TY - GEN
T1 - FlatDD
T2 - 53rd International Conference on Parallel Processing, ICPP 2024
AU - Jiang, Shui
AU - Fu, Rongliang
AU - Burgholzer, Lukas
AU - Wille, Robert
AU - Ho, Tsung Yi
AU - Huang, Tsung Wei
N1 - Publisher Copyright:
© 2024 Owner/Author.
PY - 2024/8/12
Y1 - 2024/8/12
N2 - Quantum circuit simulator (QCS) is essential for designing quantum algorithms because it assists researchers in understanding how quantum operations work without access to expensive quantum computers. Traditional array-based QCSs suffer from exponential time and memory complexities. To address this problem, Decision Diagram (DD) was introduced to compress simulation data by exploring the circuit regularity. However, for irregular circuit structures, DD-based simulation incurs significant runtime and memory overhead. To overcome this challenge, we present FlatDD, a high-performance QCS that capitalizes on the strength of both DD- and array-based approaches. FlatDD parallelizes the simulation workload at multiple levels and leverages an efficient caching technique to reuse historical results. To further enhance the simulation performance for deep circuits, FlatDD introduces a gate-fusion algorithm to reduce the computational cost. Compared to state-of-the-art QCSs on commonly used quantum circuits, FlatDD achieves 34.81 × speed-up and 1.93 × memory reduction.
AB - Quantum circuit simulator (QCS) is essential for designing quantum algorithms because it assists researchers in understanding how quantum operations work without access to expensive quantum computers. Traditional array-based QCSs suffer from exponential time and memory complexities. To address this problem, Decision Diagram (DD) was introduced to compress simulation data by exploring the circuit regularity. However, for irregular circuit structures, DD-based simulation incurs significant runtime and memory overhead. To overcome this challenge, we present FlatDD, a high-performance QCS that capitalizes on the strength of both DD- and array-based approaches. FlatDD parallelizes the simulation workload at multiple levels and leverages an efficient caching technique to reuse historical results. To further enhance the simulation performance for deep circuits, FlatDD introduces a gate-fusion algorithm to reduce the computational cost. Compared to state-of-the-art QCSs on commonly used quantum circuits, FlatDD achieves 34.81 × speed-up and 1.93 × memory reduction.
KW - Decision Diagram
KW - Exponentially Weighted Moving Average
KW - Quantum Circuit Simulation
UR - http://www.scopus.com/inward/record.url?scp=85202430974&partnerID=8YFLogxK
U2 - 10.1145/3673038.3673073
DO - 10.1145/3673038.3673073
M3 - Conference contribution
AN - SCOPUS:85202430974
T3 - ACM International Conference Proceeding Series
SP - 388
EP - 399
BT - 53rd International Conference on Parallel Processing, ICPP 2024 - Main Conference Proceedings
PB - Association for Computing Machinery
Y2 - 12 August 2024 through 15 August 2024
ER -