Toward Complete Emulation of Quantum Algorithms Using High-Performance Reconfigurable Computing
Advisor
Event Type
Doctoral Showcase
Posters
TP
TimeThursday, 18 November 202111:30am - 12pm CST
Location222
DescriptionQuantum computing, at its current nascent stage, has many critical problems that require investigation. For instance, solving real-world problems require quantum circuits with large depth which are difficult, often impossible to implement on a quantum computer due to decoherence noise. More specifically, developing I/O-intensive quantum applications require large amount of classical data to be transferred to the quantum domain. Performing classical-to-quantum (C2Q) data-encoding requires deep, complex circuits. Current quantum devices have low qubit counts and often produce faulty or noisy measurements, which results in heavy dependency on costly, resource-hungry simulators. This work proposes a cost-effective, hardware-based emulation methodology for quantum computation, using Field Programmable Gate Arrays (FPGA) technology. The proposed methodology enables emulation of quantum computation with high-scalability, high-accuracy, and high-throughput, compared to existing FPGA-based emulators. The emulation framework includes C2Q data-encoding circuit models that are highly optimized, compared to existing methods. The proposed emulation methodology allows investigating optimizations for several important quantum algorithms such as Quantum Fourier Transform (QFT), Quantum Grover’s Search (QGS), and Quantum Haar Transform (QHT). In addition, it facilitated developing novel extensions to conventional QGS to enable search for dynamically changing data, as well as developing decoherence-optimized circuits for QHT. Furthermore, a novel technique is presented combining QHT and dynamic, multi-pattern QGS to perform dimension reduction for pattern matching and recognition on high-resolution, spatio-spectral data. Emulation architectures are designed and implemented on a high-performance reconfigurable computer (HPRC). As a case study, quantum image processing using multi-spectral images is investigated in experiments.
Archive view