Quantum Computing and Simulation Platform for HEP at IHEP

Yujiang Bi, Lu Wang, Wei Sun, Qingbao Hu2022 · DOI

Abstract

With the dramatic growth of experimental data in high-energy physics (HEP) and the increasing demand for accuracy in physical analysis, classical computing models are reaching their limits. Quantum computers offer a potential solution to these bottlenecks, yet current hardware remains in the Noisy Intermediate-Scale Quantum (NISQ) era, characterized by few qubits and high error rates. To bridge this gap, the Institute of High Energy Physics (IHEP) has developed a full-stack development and simulation platform. By leveraging classical high-performance computing (HPC) to simulate quantum systems, the platform provides a stable environment for researchers to develop quantum algorithms for HEP applications, such as Lattice QCD and event classification.

Introduction

High Energy Physics is facing a significant data crisis. Experiments at the Large Hadron Collider (LHC) and the proposed Circular Electron Positron Collider (CEPC) generate volumes of data that outstrip classical processing capabilities. Quantum computing (QC) theoretically addresses these challenges, particularly for NP-hard problems like the sign problem in Lattice Quantum Chromodynamics (Lattice QCD). However, because physical quantum hardware is currently unreliable, classical simulators are essential for algorithm development, providing "perfect" qubits for logic testing and controlled noise modeling.

System Architecture and Methods

The IHEP Quantum Computing and Simulation Platform is designed as a four-layer stack to provide a seamless transition from high-level applications to heterogeneous hardware backends.

The four-layer architecture of the IHEP quantum simulation platform, spanning from HEP applications to classical hardware backends.
  1. Application Layer: Targets HEP-specific tasks including noise filtering for LHAASO and track recognition for BESIII/CEPC.
  2. Framework Layer: Offers a user-friendly interface featuring a Jupyter-based environment and a drag-and-drop GUI to lower the entry barrier for physicists.
  3. Computing Engine: Integrates open-source SDKs like Qiskit, Cirq, and QPanda, handling circuit simulation, compilation, and optimization.
  4. Hardware Platform: A heterogeneous backend utilizing IHEP’s classical resources, including NVIDIA Tesla V100 GPUs and ARM-based Kunpeng 920 clusters.

To maximize efficiency, the team is optimizing simulation workloads for GPUs using CUDA and AMD’s ROCm, and implementing vectorization on ARM processors via NEON instructions.

Performance and Scalability

A primary challenge in quantum simulation is the exponential growth of memory requirements. Simulating a 36-qubit system using the state vector (SV) method requires approximately 1 TB of memory. As the number of qubits increases, the memory demand quickly hits the "50-qubit wall," where even the world's most powerful supercomputers struggle to maintain a full state vector.

Comparison of memory requirements for state vector (SV) and tensor network contraction (TNC) methods across various supercomputing platforms.

Currently, IHEP's interactive cluster supports real-time simulations of up to 20 qubits. More complex circuits are offloaded to an HPC cluster equipped with V100 GPUs. The platform is designed to be hardware-agnostic, with ongoing work to port Qiskit to AMD GPUs and ARM architectures to ensure flexibility as classical hardware evolves.

Future Directions

The platform is currently live, allowing users to deploy containerized simulation environments via Kubernetes. Future development will focus on:

  • Lattice QCD Simulation: Moving beyond simple algorithms to simulate complex fermion and gluon fields.
  • Advanced Noise Modeling: Implementing sophisticated noise simulations to better reflect the behavior of real NISQ hardware.
  • Public Education: Launching a web portal dedicated to public training and education in quantum computing for the physics community.

Summary updated

Abstract

With the dramatic growth of experimental data in high-energy physics and the increasing demand for accuracy of physical analysis results, the current classical computing model and computing power can no longer meet the needs of data processing and analysis. As a new computing device with great potential, quantum computers have become one of the possible ways to solve the bottleneck of high-energy physics data processing and computation, though quantum computers are still far away from general computing. At such a stage that quantum computers have so many shortcomings like few physical qubits, short coherent time, and poor fidelity to solve most of practical problems, it's necessary to build a development and simulation platform using classical computers to accelerate research on quantum algorithms and applications. We've set up an interactive development platform for quantum algorithms research and quantum computing simulation, and are preparing one web portal for public education and training on quantum computing. In the future, we will pay more attention to quantum computing applications in HEP and QCD simulation.
BibTeX
@misc{bi2022quantum,
  author = {Yujiang Bi and Lu Wang and Wei Sun and Qingbao Hu},
  title = {{Quantum Computing and Simulation Platform for HEP at IHEP}},
  year = {2022},
  doi = {10.22323/1.415.0002},
}
Figure