[ { "50.00 percentile latency (ns)": 3374572, "90.00 percentile latency (ns)": 3540689, "90th percentile latency (ns)": 3540689, "95.00 percentile latency (ns)": 3569535, "97.00 percentile latency (ns)": 3585031, "99.00 percentile latency (ns)": 3642882, "99.90 percentile latency (ns)": 5966328, "Max latency (ns)": 9757648, "Mean latency (ns)": 3393792, "Min duration satisfied": "Yes", "Min latency (ns)": 2768533, "Min queries satisfied": "Yes", "Mode": "Performance", "QPS w/ loadgen overhead": 289.37, "QPS w/o loadgen overhead": 294.66, "Result is": "VALID", "SUT name": "IVA FPGA1", "Scenario": "singlestream", "accelerator_frequency": "250MHz", "accelerator_memory_capacity": "16 GB per accelerator", "accelerator_memory_configuration": "HBM", "accelerator_model_name": "IVA_FPGA_1", "accelerator_on-chip_memories": "16x 64K", "accelerators_per_node": 1, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "characteristics.90th_percentile_latency_ms": 3.540689, "characteristics.90th_percentile_latency_ns": 3540689.0, "characteristics.90th_percentile_latency_s": 0.003540689, "characteristics.90th_percentile_latency_us": 3540.689, "characteristics.accuracy": 70.14, "characteristics.good": 35070, "characteristics.total": 50000, "ck_system": "iva-fpga-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.90th_percentile_latency_ms", "dim_y_default": "characteristics.accuracy", "dim_y_maximize": true, "division": "open", "formal_model": "mobilenet", "formal_model_accuracy": 99.0, "formal_model_link": "https://github.com/mlcommons/ck-mlops/tree/main/package", "framework": "Tensorflow 1.15.0", "host_memory_capacity": "16 GB", "host_processor_core_count": 120, "host_processor_model_name": "AMD EPYC 7V13 64-Core Processor", "host_processors_per_node": 1, "host_storage_capacity": "500 GB", "host_storage_type": "SSD", "hw_notes": "Alveo U200 Data Center Accelerator Card(A-U200-A64G-PQ-G)", "informal_model": "mobilenet", "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": 1024, "mlperf_version": 0.7, "normalize_cores": 1, "normalize_processors": 1, "note_code": "https://github.com/mlcommons/inference_results_v0.7/tree/master/open/IVATech/code", "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/open/IVATech/results/iva-fpga-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": "iva_tpu 4.1.4, libtpu 7.1.15, tpu_quantize 3.2.1, tpu-compiler 13.1.0, gcc 10.1.0, tbb 2020.3; 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": 12786827339337101903, "retraining": "N", "sample_index_rng_seed": 12640797754436136668, "samples_per_query": 1, "schedule_rng_seed": 3135815929913719677, "starting_weights_filename": "https://zenodo.org/record/2269307/files/mobilenet_v1_1.0_224.tgz", "status": "RDI", "submitter": "IVATech", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/IVATech", "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/iva-fpga-1", "system_name": "Microsoft Corporation 7.0 (Virtual Machine)", "system_type": "edge", "target_latency (ns)": 0, "target_qps": 500, "task": "image classification", "task2": "image classification", "total_cores": 120, "uid": "b0a3095a74d3caa8", "use_accelerator": true, "weight_data_types": "int8,fp16", "weight_transformations": "quantization" }, { "50.00 percentile latency (ns)": 12062959, "90.00 percentile latency (ns)": 12230138, "90th percentile latency (ns)": 12230138, "95.00 percentile latency (ns)": 12246475, "97.00 percentile latency (ns)": 12258464, "99.00 percentile latency (ns)": 12345478, "99.90 percentile latency (ns)": 13598707, "Max latency (ns)": 14198754, "Mean latency (ns)": 12081844, "Min duration satisfied": "Yes", "Min latency (ns)": 11555095, "Min queries satisfied": "Yes", "Mode": "Performance", "QPS w/ loadgen overhead": 82.37, "QPS w/o loadgen overhead": 82.77, "Result is": "VALID", "SUT name": "IVA FPGA1", "Scenario": "singlestream", "accelerator_frequency": "250MHz", "accelerator_memory_capacity": "16 GB per accelerator", "accelerator_memory_configuration": "HBM", "accelerator_model_name": "IVA_FPGA_1", "accelerator_on-chip_memories": "16x 64K", "accelerators_per_node": 1, "accuracy_log_probability": 0, "accuracy_log_rng_seed": 0, "accuracy_log_sampling_target": 0, "characteristics.90th_percentile_latency_ms": 12.230138, "characteristics.90th_percentile_latency_ns": 12230138.0, "characteristics.90th_percentile_latency_s": 0.012230138, "characteristics.90th_percentile_latency_us": 12230.138, "characteristics.accuracy": 75.756, "characteristics.good": 37878, "characteristics.total": 50000, "ck_system": "iva-fpga-1", "ck_used": false, "dataset": "ImageNet 2012", "dataset_link": "https://github.com/ctuning/ck/blob/master/docs/mlperf-automation/datasets/imagenet2012.md", "dim_x_default": "characteristics.90th_percentile_latency_ms", "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": "Tensorflow 1.15.0", "host_memory_capacity": "16 GB", "host_processor_core_count": 4, "host_processor_model_name": "Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz", "host_processors_per_node": 1, "host_storage_capacity": "500 GB", "host_storage_type": "SSD", "hw_notes": "Alveo U200 Data Center Accelerator Card(A-U200-A64G-PQ-G)", "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": 1024, "mlperf_version": 0.7, "normalize_cores": 1, "normalize_processors": 1, "note_code": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/IVATech/code", "note_details": "https://github.com/mlcommons/inference_results_v0.7/tree/master/closed/IVATech/results/iva-fpga-1", "number_of_nodes": 1, "operating_system": "Ubuntu 18.04", "other_software_stack": "iva_tpu 4.1.4, libtpu 7.1.15, tpu_quantize 3.2.1, tpu-compiler 13.1.0, gcc 10.1.0, tbb 2020.3", "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": 12786827339337101903, "retraining": "N", "sample_index_rng_seed": 12640797754436136668, "samples_per_query": 1, "schedule_rng_seed": 3135815929913719677, "starting_weights_filename": "https://zenodo.org/record/2535873/files/resnet50_v1.pb", "status": "RDI", "submitter": "IVATech", "submitter_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.submitter/IVATech", "system_link": "https://github.com/ctuning/ck-mlperf-inference/tree/main/bench.mlperf.system/iva-fpga-1", "system_name": "iva-fpga-1", "system_type": "edge", "target_latency (ns)": 0, "target_qps": 100, "task": "image classification", "task2": "image classification", "total_cores": 4, "uid": "769483e452fbc27e", "use_accelerator": true, "weight_data_types": "int8,fp16", "weight_transformations": "quantization" } ]