Gemmini：通过全栈集成实现系统化的深度学习架构评估
分布式、并行与集群计算
2021-07-12 v3 硬件体系结构
机器学习
性能
摘要
DNN 加速器通常在孤立状态下开发和评估，未考虑真实世界环境中跨栈、系统级的影响。这使得人们难以认识片上系统（SoC）资源争用、操作系统开销以及编程栈低效对整体性能/能效的影响。为应对这一挑战，我们提出 Gemmini，一个开源*、全栈 DNN 加速器生成器。Gemmini 从灵活的架构模板生成广泛设计空间的高效 ASIC 加速器，同时生成灵活的编程栈以及带有共享资源、能捕捉系统级效应的完整 SoC。Gemmini 生成的加速器也已流片，在各种 DNN 基准测试上相较高性能 CPU 实现了高达三个数量级的加速。* https://github.com/ucb-bar/gemmini
引用
@article{arxiv.1911.09925,
title = {Gemmini: Enabling Systematic Deep-Learning Architecture Evaluation via Full-Stack Integration},
author = {Hasan Genc and Seah Kim and Alon Amid and Ameer Haj-Ali and Vighnesh Iyer and Pranav Prakash and Jerry Zhao and Daniel Grubb and Harrison Liew and Howard Mao and Albert Ou and Colin Schmidt and Samuel Steffl and John Wright and Ion Stoica and Jonathan Ragan-Kelley and Krste Asanovic and Borivoje Nikolic and Yakun Sophia Shao},
journal= {arXiv preprint arXiv:1911.09925},
year = {2021}
}
备注
To appear at the 58th IEEE/ACM Design Automation Conference (DAC), December 2021, San Francisco, CA, USA