Loading...
Search
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
Digital Object List
Bookmark
Review List
Fast and Scalable Quantum Circuit Simulation on Cluster of Multi-Core and Many-Core Platforms
Ahmadzadeh Chaghooshi, Armin | 2025
110
Viewed
- Type of Document: Ph.D. Dissertation
- Language: English
- Document No: 58499 (52)
- University: Sharif University of Technology, International Campus, Kish Island
- Department: Science and Engineering
- Advisor(s): Sarbazi Azad, Hamid
- Abstract:
- Quantum computing is an emerging and promising computational paradigm that pro- vides substantial speedup for important applications such as integer factorization, database search, and machine learning. Since practical quantum computers remain in their infancy, researchers must rely on classical machines to model quantum behav- ior and design algorithms—simulations that impose exponentially escalating demands on both memory and processing power. In this study, we propose a novel method for distributing the computation load (based on Dynamic Load Partitioning (DLP)) of the simulation among CPUs and GPUs in a hybrid computer system to reduce the required memory and computation time. Our approach employs a hybrid platform that simulates the quantum circuit in two phases using parallel array-based and recur- sive methods. Our experimental results demonstrate significant speedups over existing implementations, achieving a 96X speed-up over a recursive implementation in a GPU and a 12.9X speed-up over the state-of-the-art parallel implementation on a multi- node cluster. Moreover, our approach is significantly more energy efficient than the state-of-the-art method by 55X. While the state-of-the-art simulator necessitates over 2048 compute nodes to effectively simulate 40 qubits, our method achieves the same simulation using a single GPU, drastically reducing equipment costs. Furthermore, the diverse CPU and GPU simulation configurations, encompass- ing various combinations and parameters such as qubit size, memory capacity, circuit depth, GPU performance, resource heterogeneity, and load imbalance pose additional challenges. Addressing these complexities requires an exhaustive exploration of the design space, which is impractical within a reasonable timeframe. Therefore, given the multitude of parameters and the analysis of influential factors, having an analytical model for selecting the proper configuration is desirable and even essential for large systems. To do so, we propose a novel analytical performance model for quantum circuit simulation on a hybrid CPU-GPU platform with arbitrary sizes and parame- ters. The model analyzes the execution time of individual GPU kernels and the impact of major micro-architecture features on overall performance. By employing DLP and the heterogeneous multi-GPU kernel, performance bottlenecks are accurately identi- fied, and execution time is estimated. The proposed model can be employed to pro- vide insights into scalability, efficiency, and load balancing in hybrid parallel systems, supporting code optimization and development of efficient quantum algorithms and advanced quantum circuit simulation on hybrid parallel architectures
- Keywords:
- High Performance Computing ; Quantum Computation ; Simulation ; Distributed Simulation ; Analytical Modeling ; Hybrid Central Processing Unit-Graphics Procssing Unit (CPU-GPU)Systems
محتواي کتاب