arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1908.03072cs.LGcs.ARcs.DCcs.NE
收藏

TensorDIMM:一种实用的近存储器处理架构,用于深度学习中的嵌入和张量运算

TensorDIMM: A Practical Near-Memory Processing Architecture for Embeddings and Tensor Operations in Deep Learning

KAIST(韩国科学技术院)

Youngeun Kwon, Yunjae Lee, Minsoo Rhu

摘要:

最近几个超标量的研究指出,嵌入层是当今计算机中部署的内存最密集的深度学习(DL)算法。本文讨论了嵌入层和相关张量操作的内存容量和带宽挑战。我们提出了我们的垂直集成的硬件/软件协同设计,其中包括一个定制的DIMM模块,增强了为DL张量操作量身定制的近数据处理内核。这些自定义DIRECT作为远程内存池填充在以GPU为中心的系统互连内部,允许GPU利用可扩展的内存带宽和容量扩展。我们的建议在真正的DL系统上的原型实现显示出平均6.2- 17.6倍的性能提高最先进的推荐系统。

英文摘要:

Recent studies from several hyperscalars pinpoint to embedding layers as the most memory-intensive deep learning (DL) algorithm being deployed in today's datacenters. This paper addresses the memory capacity and bandwidth challenges of embedding layers and the associated tensor operations. We present our vertically integrated hardware/software co-design, which includes a custom DIMM module enhanced with near-data processing cores tailored for DL tensor operations. These custom DIMMs are populated inside a GPU-centric system interconnect as a remote memory pool, allowing GPUs to utilize for scalable memory bandwidth and capacity expansion. A prototype implementation of our proposal on real DL systems shows an average 6.2-17.6x performance improvement on state-of-the-art recommender systems.

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。

↑