Efficient and accurate Word2Vec implementations in GPU and shared-memory multicore architectures

Trevor M. Simonton, G. Alaghband
{"title":"Efficient and accurate Word2Vec implementations in GPU and shared-memory multicore architectures","authors":"Trevor M. Simonton, G. Alaghband","doi":"10.1109/HPEC.2017.8091076","DOIUrl":null,"url":null,"abstract":"Word2Vec is a popular set of machine learning algorithms that use a neural network to generate dense vector representations of words. These vectors have proven to be useful in a variety of machine learning tasks. In this work, we propose new methods to increase the speed of the Word2Vec skip gram with hierarchical softmax architecture on multi-core shared memory CPU systems, and on modern NVIDIA GPUs with CUDA. We accomplish this on multi-core CPUs by batching training operations to increase thread locality and to reduce accesses to shared memory. We then propose new heterogeneous NVIDIA GPU CUDA implementations of both the skip gram hierarchical softmax and negative sampling techniques that utilize shared memory registers and in-warp shuffle operations for maximized performance. Our GPU skip gram with negative sampling approach produces a higher quality of word vectors than previous GPU implementations, and our flexible skip gram with hierarchical softmax implementation achieves a factor of 10 speedup of the existing methods.","PeriodicalId":364903,"journal":{"name":"2017 IEEE High Performance Extreme Computing Conference (HPEC)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2017-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"8","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2017 IEEE High Performance Extreme Computing Conference (HPEC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/HPEC.2017.8091076","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 8

Abstract

Word2Vec is a popular set of machine learning algorithms that use a neural network to generate dense vector representations of words. These vectors have proven to be useful in a variety of machine learning tasks. In this work, we propose new methods to increase the speed of the Word2Vec skip gram with hierarchical softmax architecture on multi-core shared memory CPU systems, and on modern NVIDIA GPUs with CUDA. We accomplish this on multi-core CPUs by batching training operations to increase thread locality and to reduce accesses to shared memory. We then propose new heterogeneous NVIDIA GPU CUDA implementations of both the skip gram hierarchical softmax and negative sampling techniques that utilize shared memory registers and in-warp shuffle operations for maximized performance. Our GPU skip gram with negative sampling approach produces a higher quality of word vectors than previous GPU implementations, and our flexible skip gram with hierarchical softmax implementation achieves a factor of 10 speedup of the existing methods.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
在GPU和共享内存多核架构中实现高效准确的Word2Vec
Word2Vec是一套流行的机器学习算法,它使用神经网络生成单词的密集向量表示。这些向量已被证明在各种机器学习任务中是有用的。在这项工作中,我们提出了新的方法来提高Word2Vec跳过gram的速度与分层softmax架构在多核共享内存CPU系统和现代NVIDIA gpu CUDA。我们通过批处理训练操作在多核cpu上实现这一点,以增加线程局部性并减少对共享内存的访问。然后,我们提出了新的异构NVIDIA GPU CUDA实现,包括跳过克分层softmax和负采样技术,这些技术利用共享内存寄存器和warp shuffle操作来最大化性能。我们采用负采样方法的GPU跳过克比以前的GPU实现产生更高质量的词向量,并且我们采用分层softmax实现的灵活跳过克比现有方法实现了10倍的加速。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
Optimized task graph mapping on a many-core neuromorphic supercomputer Software-defined extreme scale networks for bigdata applications Power-aware computing: Measurement, control, and performance analysis for Intel Xeon Phi xDCI, a data science cyberinfrastructure for interdisciplinary research Leakage energy reduction for hard real-time caches
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1