arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2010.05037cs.ARcs.DCcs.IR
收藏

深度推荐系统的跨栈可重用性分析

Cross-Stack Workload Characterization of Deep Recommendation Systems

Samuel Hsia, Udit Gupta, Mark Wilkening, Carole-Jean Wu, Gu-Yeon Wei, David Brooks

摘要:

基于深度学习的推荐系统构成了大多数个性化云服务的支柱。虽然计算机体系结构社区最近开始注意到深度推荐推理,但由此产生的解决方案采用了截然不同的方法-从近内存处理到大规模优化。为了更好地设计用于深度推荐推理的未来硬件系统,我们必须首先系统地检查和表征跨执行堆栈的不同级别的设计决策的底层系统级影响。在本文中,我们在执行堆栈的三个不同层次上描述了八个具有行业代表性的深度推荐模型:算法和软件,系统平台和硬件微架构。通过这种跨栈表征,我们首先表明系统部署选择(即,CPU或GPU,批量大小粒度)可以为我们提供高达15倍的加速。为了更好地理解进一步优化的瓶颈,我们研究了软件操作员的使用分解以及CPU前端和后端微架构的低效率。最后,我们对关键算法模型架构特征与硬件瓶颈之间的相关性进行建模,揭示了每个硬件瓶颈背后没有一个占主导地位的算法组件。

英文摘要:

Deep learning based recommendation systems form the backbone of most personalized cloud services. Though the computer architecture community has recently started to take notice of deep recommendation inference, the resulting solutions have taken wildly different approaches - ranging from near memory processing to at-scale optimizations. To better design future hardware systems for deep recommendation inference, we must first systematically examine and characterize the underlying systems-level impact of design decisions across the different levels of the execution stack. In this paper, we characterize eight industry-representative deep recommendation models at three different levels of the execution stack: algorithms and software, systems platforms, and hardware microarchitectures. Through this cross-stack characterization, we first show that system deployment choices (i.e., CPUs or GPUs, batch size granularity) can give us up to 15x speedup. To better understand the bottlenecks for further optimization, we look at both software operator usage breakdown and CPU frontend and backend microarchitectural inefficiencies. Finally, we model the correlation between key algorithmic model architecture features and hardware bottlenecks, revealing the absence of a single dominant algorithmic component behind each hardware bottleneck.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑