Similarity Search with Tensor Core Units
Similarity Search with Tensor Core Units
复制标题
DOI：
10.1007/978-3-030-60936-8_6
复制
发表时间：
2020-06
期刊：
影响因子：
--
通讯作者：
Thomas Dybdahl Ahle;Francesco Silvestri
中科院分区：
文献类型：
--
作者：
Thomas Dybdahl Ahle;Francesco Silvestri
张量核心单元（tcu）是为深度神经网络开发的硬件加速器，它有效地支持两个密度矩阵的乘法，其中是给定的硬件参数。在本文中，我们证明了tcu也可以加速相似搜索问题。我们提出了Johnson-Lindenstrauss降维算法和相似连接算法，通过利用tcu，实现了相对于传统方法的加速。
Tensor Core Units (TCUs) are hardware accelerators developed for deep neural networks, which efficiently support the multiplication of two densematrices, wheremis 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 aspeedup up with respect to traditional approaches.