arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2512.23917quant-phcond-mat.str-elcs.LG
收藏

张量计算接口:一种面向应用、轻量级的便携式高性能张量网络应用接口

Tensor computing application programming interface: An application-oriented, lightweight API for portable high-performance tensor network computations

Department of Physics and Astronomy, California State University Northridge(加州州立大学北岭分校物理与天文学系);RIKEN Interdisciplinary Theoretical and Mathematical Sciences Program (iTHEMS)(理化学研究所交叉学科理论与数学科学项目);Computational Materials Science Research Team, RIKEN Center for Computational Science (R-CCS)(理化学研究所计算科学中心计算材料科学研究团队);Quantum Computational Science Research Team, RIKEN Center for Quantum Computing (RQC)(理化学研究所量子计算中心量子计算科学研究团队);Computational Condensed Matter Physics Laboratory, RIKEN Cluster for Pioneering Research (CPR)(理化学研究所先锋研究集群计算凝聚态物理实验室);Quantum Computing Simulation Unit, Quantum-HPC Hybrid Platform Division, RIKEN Center for Computational Science (R-CCS)(理化学研究所计算科学中心量子计算模拟单元);Next-Generation HPC Application Development Unit, Next-Generation HPC Infrastructure Development Division, RIKEN Center for Computational Science (R-CCS)(理化学研究所计算科学中心下一代HPC应用开发单元)

Rong-Yang Sun, Tomonori Shirakawa, Hidehiko Kohshiro, D. N. Sheng, Seiji Yunoki

AI总结:

本文提出张量计算接口(TCI),通过统一的类型系统和核心函数,实现跨框架的高性能张量网络应用,提升代码可移植性和性能

摘要:

张量网络(TNs)是量子科学和人工智能中的核心计算工具。然而,缺乏统一的软件接口在不同张量计算框架之间严重限制了张量网络应用的可移植性,将算法开发耦合到特定的硬件和软件后端。为了解决这一挑战,我们引入了张量计算接口(TCI)——一种面向应用、轻量级的应用程序编程接口,旨在实现框架无关、高性能的张量网络应用。TCI提供了一个明确的类型系统,抽象出张量对象,并提供了一组最小但表达力强的核心函数,涵盖基本的张量操作和张量线性代数运算。通过在代表性张量网络应用上的数值演示,我们展示出编写针对TCI的代码可以无缝迁移至异构硬件和软件平台,同时实现与原生框架实现相当的性能。我们进一步基于Cytnx发布了TCI的开源实现,证明了其实用性和与现有张量计算框架的集成便捷性。

英文摘要:

Tensor networks (TNs) are a central computational tool in quantum science and artificial intelligence. However, the lack of a unified software interface across tensor-computing frameworks severely limits the portability of TN applications, coupling algorithmic development to specific hardware and software back ends. To address this challenge, we introduce the Tensor Computing Application Programming Interface (TCAPI)---an application-oriented, lightweight API designed to enable framework-independent, high-performance TN applications, currently specified for \texttt{C++} and \texttt{Python}. TCAPI provides well-defined type-system conventions that abstract tensor objects, together with a minimal yet expressive set of core functions for essential tensor manipulations and tensor linear-algebra operations, with operation semantics shared across language specifications. Through numerical demonstrations of representative tensor-network applications implemented in \texttt{C++} using TCAPI, we show that code written against TCAPI can be migrated seamlessly across heterogeneous hardware and software platforms while achieving performance comparable to native framework implementations. We further provide open-source implementations of TCAPI based on \textit{Cytnx} and \textit{ITensor} in \texttt{C++}, and on \textit{NumPy} in \texttt{Python}, demonstrating its practicality and ease of integration with existing tensor-computing frameworks.

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。

↑