High-Performance Tensor Network Library for Accelerating Quantum Science via cuQuantum SDK
ORAL
Abstract
Tensor network methods have emerged as a powerful foundation and effective research tool for exploring the fundamental properties of quantum systems. The availability of a fast, scalable software is vital for quantum algorithm researchers, as well as for quantum hardware engineers.
NVIDIA’s cuQuantum SDK, cuTensorNet component, is designed to accelerate and scale simulations developed by the quantum science community. cuTensorNet enables researchers to leverage highly efficient, GPU-optimized software components tailored for NVIDIA platforms. Our goal is to accelerate and augment the potential of quantum science to enable new discoveries and deepen our understanding of complex physical systems.
cuTensorNet provides comprehensive support for tensor network based methods, including approximate algorithms and projection methods based on matrix product states as well as differentiation techniques and other factorizable tensor representations. All capabilities are accessible through both Python and C APIs, ensuring flexibility and ease of integration.
NVIDIA’s cuQuantum SDK, cuTensorNet component, is designed to accelerate and scale simulations developed by the quantum science community. cuTensorNet enables researchers to leverage highly efficient, GPU-optimized software components tailored for NVIDIA platforms. Our goal is to accelerate and augment the potential of quantum science to enable new discoveries and deepen our understanding of complex physical systems.
cuTensorNet provides comprehensive support for tensor network based methods, including approximate algorithms and projection methods based on matrix product states as well as differentiation techniques and other factorizable tensor representations. All capabilities are accessible through both Python and C APIs, ensuring flexibility and ease of integration.
*NVIDIA
–
Publication: https://arxiv.org/abs/2308.01999
Presenters
-
Azzam Haidar
- NVIDIA Corporation