cuTensorNet: A High-Performance Library for Tensor Network Computations

Welcome to the cuTensorNet library documentation!

NVIDIA cuTensorNet is a high-performance library for tensor network computations, a component of the NVIDIA cuQuantum SDK.

Download: https://developer.nvidia.com/cuQuantum-downloads

Key Features

• Based on NVIDIA’s high-performance tensor linear algebra library: cuTENSOR

• Provides APIs for:

  • Creating a tensor network

  • Finding a low-cost path for any given tensor network contraction

  • Finding a low-cost slicing to meet specified memory constraints

  • Tuning the path finder’s configuration for better performance

  • Performing tensor network contractions

  • Logging

Support

• Supported GPU Architectures: Volta and Ampere

• Supported OSes: Linux

• Supported CPU Architectures: x86_64, ARM64, ppc64le

Prerequisites

• CUDA Toolkit v11.4 (or above) and compatible driver (see CUDA Driver Release Notes).

• cuTENSOR v1.4.0 (or above)

• NVIDIA HPC SDK v21.11 (or above)

Contents

• Release Notes
  • cuTensorNet v0.1.0
  • cuTensorNet v0.0.1
• Overview
  • Introduction to tensor networks
  • Contraction pathfinder
  • Hyper-optimizer
  • Supported data types
  • Reference
• Getting Started
  • Installation and Compilation
  • Code Example
• API Reference
  • cuTensorNet Data Types
  • cuTensorNet Functions
• Acknowledgements