Albireo: Energy-Efficient Acceleration of Convolutional Neural Networks via Silicon Photonics

计算机科学 光子学 高效能源利用 多路复用 可扩展性 多播 硅光子学 能源消耗 吞吐量 计算机体系结构 电子工程 计算机网络 电信 电气工程 无线 物理 光电子学 工程类 数据库
作者
Kyle Shiflett,Avinash Karanth,Răzvan Bunescu,Ahmed Louri
出处
期刊: 卷期号:: 860-873 被引量:53
标识
DOI:10.1109/isca52012.2021.00072
摘要

With the end of Dennard scaling, highly-parallel and specialized hardware accelerators have been proposed to improve the throughput and energy-efficiency of deep neural network (DNN) models for various applications. However, collective data movement primitives such as multicast and broadcast that are required for multiply-and-accumulate (MAC) computation in DNN models are expensive, and require excessive energy and latency when implemented with electrical networks. This consequently limits the scalability and performance of electronic hardware accelerators. Emerging technology such as silicon photonics can inherently provide efficient implementation of multicast and broadcast operations, making photonics more amenable to exploit parallelism within DNN models. Moreover, when coupled with other unique features such as low energy consumption, high channel capacity with wavelength-division multiplexing (WDM), and high speed, silicon photonics could potentially provide a viable technology for scaling DNN acceleration.In this paper, we propose Albireo, an analog photonic architecture for scaling DNN acceleration. By characterizing photonic devices such as microring resonators (MRRs) and Mach-Zehnder modulators (MZM) using photonic simulators, we develop realistic device models and outline their capability for system level acceleration. Using the device models, we develop an efficient broadcast combined with multicast data distribution by leveraging parameter sharing through unique WDM dot product processing. We evaluate the energy and throughput performance of Albireo on DNN models such as ResNet18, MobileNet and VGG16. When compared to cur-rent state-of-the-art electronic accelerators, Albireo increases throughput by 110 X, and improves energy-delay product (EDP) by an average of 74 X with current photonic devices. Furthermore, by considering moderate and aggressive photonic scaling, the proposed Albireo design shows that EDP can be reduced by at least 229 X.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
刚刚
123发布了新的文献求助10
刚刚
成就煎饼发布了新的文献求助10
刚刚
余香的应助被温暖马里奥采纳,获得10
刚刚
充电宝的应助被风中的宛白采纳,获得10
刚刚
1秒前
科研中发布了新的文献求助10
1秒前
英姑的应助被gxh采纳,获得10
1秒前
小嘿嘿完成签到,获得积分10
1秒前
王少雄发布了新的文献求助10
1秒前
2秒前
小明同学发布了新的文献求助10
2秒前
张冠黄戴发布了新的文献求助10
2秒前
3秒前
令狐惜海的应助被刘一一采纳,获得20
3秒前
干饭啦发布了新的文献求助10
3秒前
3秒前
4秒前
4秒前
淡淡的忆彤完成签到,获得积分10
5秒前
5秒前
null的应助被自觉的亦竹采纳,获得10
6秒前
大白菜心完成签到,获得积分10
6秒前
6秒前
搜集达人的应助被h11采纳,获得10
6秒前
Hello的应助被小小蚂蚁采纳,获得10
6秒前
FashionBoy的应助被成就煎饼采纳,获得10
6秒前
照清完成签到 ,获得积分10
7秒前
阔达的阑香完成签到 ,获得积分20
7秒前
呵呵哒完成签到,获得积分10
7秒前
7秒前
7秒前
8秒前
123发布了新的文献求助10
8秒前
舒心无剑完成签到,获得积分10
8秒前
无极微光的应助被孙禹采纳,获得20
9秒前
小狗发布了新的文献求助10
9秒前
wg发布了新的文献求助10
9秒前
呵呵哒发布了新的文献求助10
9秒前
陌小石完成签到 ,获得积分0
9秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
John McDowells Bildungstheorie 800
八旗制度の研究 600
What Else Can I Tell You?: A Pragmatic Study of English Rhetorical Questions as Discursive and Argumentative Acts 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8163746
求助须知:如何正确求助?哪些是违规求助? 9557874
关于积分的说明 21742203
捐赠科研通 7754407
什么是DOI,文献DOI怎么找? 3428345
关于科研通互助平台的介绍 2553000
邀请新用户注册赠送积分活动 2472945