Lego: Dynamic Tensor-Splitting Multi-Tenant DNN Models on Multi-Chip-Module Architecture

Zhou Yu Xuan, Ching Jui Lee, Tsung Tai Yeh

研究成果: Conference contribution同行評審

摘要

Modern deep neural network (DNN) accelerators target the acceleration of a single DNN model and limit the throughput for multi-tenant DNN data center applications. The multi-chip-module (MCM) architecture breaks a monolithic accelerator into multiple small chiplets. The MCM is a promising approach that dispatches DNN models across chiplets with equal PEs. However, it is challenging to distribute data of DNN model layers with different parameters across chiplets while maximizing the chiplet utilization. This work proposes Lego MCM architecture that dynamically adapts to the size of DNN model layers and improves the throughput of multi-tenant DNN applications by increasing the chiplet utilization. Lego's dynamic scheduler achieves the geometric average 1.51× speedup over a monolithic DNN accelerator.

原文English
主出版物標題Proceedings - International SoC Design Conference 2022, ISOCC 2022
發行者Institute of Electrical and Electronics Engineers Inc.
頁面173-174
頁數2
ISBN(電子)9781665459716
DOIs
出版狀態Published - 2022
事件19th International System-on-Chip Design Conference, ISOCC 2022 - Gangneung-si, 韓國
持續時間: 19 10月 202222 10月 2022

出版系列

名字Proceedings - International SoC Design Conference 2022, ISOCC 2022

Conference

Conference19th International System-on-Chip Design Conference, ISOCC 2022
國家/地區韓國
城市Gangneung-si
期間19/10/2222/10/22

指紋

深入研究「Lego: Dynamic Tensor-Splitting Multi-Tenant DNN Models on Multi-Chip-Module Architecture」主題。共同形成了獨特的指紋。

引用此