Lenovo ThinkSystem SR670 Server", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 115, "task": "image segmentation", "task2": "image segmentation", "total_cores": 56, "uid": "104ef7c49f2b6f01", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 315155231623, "90.00 percentile latency (ns)": 567847093947, "95.00 percentile latency (ns)": 599401199468, "97.00 percentile latency (ns)": 612035542134, "99.00 percentile latency (ns)": 624677753107, "99.90 percentile latency (ns)": 630376683138, "Max latency (ns)": 630992127947, "Mean latency (ns)": 315219106609, "Min duration satisfied": "Yes", "Min latency (ns)": 142303911, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 120.287, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "48 GB", "accelerator_memory_configuration": "GDDR6", "accelerator_model_name": "NVIDIA A30", "accelerator_on-chip_memories": "", "accelerators_per_node": 4, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.7803, "characteristics.mean": 0.85443, "characteristics.samples_per_second": 120.287, "characteristics.samples_per_second.normalized_per_core": 30.07175, "characteristics.samples_per_second.normalized_per_processor": 30.07175, "characteristics.tumor core": 0.8697, "characteristics.whole tumor": 0.9133, "ck_system": "A30x4_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 8.0.2, CUDA 11.3", "host_memory_capacity": "704 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) Platinum 8280 CPU @ 2.60GHz", "host_processors_per_node": 2, "host_storage_capacity": "2 TB", "host_storage_type": "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": 4, "normalize_processors": 4, "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/results/A30x4_TRT", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Ubuntu 20.04.4", "other_hardware": "", "other_software_stack": "TensorRT 8.0.2, CUDA 11.3, cuDNN 8.2.1, Driver 470.57.02", "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": 75900, "schedule_rng_seed": 10051496985653635065, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "sw_notes": "", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/A30x4_TRT", "system_name": "Lenovo ThinkSystem SR670 Server", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 115, "task": "image segmentation", "task2": "image segmentation", "total_cores": 56, "uid": "9a599aceb8004bf0", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 1806956768336, "90.00 percentile latency (ns)": 3262788632279, "95.00 percentile latency (ns)": 3443183252055, "97.00 percentile latency (ns)": 3514828338529, "99.00 percentile latency (ns)": 3586573837780, "99.90 percentile latency (ns)": 3618725751645, "Max latency (ns)": 3622139931143, "Mean latency (ns)": 1809066677261, "Min duration satisfied": "Yes", "Min latency (ns)": 1314801171, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "Triton_Server", "Samples per second": 6.78494, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "N/A", "accelerator_memory_configuration": "", "accelerator_model_name": "N/A", "accelerator_on-chip_memories": "", "accelerators_per_node": 0, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.7876, "characteristics.mean": 0.8562, "characteristics.samples_per_second": 6.78494, "characteristics.samples_per_second.normalized_per_core": 0.08481174999999999, "characteristics.samples_per_second.normalized_per_processor": 3.39247, "characteristics.tumor core": 0.8679, "characteristics.whole tumor": 0.9131, "ck_system": "Triton_CPU_2S_8380x1", "ck_used": false, "cooling": "Liquid - Lenovo Neptune liquid 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": "OpenVino 2021", "host_memory_capacity": "", "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 @ 2.20GHz", "host_processors_per_node": 2, "host_storage_capacity": "1 TB", "host_storage_type": "", "hw_notes": "", "informal_model": "3d-unet-99", "input_data_types": "fp32", "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": 80, "normalize_processors": 2, "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/results/Triton_CPU_2S_8380x1", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Ubuntu 20.04.2 LTS", "other_hardware": "", "other_software_stack": "OpenVino 2021, Triton 21.07", "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": 24576, "schedule_rng_seed": 10051496985653635065, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "sw_notes": "CPU Inference on Triton Inference Server", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/Triton_CPU_2S_8380x1", "system_name": "Lenovo ThinkSystem SD650-v2 (Ice Lake running Triton)", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 6.5, "task": "image segmentation", "task2": "image segmentation", "total_cores": 80, "uid": "820e12c3ccc472f9", "use_accelerator": false, "weight_data_types": "fp32", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 1806956768336, "90.00 percentile latency (ns)": 3262788632279, "95.00 percentile latency (ns)": 3443183252055, "97.00 percentile latency (ns)": 3514828338529, "99.00 percentile latency (ns)": 3586573837780, "99.90 percentile latency (ns)": 3618725751645, "Max latency (ns)": 3622139931143, "Mean latency (ns)": 1809066677261, "Min duration satisfied": "Yes", "Min latency (ns)": 1314801171, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "Triton_Server", "Samples per second": 6.78494, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "N/A", "accelerator_memory_configuration": "", "accelerator_model_name": "N/A", "accelerator_on-chip_memories": "", "accelerators_per_node": 0, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.7876, "characteristics.mean": 0.8562, "characteristics.samples_per_second": 6.78494, "characteristics.samples_per_second.normalized_per_core": 0.08481174999999999, "characteristics.samples_per_second.normalized_per_processor": 3.39247, "characteristics.tumor core": 0.8679, "characteristics.whole tumor": 0.9131, "ck_system": "Triton_CPU_2S_8380x1", "ck_used": false, "cooling": "Liquid - Lenovo Neptune liquid 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": "OpenVino 2021", "host_memory_capacity": "", "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 @ 2.20GHz", "host_processors_per_node": 2, "host_storage_capacity": "1 TB", "host_storage_type": "", "hw_notes": "", "informal_model": "3d-unet-99.9", "input_data_types": "fp32", "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": 80, "normalize_processors": 2, "note_code": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.1/tree/master/closed/Lenovo/results/Triton_CPU_2S_8380x1", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Ubuntu 20.04.2 LTS", "other_hardware": "", "other_software_stack": "OpenVino 2021, Triton 21.07", "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": 24576, "schedule_rng_seed": 10051496985653635065, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "sw_notes": "CPU Inference on Triton Inference Server", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/Triton_CPU_2S_8380x1", "system_name": "Lenovo ThinkSystem SD650-v2 (Ice Lake running Triton)", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 6.5, "task": "image segmentation", "task2": "image segmentation", "total_cores": 80, "uid": "cb74bc6962eff32b", "use_accelerator": false, "weight_data_types": "fp32", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 340686764646, "90.00 percentile latency (ns)": 614428604242, "95.00 percentile latency (ns)": 648648446655, "97.00 percentile latency (ns)": 662321395888, "99.00 percentile latency (ns)": 676021917156, "99.90 percentile latency (ns)": 682169000933, "Max latency (ns)": 682868663143, "Mean latency (ns)": 340865376074, "Min duration satisfied": "Yes", "Min latency (ns)": 115000717, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 123.713, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "48 GB", "accelerator_memory_configuration": "GDDR6", "accelerator_model_name": "NVIDIA A40", "accelerator_on-chip_memories": "", "accelerators_per_node": 4, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.7799, "characteristics.mean": 0.85383, "characteristics.samples_per_second": 123.713, "characteristics.samples_per_second.normalized_per_core": 30.92825, "characteristics.samples_per_second.normalized_per_processor": 30.92825, "characteristics.tumor core": 0.8684, "characteristics.whole tumor": 0.9132, "ck_system": "A40x4_TRT", "ck_used": true, "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 7.2.3, CUDA 11.1", "host_memory_capacity": "704 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": "2 TB", "host_storage_type": "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": 1.0, "network_speed_mbit": "", "nics_enabled_connected": "", "nics_enabled_firmware": "", "nics_enabled_os": "", "normalize_cores": 4, "normalize_processors": 4, "note_code": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/results/A40x4_TRT", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Ubuntu 18.04.5 LTS (Linux-5.4.0-1055-azure-x86_64-with-Ubuntu-18.04-bionic)", "other_hardware": "", "other_software_stack": "TensorRT 7.2.3, CUDA 11.1, cuDNN 8.1.1, Driver 460.39; GCC 7.5.0; Python 3.7.10", "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": 7322528924094909334, "retraining": "N", "sample_index_rng_seed": 1570999273408051088, "samples_per_query": 84480, "schedule_rng_seed": 3507442325620259414, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "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/A40x4_TRT", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 128, "task": "image segmentation", "task2": "image segmentation", "total_cores": 240, "uid": "099980d1e589faf1", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 340599518293, "90.00 percentile latency (ns)": 614329285131, "95.00 percentile latency (ns)": 648555602707, "97.00 percentile latency (ns)": 662229475962, "99.00 percentile latency (ns)": 675929351659, "99.90 percentile latency (ns)": 682092499369, "Max latency (ns)": 682775400692, "Mean latency (ns)": 340795266288, "Min duration satisfied": "Yes", "Min latency (ns)": 115000563, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 123.73, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "48 GB", "accelerator_memory_configuration": "GDDR6", "accelerator_model_name": "NVIDIA A40", "accelerator_on-chip_memories": "", "accelerators_per_node": 4, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.7799, "characteristics.mean": 0.85383, "characteristics.samples_per_second": 123.73, "characteristics.samples_per_second.normalized_per_core": 30.9325, "characteristics.samples_per_second.normalized_per_processor": 30.9325, "characteristics.tumor core": 0.8684, "characteristics.whole tumor": 0.9132, "ck_system": "A40x4_TRT", "ck_used": true, "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 7.2.3, CUDA 11.1", "host_memory_capacity": "704 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": "2 TB", "host_storage_type": "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": 1.0, "network_speed_mbit": "", "nics_enabled_connected": "", "nics_enabled_firmware": "", "nics_enabled_os": "", "normalize_cores": 4, "normalize_processors": 4, "note_code": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/results/A40x4_TRT", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Ubuntu 18.04.5 LTS (Linux-5.4.0-1055-azure-x86_64-with-Ubuntu-18.04-bionic)", "other_hardware": "", "other_software_stack": "TensorRT 7.2.3, CUDA 11.1, cuDNN 8.1.1, Driver 460.39; GCC 7.5.0; Python 3.7.10", "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": 7322528924094909334, "retraining": "N", "sample_index_rng_seed": 1570999273408051088, "samples_per_query": 84480, "schedule_rng_seed": 3507442325620259414, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "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/A40x4_TRT", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 128, "task": "image segmentation", "task2": "image segmentation", "total_cores": 240, "uid": "ccf8c8fcf64d279d", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 405272763897, "90.00 percentile latency (ns)": 727149480282, "95.00 percentile latency (ns)": 767493532990, "97.00 percentile latency (ns)": 783613309803, "99.00 percentile latency (ns)": 799691579217, "99.90 percentile latency (ns)": 806911621214, "Max latency (ns)": 807716328512, "Mean latency (ns)": 405165671637, "Min duration satisfied": "Yes", "Min latency (ns)": 115176641, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 173.229, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "40 GB", "accelerator_memory_configuration": "HBM2", "accelerator_model_name": "NVIDIA A100-PCIe-40GB", "accelerator_on-chip_memories": "", "accelerators_per_node": 4, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.78, "characteristics.mean": 0.85387, "characteristics.samples_per_second": 173.229, "characteristics.samples_per_second.normalized_per_core": 43.30725, "characteristics.samples_per_second.normalized_per_processor": 43.30725, "characteristics.tumor core": 0.8684, "characteristics.whole tumor": 0.9132, "ck_system": "A100-PCIex4_TRT", "ck_used": true, "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 7.2.3, CUDA 11.1", "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": "2 TB", "host_storage_type": "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": 1.0, "network_speed_mbit": "", "nics_enabled_connected": "", "nics_enabled_firmware": "", "nics_enabled_os": "", "normalize_cores": 4, "normalize_processors": 4, "note_code": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/results/A100-PCIex4_TRT", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Ubuntu 18.04.5 LTS (Linux-5.4.0-1055-azure-x86_64-with-Ubuntu-18.04-bionic)", "other_hardware": "", "other_software_stack": "TensorRT 7.2.3, CUDA 11.1, cuDNN 8.1.1, Driver 460.32.03, DALI 0.30.0; GCC 7.5.0; Python 3.7.10", "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": 7322528924094909334, "retraining": "N", "sample_index_rng_seed": 1570999273408051088, "samples_per_query": 139920, "schedule_rng_seed": 3507442325620259414, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "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/A100-PCIex4_TRT", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 212, "task": "image segmentation", "task2": "image segmentation", "total_cores": 240, "uid": "b2595cf114f7daac", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 405272763897, "90.00 percentile latency (ns)": 727149480282, "95.00 percentile latency (ns)": 767493532990, "97.00 percentile latency (ns)": 783613309803, "99.00 percentile latency (ns)": 799691579217, "99.90 percentile latency (ns)": 806911621214, "Max latency (ns)": 807716328512, "Mean latency (ns)": 405165671637, "Min duration satisfied": "Yes", "Min latency (ns)": 115176641, "Min queries satisfied": "Yes", "Mode": "PerformanceOnly", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 173.229, "Scenario": "offline", "accelerator_frequency": "", "accelerator_host_interconnect": "", "accelerator_interconnect": "", "accelerator_interconnect_topology": "", "accelerator_memory_capacity": "40 GB", "accelerator_memory_configuration": "HBM2", "accelerator_model_name": "NVIDIA A100-PCIe-40GB", "accelerator_on-chip_memories": "", "accelerators_per_node": 4, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "boot_firmware_version": "", "characteristics.enhancing tumor": 0.78, "characteristics.mean": 0.85387, "characteristics.samples_per_second": 173.229, "characteristics.samples_per_second.normalized_per_core": 43.30725, "characteristics.samples_per_second.normalized_per_processor": 43.30725, "characteristics.tumor core": 0.8684, "characteristics.whole tumor": 0.9132, "ck_system": "A100-PCIex4_TRT", "ck_used": true, "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 7.2.3, CUDA 11.1", "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": "2 TB", "host_storage_type": "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": 1.0, "network_speed_mbit": "", "nics_enabled_connected": "", "nics_enabled_firmware": "", "nics_enabled_os": "", "normalize_cores": 4, "normalize_processors": 4, "note_code": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v1.0/tree/master/closed/Lenovo/results/A100-PCIex4_TRT", "number_of_nodes": 1, "number_of_type_nics_installed": "", "operating_system": "Ubuntu 18.04.5 LTS (Linux-5.4.0-1055-azure-x86_64-with-Ubuntu-18.04-bionic)", "other_hardware": "", "other_software_stack": "TensorRT 7.2.3, CUDA 11.1, cuDNN 8.1.1, Driver 460.32.03, DALI 0.30.0; GCC 7.5.0; Python 3.7.10", "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": 7322528924094909334, "retraining": "N", "sample_index_rng_seed": 1570999273408051088, "samples_per_query": 139920, "schedule_rng_seed": 3507442325620259414, "starting_weights_filename": "224_224_160_dyanmic_bs.onnx", "status": "available", "submitter": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "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/A100-PCIex4_TRT", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 212, "task": "image segmentation", "task2": "image segmentation", "total_cores": 240, "uid": "ca833ae0c784e80a", "use_accelerator": true, "weight_data_types": "int8", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 79565579286, "90.00 percentile latency (ns)": 143452728603, "95.00 percentile latency (ns)": 151459415300, "97.00 percentile latency (ns)": 154646527284, "99.00 percentile latency (ns)": 157849752607, "99.90 percentile latency (ns)": 159284308159, "Max latency (ns)": 159439461424, "Mean latency (ns)": 79625330399, "Min duration satisfied": "Yes", "Min latency (ns)": 68150773, "Min queries satisfied": "Yes", "Mode": "Performance", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 154.14, "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": 0.7796, "characteristics.mean": 0.8538, "characteristics.samples_per_second": 154.14, "characteristics.samples_per_second.normalized_per_core": 38.535, "characteristics.samples_per_second.normalized_per_processor": 38.535, "characteristics.tumor core": 0.8693, "characteristics.whole tumor": 0.9125, "ck_system": "A100-PCIex4_TRT72", "ck_used": true, "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": "770 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": "25.6 TB", "host_storage_type": "SAS 2.5 SSD", "hw_notes": "ECC off", "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/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Lenovo/results/A100-PCIex4_TRT72", "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": "TensorRT 7.2, CUDA 11.0 Update 1, cuDNN 8.0.2, DALI 0.25.0; GCC 7.5.0; Python 3.7.10", "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": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "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/A100-PCIex4_TRT72", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 200, "task": "image segmentation", "task2": "image segmentation", "total_cores": 240, "uid": "65b9407e9181d071", "use_accelerator": true, "weight_data_types": "int8,fp16", "weight_transformations": "quantization, affine fusion" }, { "50.00 percentile latency (ns)": 79565579286, "90.00 percentile latency (ns)": 143452728603, "95.00 percentile latency (ns)": 151459415300, "97.00 percentile latency (ns)": 154646527284, "99.00 percentile latency (ns)": 157849752607, "99.90 percentile latency (ns)": 159284308159, "Max latency (ns)": 159439461424, "Mean latency (ns)": 79625330399, "Min duration satisfied": "Yes", "Min latency (ns)": 68150773, "Min queries satisfied": "Yes", "Mode": "Performance", "Result is": "VALID", "SUT name": "LWIS_Server", "Samples per second": 154.14, "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": 0.7796, "characteristics.mean": 0.8538, "characteristics.samples_per_second": 154.14, "characteristics.samples_per_second.normalized_per_core": 38.535, "characteristics.samples_per_second.normalized_per_processor": 38.535, "characteristics.tumor core": 0.8693, "characteristics.whole tumor": 0.9125, "ck_system": "A100-PCIex4_TRT72", "ck_used": true, "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": "770 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": "25.6 TB", "host_storage_type": "SAS 2.5 SSD", "hw_notes": "ECC off", "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/Lenovo/code", "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/Lenovo/results/A100-PCIex4_TRT72", "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": "TensorRT 7.2, CUDA 11.0 Update 1, cuDNN 8.0.2, DALI 0.25.0; GCC 7.5.0; Python 3.7.10", "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": "Lenovo", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Lenovo", "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/A100-PCIex4_TRT72", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "system_type": "datacenter", "target_latency (ns)": 0, "target_qps": 200, "task": "image segmentation", "task2": "image segmentation", "total_cores": 240, "uid": "786d468197aba9fa", "use_accelerator": true, "weight_data_types": "int8,fp16", "weight_transformations": "quantization, affine fusion" } ]