Quantum Transpiler Optimization: On the Development, Implementation, and Use of a Quantum Research Testbed
Abstract
Quantum computing research is at the cusp of a paradigm shift. As the complexity of quantum systems increases, so does the complexity of research procedures for creating and testing layers of the quantum software stack. However, the tools used to perform these tasks have not experienced the increase in capability required to effectively handle the development burdens involved. This case is made particularly clear in the context of IBM QX Transpiler optimization algorithms and functions. IBM QX systems use the Qiskit library to create, transform, and execute quantum circuits. As coherence times and hardware qubit counts increase and qubit topologies become more complex, so does orchestration of qubit mapping and qubit state movement across these topologies. The transpiler framework used to create and test improved algorithms have not kept pace. A test bed is proposed to provide abstractions to create and test transpiler routines. The development process is analyzed and implemented, from design principles through requirements analysis and verification testing. Additionally, limitations of existing transpiler algorithms are identified and initial results are provided that suggest more effective algorithms for qubit mapping and state movement.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 26, 2020
- Accession Number
- AD1102933
Entities
People
- Brandon K Kamaka
Organizations
- Air Force Institute of Technology