Quantum Computing and High Performance Computing
Abstract
GE Global Research has enhanced a previously developed general-purpose quantum computer simulator, improving its efficiency and increasing its functionality. Matrix multiplication operations in the simulator were optimized by taking advantage of the particular structure of the matrices, significantly reducing the number of operations and memory overhead. The remaining operations were then distributed over a cluster, allowing feasible compute times for large quantum systems. The simulator was augmented to evaluate a step-by-step comparison of a quantum algorithm's ideal execution to its real-world performance, including errors. To facilitate the study of error propagation in a quantum system, the simulator s graphical user interface was enhanced to visualize the differences at each step in the algorithm s execution. To verify the simulator s accuracy, three ion trap-based experiments were simulated. The simulator output closely matches experimentalist s results, indicating that the simulator can accurately model such devices. Finally, alternative hardware platforms were researched to further improve the simulator performance. An FPGA-based accelerator was designed and simulated, resulting in substantial performance improvements over the original simulator. Together, this research produced a highly efficient quantum computer simulator capable of accurately modeling arbitrary algorithms on any hardware device.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2006
- Accession Number
- ADA462065
Entities
People
- Brent H. Allen
- Joseph Shultz
- Kareem S. Aggour
- Michael Lapinski
- Robert M. Mattheyses
Organizations
- General Electric