np from tvm import relay 2. Load a pretrained network mod, params = relay.testing.mobilenet.get_workload(batch_size=1) 3. Partition and build the network with an external codegen mod = relay.build_extern(mod ib//graph_annotator.py ● Apply the annotator to a workload: mod, params = relay.testing.mobilenet.get_workload(batch_size=1) mod[‘main’] = MyAnnotator().visit(mod[‘main’]) mod = relay

Applications  Evaluate in larger scope of real datasets and query workloads  Examples: Customer workload, ML pipelines (dataprep, training, eval) Experiments and Result Presentation 7 706.015 Introduction many ways how to do it wrong  Basic Planning  Which data / data sets should be used?  Which workload / queries should be run?  Which hardware & software should be used?  Metrics: What to measure Reproducibility  Verify robustness of results wrt parameters and environments  Examples: different data and workload characteristics, hardware Reproducibility and RDM 26 706.015 Introduction to Scientific Writing

让 1 splits the workload into thread <| | Apaday+my blocks Bly zx) https://docstvm ai/ PVR TOPI Alibaba ALLabs 阿里巴巴人工智能实验室 Blocking Splits the workload into thread blocks (work groups) and individual threads (work items) Processing Element

45 .31让工 1.31 -35 1 129 中131 124有23152136 2 1.14 am omo oo Convolution Workload Performance AiOS 1驱动万物智能 Alios TVM @ ARM CPU INT8 Depthwise Convolution 。， NHWC layout 。 Using 33. 1.15 116 111 09工08 工区 0.77 0.77 | | | Depthwise Convolution Workload Performance Alios TVM @ ARM CPU INT8 Performance Comparison @ rasp 3b+ AARCH64 aoo0 8.87 sm ao

PerformanceNVIDIA GPU Performance = +225x Over Eight Years 106 1 GPT-MoE Inference Workload = A type of workload where a GPT-style model with a Mixture-of-Experts (MoE) architecture is used for inference over eight years while requiring 4x fewer GPUs… $1B Data Center Comparison GPT-MoE Inference Workload1 …Inference token capacity +27,500x over eight years, implying +30,000x higher theoretical

func = remote_mod[“npufunction0"] func(remote_a, remote_b)Virtual Machine: Supporting Dynamic Workload Dynamic shape workloads More runtime objects: Arrays, Tuples, Trees, ADTs Minimum runtime for

projects, with a total refactor of the software with new functions, redesign of the internal window manager and event/tool/data handling system, and new Python API. The final version of this project was Blender Blender. These settings are typically found in conjunction with the Window Manager settings. (Gnome or KDE.) Install from Package Manager Some Linux distributions may have a specific package for Blender in consistency with other packages on the system and may provide other features (given by the package manager), such as listing of packages, update notifications and automatic menu configuration. Be aware, though

projects, with a total refactor of the software with new functions, redesign of the internal window manager and event/tool/data handling system, and new Python API. The final version of this project was Blender Blender. These settings are typically found in conjunction with the Window Manager settings. (Gnome or KDE.) Install from Package Manager Some Linux distributions may have a specific package for Blender in consistency with other packages on the system and may provide other features (given by the package manager), such as listing of packages, update notifications and automatic menu configuration. Be aware, though

projects, with a total refactor of the software with new functions, redesign of the internal window manager and event/tool/data handling system, and new Python API. The final version of this project was Blender Blender. These settings are typically found in conjunction with the Window Manager settings. (Gnome or KDE.) Install from Package Manager Some Linux distributions may have a specific package for Blender in consistency with other packages on the system and may provide other features (given by the package manager), such as listing of packages, update notifications and automatic menu configuration. Be aware, though

projects, with a total refactor of the software with new functions, redesign of the internal window manager and event/tool/data handling system, and new Python API. The final version of this project was Blender Blender. These settings are typically found in conjunction with the Window Manager settings. (Gnome or KDE.) Install from Package Manager Some Linux distributions may have a specific package for Blender in consistency with other packages on the system and may provide other features (given by the package manager), such as listing of packages, update notifications and automatic menu configuration. Be aware, though