[ { "50.00 percentile": 1, "50.00 percentile latency (ns)": 28844483, "90.00 percentile": 1, "90.00 percentile latency (ns)": 29680775, "95.00 percentile": 1, "95.00 percentile latency (ns)": 29951402, "97.00 percentile": 1, "97.00 percentile latency (ns)": 30056945, "99.00 percentile": 1, "99.00 percentile latency (ns)": 30243666, "99.90 percentile": 2, "99.90 percentile latency (ns)": 32790881, "Max latency (ns)": 36928259, "Mean latency (ns)": 28970685, "Min duration satisfied": "Yes", "Min latency (ns)": 27642563, "Min queries satisfied": "Yes", "Mode": "Performance", "Per-sample latency": "", "Performance constraints satisfied": "Yes", "Result is": "VALID", "SUT name": "HABANA_Runner\u0000", "Samples per query": 8, "Scenario": "multistream", "accelerator_frequency": "-", "accelerator_host_interconnect": "-", "accelerator_interconnect": "-", "accelerator_interconnect_topology": "-", "accelerator_memory_capacity": "16GB", "accelerator_memory_configuration": "none", "accelerator_model_name": "Goya", "accelerator_on-chip_memories": "-", "accelerators_per_node": 1, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "characteristics.mAP": 19.81, "characteristics.samples_per_query": 8, "characteristics.samples_per_query.normalized_per_core": 8.0, "characteristics.samples_per_query.normalized_per_processor": 8.0, "ck_system": "Goya_med_latency", "ck_used": true, "dataset": "COCO 2017 (300x300)", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/coco2017.md", "dim_x_default": "characteristics.samples_per_query", "dim_y_default": "characteristics.mAP", "dim_y_maximize": true, "division": "open", "formal_model": "ssd-large", "formal_model_accuracy": 99.0, "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package", "framework": "Synapse-V0.2.0", "host_memory_capacity": "64GB", "host_processor_core_count": 120, "host_processor_frequency": "2.20GHz", "host_processor_model_name": "AMD EPYC 7V13 64-Core Processor", "host_processors_per_node": 40, "host_storage_capacity": "1TB", "host_storage_type": "SSD", "informal_model": "ssd-large", "input_data_types": "INT8", "key.accuracy": "characteristics.mAP", "max_async_queries": 1, "max_duration (ms)": 0, "max_query_count": 0, "min_duration (ms)": 60000, "min_query_count": 270336, "mlperf_version": 0.5, "normalize_cores": 1, "normalize_processors": 1, "note_code": "https://github.com/mlcommons/inference_results_v0.5/tree/master/open/Habana/code", "note_details": "https://github.com/mlcommons/inference_results_v0.5/tree/master/open/Habana/results/Goya_med_latency", "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": "Synapse-V0.2.0; GCC 7.5.0; Python 3.7.10", "performance_issue_same": true, "performance_issue_same_index": 0, "performance_issue_unique": true, "performance_sample_count": 64, "print_timestamps": true, "problem": false, "qsl_rng_seed": 3133965575612453542, "retraining": "No", "sample_index_rng_seed": 665484352860916858, "samples_per_query": 8, "schedule_rng_seed": 3622009729038561421, "starting_weights_filename": "onnx-model from zenado", "status": "available", "submitter": "Habana", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Habana", "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/Goya_med_latency", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "target_latency (ns)": 31250000, "target_qps": 32, "task": "object detection", "task2": "object detection", "total_cores": 4800, "uid": "7b4450a6b0a689c3", "use_accelerator": true, "weight_data_types": "INT8", "weight_transformations": "Asymmetric Quantization" }, { "50.00 percentile": 1, "50.00 percentile latency (ns)": 14781894, "90.00 percentile": 1, "90.00 percentile latency (ns)": 15416775, "95.00 percentile": 1, "95.00 percentile latency (ns)": 15695602, "97.00 percentile": 1, "97.00 percentile latency (ns)": 15830298, "99.00 percentile": 1, "99.00 percentile latency (ns)": 15963426, "99.90 percentile": 1, "99.90 percentile latency (ns)": 16300190, "Max latency (ns)": 18029176, "Mean latency (ns)": 14864410, "Min duration satisfied": "Yes", "Min latency (ns)": 13987812, "Min queries satisfied": "Yes", "Mode": "Performance", "Per-sample latency": "", "Performance constraints satisfied": "Yes", "Result is": "VALID", "SUT name": "HABANA_Runner\u0000", "Samples per query": 4, "Scenario": "multistream", "accelerator_frequency": "-", "accelerator_host_interconnect": "-", "accelerator_interconnect": "-", "accelerator_interconnect_topology": "-", "accelerator_memory_capacity": "16GB", "accelerator_memory_configuration": "none", "accelerator_model_name": "Goya", "accelerator_on-chip_memories": "-", "accelerators_per_node": 1, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "characteristics.mAP": 19.81, "characteristics.samples_per_query": 4, "characteristics.samples_per_query.normalized_per_core": 4.0, "characteristics.samples_per_query.normalized_per_processor": 4.0, "ck_system": "Goya_fast_latency", "ck_used": true, "dataset": "COCO 2017 (300x300)", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/coco2017.md", "dim_x_default": "characteristics.samples_per_query", "dim_y_default": "characteristics.mAP", "dim_y_maximize": true, "division": "open", "formal_model": "ssd-large", "formal_model_accuracy": 99.0, "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package", "framework": "Synapse-V0.2.0", "host_memory_capacity": "64GB", "host_processor_core_count": 120, "host_processor_frequency": "2.20GHz", "host_processor_model_name": "AMD EPYC 7V13 64-Core Processor", "host_processors_per_node": 40, "host_storage_capacity": "1TB", "host_storage_type": "SSD", "informal_model": "ssd-large", "input_data_types": "INT8", "key.accuracy": "characteristics.mAP", "max_async_queries": 1, "max_duration (ms)": 0, "max_query_count": 0, "min_duration (ms)": 60000, "min_query_count": 270336, "mlperf_version": 0.5, "normalize_cores": 1, "normalize_processors": 1, "note_code": "https://github.com/mlcommons/inference_results_v0.5/tree/master/open/Habana/code", "note_details": "https://github.com/mlcommons/inference_results_v0.5/tree/master/open/Habana/results/Goya_fast_latency", "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": "Synapse-V0.2.0; GCC 7.5.0; Python 3.7.10", "performance_issue_same": true, "performance_issue_same_index": 0, "performance_issue_unique": true, "performance_sample_count": 64, "print_timestamps": true, "problem": false, "qsl_rng_seed": 3133965575612453542, "retraining": "No", "sample_index_rng_seed": 665484352860916858, "samples_per_query": 4, "schedule_rng_seed": 3622009729038561421, "starting_weights_filename": "onnx-model from zenado", "status": "available", "submitter": "Habana", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Habana", "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/Goya_fast_latency", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "target_latency (ns)": 16666666, "target_qps": 60, "task": "object detection", "task2": "object detection", "total_cores": 4800, "uid": "aed81a60a4501a87", "use_accelerator": true, "weight_data_types": "INT8", "weight_transformations": "Asymmetric Quantization" }, { "50.00 percentile": 1, "50.00 percentile latency (ns)": 64346890, "90.00 percentile": 1, "90.00 percentile latency (ns)": 65179373, "95.00 percentile": 1, "95.00 percentile latency (ns)": 65413688, "97.00 percentile": 1, "97.00 percentile latency (ns)": 65560526, "99.00 percentile": 1, "99.00 percentile latency (ns)": 66383840, "99.90 percentile": 2, "99.90 percentile latency (ns)": 69223871, "Max latency (ns)": 72476929, "Mean latency (ns)": 64387125, "Min duration satisfied": "Yes", "Min latency (ns)": 61821336, "Min queries satisfied": "Yes", "Mode": "Performance", "Per-sample latency": "", "Performance constraints satisfied": "Yes", "Result is": "VALID", "SUT name": "HABANA_Runner\u0000", "Samples per query": 18, "Scenario": "multistream", "accelerator_frequency": "-", "accelerator_host_interconnect": "-", "accelerator_interconnect": "-", "accelerator_interconnect_topology": "-", "accelerator_memory_capacity": "16GB", "accelerator_memory_configuration": "none", "accelerator_model_name": "Goya", "accelerator_on-chip_memories": "-", "accelerators_per_node": 1, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "characteristics.mAP": 19.81, "characteristics.samples_per_query": 18, "characteristics.samples_per_query.normalized_per_core": 18.0, "characteristics.samples_per_query.normalized_per_processor": 18.0, "ck_system": "Goya_1", "ck_used": true, "dataset": "COCO 2017 (300x300)", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/coco2017.md", "dim_x_default": "characteristics.samples_per_query", "dim_y_default": "characteristics.mAP", "dim_y_maximize": true, "division": "closed", "formal_model": "ssd-mobilenet", "formal_model_accuracy": 99.0, "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package", "framework": "Synapse-V0.2.0", "host_memory_capacity": "64GB", "host_processor_core_count": 120, "host_processor_frequency": "2.20GHz", "host_processor_model_name": "AMD EPYC 7V13 64-Core Processor", "host_processors_per_node": 2, "host_storage_capacity": "1TB", "host_storage_type": "SSD", "informal_model": "ssd-large", "input_data_types": "INT8", "key.accuracy": "characteristics.mAP", "max_async_queries": 1, "max_duration (ms)": 0, "max_query_count": 0, "min_duration (ms)": 60000, "min_query_count": 270336, "mlperf_version": 0.5, "normalize_cores": 1, "normalize_processors": 1, "note_code": "https://github.com/mlcommons/inference_results_v0.5/tree/master/closed/Habana/code", "note_details": "https://github.com/mlcommons/inference_results_v0.5/tree/master/closed/Habana/results/Goya_1", "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": "Synapse-V0.2.0; GCC 7.5.0; Python 3.7.10", "performance_issue_same": true, "performance_issue_same_index": 0, "performance_issue_unique": true, "performance_sample_count": 64, "print_timestamps": true, "problem": false, "qsl_rng_seed": 3133965575612453542, "retraining": "No", "sample_index_rng_seed": 665484352860916858, "samples_per_query": 18, "schedule_rng_seed": 3622009729038561421, "starting_weights_filename": "onnx-model from zenado", "status": "available", "submitter": "Habana", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/Habana", "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/Goya_1", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "target_latency (ns)": 66666666, "target_qps": 15, "task": "object detection", "task2": "object detection", "total_cores": 240, "uid": "6cec42fea4866a96", "use_accelerator": true, "weight_data_types": "INT8", "weight_transformations": "Asymmetric Quantization" } ]