Open Discussion on Project Planning Planning in an Agile Environment Key Tenets  Planning should focus strongly on the near term  Eliminate waste caused by planning for eventualities that never come

AI Usage + Cost + Loss Growth = Unprecedented • AI Monetization Threats = Rising Competition + Open-Source Momentum + China’s Rise • AI & Physical World Ramps = Fast + Data-Driven • Global Internet User User + Usage + CapEx Growth = Unprecedented Developers in Leading Chipmaker’s Ecosystem 1 2.1 Source: Leading Chipmaker Details on Page 38 AI User + Usage + CapEx Growth = Unprecedented 2.2 Internet mobile app users. App not available in select countries, including China and Russia, as of 5/25. Source: United Nations / International Telecommunications Union (3/25), Sensor Tower (5/25) 0 Years In

track progress during a sprint. Figure 3: Example Burn Down Chart Program Backlog – Primary source of all requirements/desired functionality for the program Release Backlog – Subset of the program features the team considers relevant to building the product. The program backlog serves as the primary source for all program requirements and user stories, and the team must prioritize the contents to ensure the acquisition leadership, contracting, test, Certification and Accreditation (C&A), the user community, external systems, and cost and financial support.11 Figure 5: Potential Agile Team Construct

Inference engines DL Compilers Kenrel Libraries Hardware CuDNN NNPack MKL-DNN Hand optimized Open source, automated end-to- end optimization framework for deep learning.TVM Stack High-Level Differentiable tvm.sum(A[k, y] * B[k], axis=k)) HW Interface Specification by Tensor Expression TensorizationVTA: Open & Flexible Deep Learning Accelerator • Runtime JIT compile accelerator micro code • Support heterogenous Interface (ISA) VTA MicroArchitecture VTA Simulator} compiler, driver, hardware design full stack open source Current TVM Stack VTA Runtime & JIT CompilerTSIM: Support for Future Hardware Current TVM Stack

Q OctoML Open Source at O〇ctoML TVM Meetup 11/8/2019 Jared Roesch OctoML is a new company building DL deployment solutions using the Apache (incubating) TVM project. A goal is to nurture the TVM community design and machine learning tr tvm 。 @zxnet 和os 全 W Open Source at OctoML ee We are big believers in the power of open source o 5S$ponsoring multiple employees to contribute to TVML. ee Today octoML Coalesced t1: Tensor t2: Tensor t3: Tensor 13 Acknowledgments e The Apache(incubating) community members. e ASF Mentors and PMC members who make this awesome project Possiblel ee AWS for hosting

greatest uncertainty—the beginning of the initiative.  Working in an Agile way, we can leave options open and be prepared to accommodate a range of scenarios as the future unfolds.  Jez Humble and his decisions in the normal course of work.  Let’s say that we are deciding between two different open source products for building a piece of the system and do not know enough of their impacts to make the working support a community approach to IT, where IT leaders achieve their objectives by mobilizing the skills and passions of a broad community and encourage the members of that community to work together

reduce cost and risk.  Custom code is almost not custom these days. A developer incorporates open source frameworks, uses standardized design patterns, and orchestrates services that are already available working support a community approach to IT, where IT leaders achieve their objectives by mobilizing the skills and passions of a broad community and encourage the members of that community to work together

hosted models • MXNet Pre-quantized Models • Tested internally with MxNet + MKLDNN path • Will open RFC in a month© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Evaluation opt© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Conclusion • TVM community is pursuing both Automatic- and Pre-quantized model support. Contributions are welcomed. • We VNNI, ARM Dot, Nvidia DP4A • Full pipeline is available. Please try it and give suggestions. • Open-source discussions formed the foundations of both the approaches.

(Qualcomm) PART TWO Alios TVM @ ARM CPU AiOS 1驱动万物智能 Alios TVMQOARM CPU 。 Support TFLite ( Open Source and Upstream Master ) 。， Optimize on INT8 & FP32 AiiOS ! 驱动万物智能 Alios TVM @ ARM CPU INT8 * Kernel Offload to DSP ，loop nests marked as pipeline 。， Implement complete Hexagon runtime based on community PR. ADSPRPC Framework Applications Processor | | DSP Processor

. . . . . . . . . . . . . . . . . . . . . . . . . 1624 101.26 Fixing precompilation hangs due to open tasks or IO . . . . . . . . . . . . . . . . . 1626 102 Developing/debugging Julia’s C code 1630 In addition to the above, some advantages of Julia over comparable systems include: • Free and open source (MIT licensed) • User-defined types are as fast and compact as built-ins • No need to vectorize Started Julia installation is straightforward, whether using precompiled binaries or compiling from source. Download and install Julia by following the instructions at https://julialang.org/install/. If