.. _Tutorials: ====================================================================================================== MAESTRO Tutorial - `MICRO 2020 `_ ====================================================================================================== MAESTRO: A Data-Centric Approach for Hardware and Mapping Explorations for Deep Learning Accelerators ---------------------------------------------------------------------------------------------------------------- Date: October 17, 2020 | Organizers ------------------------------------ - `Tushar Krishna `_ (Georgia Tech) - `Michael Pellauer `_ (NVIDIA) - `Prasanth Chatarasi `_ (Georgia Tech) - `Geonhwa Jeong `_ (Georgia Tech) - `Sheng-Chun (Felix) Kao `_ (Georgia Tech) | Overview ------------------------------------ .. image:: imgs/maestro_overview.png | The efficiency of a deep neural network (DNN) accelerator depends on three factors—mapping, DNN layers, and hardware—constructing a extremely complicated design space of DNN accelerators. To demystify this design space and guide the DNN accelerator design for better efficiency, we present an analytical cost model, MAESTRO (MICRO 2019 + IEEE Micro Top Picks 2020). MAESTRO receives the DNN model description, hardware resources information, and mapping (described in a data-centric representation) as inputs. The data-centric representation consists of three directives that enable concise description of mappings in a compiler-friendly form. MAESTRO analyzes various forms of data reuse in an accelerator based on inputs quickly and generates more than 20 statistics including total latency, energy, throughput, etc., as outputs. We also present various optimization tools for automated hardware design-space and mapping-space exploration enabled by MAESTRO’s fast analysis. | Schedule (Eastern Time) ------------------------------------ ===================== ====================================================================== ================ ================== **Time** **Agenda** **Presenter** **Resources** 10:00 - 10:10 **Welcome + Intro to DNNs** Tushar [:download:`Slides `][`Video `_] 10:10 - 10:45 **DNN Dataflows** Michael [:download:`Slides `][`Video `_] \ **MAESTRO Cost Model** \ \ 10:45 - 11:15 Data-centric Mapping Directives Hyoukjun^ [:download:`Slides `][`Video `_] 11:15 - 11:30 Analytical Cost Model Hyoukjun^ [:download:`Slides `][`Video `_] 11:30 - 12:00 Compiling and Running MAESTRO Geonhwa [:download:`Slides `][`Video `_][`GitHub `_] 12:00 - 12:10 **Break** \ \ \ **Automated Design-space Exploration using MAESTRO** \ \ 12:10 - 12:25 Marvel: Mapping Space Exploration via Heuristics Prasanth [:download:`Slides `][`Video `_][GitHub - coming soon] 12:25 - 12:40 GAMMA: Mapping Space Exploration via Optimization Felix [:download:`Slides `][`Video `_][`GitHub `_] 12:40 - 12:55 ConfuciuX: Hardware Design-space Exploration via RL and Optimization Felix [:download:`Slides `][`Video `_][`GitHub `_] 12:55 - 13:00 **Wrap Up** Tushar \ ===================== ====================================================================== ================ ================== ^Pre-recorded Video | .. |test| raw:: html
We provide the docker image which includes MAESTRO binary with proper environment settings.
Resources ------------------------------------ - `MAESTRO Project GitHub `_ - `MAESTRO Website `_ - `MAESTRO YouTube Channel `_ - `Docker Image `_ which includes MAESTRO binary file with proper environment settings. | Relevant Papers ------------------------------------ - MAESTRO Cost Model - `MICRO 2019 `_ - Marvel - `Arxiv `_ - ConfuciuX - `MICRO 2020 `_ - GAMMA - `ICCAD 2020 `_