Similarity Search with Tensor Core Units

被引:5
作者
Ahle, Thomas D. [1 ,2 ]
Silvestri, Francesco [3 ]
机构
[1] IT Univ, Copenhagen, Denmark
[2] BARC, Copenhagen, Denmark
[3] Univ Padua, Padua, Italy
来源
SIMILARITY SEARCH AND APPLICATIONS, SISAP 2020 | 2020年 / 12440卷
关键词
Similarity search; Tensor core units; Dimensionality reduction; Similarity join; Locality sensitive hashing;
D O I
10.1007/978-3-030-60936-8_6
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Tensor Core Units (TCUs) are hardware accelerators developed for deep neural networks, which efficiently support the multiplication of two dense root m x root m matrices, where m is a given hardware parameter. In this paper, we show that TCUs can speed up similarity search problems as well. We propose algorithms for the Johnson-Lindenstrauss dimensionality reduction and for similarity join that, by leveraging TCUs, achieve a Omega(root m) speedup up with respect to traditional approaches.
引用
收藏
页码:76 / 84
页数:9
相关论文
共 16 条
[1]  
Ahle T.D., 2019, ARXIV190901821
[2]  
Ahle T.D., 2020, ARXIV200612608
[3]  
Ahle T.D., 2020, P 1 13 INT C SIM SEA
[4]  
Ahle TD, 2020, PROCEEDINGS OF THE THIRTY-FIRST ANNUAL ACM-SIAM SYMPOSIUM ON DISCRETE ALGORITHMS (SODA'20), P141
[5]   On the Complexity of Inner Product Similarity Join [J].
Ahle, Thomas D. ;
Pagh, Rasmus ;
Razenshteyn, Ilya ;
Silvestri, Francesco .
PODS'16: PROCEEDINGS OF THE 35TH ACM SIGMOD-SIGACT-SIGAI SYMPOSIUM ON PRINCIPLES OF DATABASE SYSTEMS, 2016, :151-164
[6]  
Ailon N., 2006, STOC'06. Proceedings of the 38th Annual ACM Symposium on Theory of Computing, P557, DOI 10.1145/1132516.1132597
[7]   Fast Dimension Reduction Using Rademacher Series on Dual BCH Codes [J].
Ailon, Nir ;
Liberty, Edo .
DISCRETE & COMPUTATIONAL GEOMETRY, 2009, 42 (04) :615-630
[8]   Limits on the Universal Method for Matrix Multiplication [J].
Alman, Josh .
34TH COMPUTATIONAL COMPLEXITY CONFERENCE (CCC 2019), 2019, 137
[9]   Brief Announcement: A Computational Model for Tensor Core Units [J].
Chowdhury, Rezaul ;
Silvestri, Francesco ;
Vella, Flavio .
PROCEEDINGS OF THE 32ND ACM SYMPOSIUM ON PARALLELISM IN ALGORITHMS AND ARCHITECTURES (SPAA '20), 2020, :519-521
[10]  
Dakkak A., 2019, P INT C SUP ICS