[ { "50.00 percentile": 1, "50.00 percentile latency (ns)": 3199609, "90.00 percentile": 1, "90.00 percentile latency (ns)": 3232846, "95.00 percentile": 1, "95.00 percentile latency (ns)": 3241216, "97.00 percentile": 1, "97.00 percentile latency (ns)": 3247895, "99.00 percentile": 1, "99.00 percentile latency (ns)": 3260154, "99.90 percentile": 1, "99.90 percentile latency (ns)": 3287938, "Max latency (ns)": 6290287, "Mean latency (ns)": 3199472, "Min duration satisfied": "Yes", "Min latency (ns)": 3017381, "Min queries satisfied": "Yes", "Mode": "Performance", "Per-sample latency": "", "Performance constraints satisfied": "Yes", "Result is": "VALID", "SUT name": "HABANA_Runner\u0000", "Samples per query": 40, "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.accuracy": 76.206, "characteristics.good": 38103, "characteristics.samples_per_query": 40, "characteristics.samples_per_query.normalized_per_core": 40.0, "characteristics.samples_per_query.normalized_per_processor": 40.0, "characteristics.total": 50000, "ck_system": "Goya_med_latency", "ck_used": true, "dataset": "ImageNet 2012", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/imagenet2012.md", "dim_x_default": "characteristics.samples_per_query", "dim_y_default": "characteristics.accuracy", "dim_y_maximize": true, "division": "open", "formal_model": "resnet", "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": "resnet", "input_data_types": "INT8", "key.accuracy": "characteristics.accuracy", "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": 1024, "print_timestamps": true, "problem": false, "qsl_rng_seed": 3133965575612453542, "retraining": "No", "sample_index_rng_seed": 665484352860916858, "samples_per_query": 40, "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)": 3389830, "target_qps": 295, "task": "image classification", "task2": "image classification", "total_cores": 4800, "uid": "504c28ab419cdabb", "use_accelerator": true, "weight_data_types": "INT8", "weight_transformations": "Asymmetric Quantization" }, { "50.00 percentile": 1, "50.00 percentile latency (ns)": 1790446, "90.00 percentile": 1, "90.00 percentile latency (ns)": 1824451, "95.00 percentile": 1, "95.00 percentile latency (ns)": 1833847, "97.00 percentile": 1, "97.00 percentile latency (ns)": 1840477, "99.00 percentile": 1, "99.00 percentile latency (ns)": 1847727, "99.90 percentile": 1, "99.90 percentile latency (ns)": 1856718, "Max latency (ns)": 2401567, "Mean latency (ns)": 1794586, "Min duration satisfied": "Yes", "Min latency (ns)": 1614182, "Min queries satisfied": "Yes", "Mode": "Performance", "Per-sample latency": "", "Performance constraints satisfied": "Yes", "Result is": "VALID", "SUT name": "HABANA_Runner\u0000", "Samples per query": 20, "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.accuracy": 76.206, "characteristics.good": 38103, "characteristics.samples_per_query": 20, "characteristics.samples_per_query.normalized_per_core": 20.0, "characteristics.samples_per_query.normalized_per_processor": 20.0, "characteristics.total": 50000, "ck_system": "Goya_fast_latency", "ck_used": true, "dataset": "ImageNet 2012", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/imagenet2012.md", "dim_x_default": "characteristics.samples_per_query", "dim_y_default": "characteristics.accuracy", "dim_y_maximize": true, "division": "open", "formal_model": "resnet", "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": "resnet", "input_data_types": "INT8", "key.accuracy": "characteristics.accuracy", "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": 1024, "print_timestamps": true, "problem": false, "qsl_rng_seed": 3133965575612453542, "retraining": "No", "sample_index_rng_seed": 665484352860916858, "samples_per_query": 20, "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)": 2000000, "target_qps": 500, "task": "image classification", "task2": "image classification", "total_cores": 4800, "uid": "b5164ff4b8fc831e", "use_accelerator": true, "weight_data_types": "INT8", "weight_transformations": "Asymmetric Quantization" }, { "50.00 percentile": 1, "50.00 percentile latency (ns)": 48932824, "90.00 percentile": 1, "90.00 percentile latency (ns)": 48971884, "95.00 percentile": 1, "95.00 percentile latency (ns)": 48977212, "97.00 percentile": 1, "97.00 percentile latency (ns)": 48980819, "99.00 percentile": 1, "99.00 percentile latency (ns)": 48988971, "99.90 percentile": 1, "99.90 percentile latency (ns)": 49011123, "Max latency (ns)": 52159315, "Mean latency (ns)": 48937168, "Min duration satisfied": "Yes", "Min latency (ns)": 48742057, "Min queries satisfied": "Yes", "Mode": "Performance", "Per-sample latency": "", "Performance constraints satisfied": "Yes", "Result is": "VALID", "SUT name": "HABANA_Runner\u0000", "Samples per query": 700, "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.accuracy": 76.206, "characteristics.good": 38103, "characteristics.samples_per_query": 700, "characteristics.samples_per_query.normalized_per_core": 700.0, "characteristics.samples_per_query.normalized_per_processor": 700.0, "characteristics.total": 50000, "ck_system": "Goya_1", "ck_used": true, "dataset": "ImageNet 2012", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/imagenet2012.md", "dim_x_default": "characteristics.samples_per_query", "dim_y_default": "characteristics.accuracy", "dim_y_maximize": true, "division": "closed", "formal_model": "resnet50-v1.5", "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": "resnet", "input_data_types": "INT8", "key.accuracy": "characteristics.accuracy", "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": 1024, "print_timestamps": true, "problem": false, "qsl_rng_seed": 3133965575612453542, "retraining": "No", "sample_index_rng_seed": 665484352860916858, "samples_per_query": 700, "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)": 50000000, "target_qps": 20, "task": "image classification", "task2": "image classification", "total_cores": 240, "uid": "78afb36c0140ea4f", "use_accelerator": true, "weight_data_types": "INT8", "weight_transformations": "Asymmetric Quantization" } ]