High-Performance integrated Quantum Computing (HPQC)
|
Quantum computers (QCs) promise to provide a platform that may be one day employed to tackle challenges such as the development of new materials (e.g., high-temperature superconductors) or more efficient solutions for simulations relevant in quantum chemistry and molecular biology. However, QCs are currently specialised processors that need to be programmed by highly-trained individuals. Only recently quantum programming languages have started to facilitate wider access to quantum computers. Yet, for end users interested in using quantum devices, no frameworks exist that would allow non-specialists to implement solutions to their own use cases for novel quantum computing hardware. Currently, these use cases are addressed by highperformance computing (HPC) facilities. As such, it is essential to develop novel solutions for integrating QCs directly with HPC systems. Yet there are currently no well defined interfaces between HPC and QC systems, neither frameworks that could build on top of a hybrid infrastructure, nor benchmarks to capture the benefits that QCs offer. Within this project the consortium will integrate, for the first time in Europe, a quantum computer directly into an Austrian HPC facility.
Building on top of existing QC and HPC infrastructure in Innsbruck, the consortium is developing a dedicated quantum acceleration ecosystem connecting QC and HPC nodes such that newly developed quantumaware code acceleration frameworks can execute proof-of-concept use cases on the hybrid infrastructure. The consortium will benchmark the quality of the code acceleration, framework, libraries, and the holistic hybrid QC/HPC infrastructure via the implementation and evaluation of key applications in the fields of QC and HPC. Additionally, improvements to the QC hardware aim to achieve performance levels at the edge for QC to provide a computational advantage over HPC systems. The consortium will realise the first fully integrated, hybrid QC/HPC infrastructure in Europe. The developed interfaces will facilitate and increase the speed of development of quantum computers, while the connection towards HPC systems will allow to train computer scientists for the first time directly on quantum-aware algorithms and use cases using quantum hardware. To the best of our knowledge, we will develop first fundamental methods for the systematic performance benchmarking of traditional and quantum-accelerated HPC code. The knowledge gained can be used later as a building block for the development of high-level languages for the hybrid Quantum/HPC programming. We aim to provide a bridge for the connection of quantum research to computer science, and subsequently strengthen the Austrian and European quantum eco-system. In this context I am leading the efforts in developing reinforcement-learning algorithms [building on prior work in Quantum 3, 215 (2019)] that can make use of the HPC infrastructure to provide real-time decoding and code switching [based on previous theoretical and experimental work in Nat. Commun. 8, 1321 (2017) and Nature 589, 220–224 (2021), repectively] for dynamical adaption to noise in the quantum hardware.
Building on top of existing QC and HPC infrastructure in Innsbruck, the consortium is developing a dedicated quantum acceleration ecosystem connecting QC and HPC nodes such that newly developed quantumaware code acceleration frameworks can execute proof-of-concept use cases on the hybrid infrastructure. The consortium will benchmark the quality of the code acceleration, framework, libraries, and the holistic hybrid QC/HPC infrastructure via the implementation and evaluation of key applications in the fields of QC and HPC. Additionally, improvements to the QC hardware aim to achieve performance levels at the edge for QC to provide a computational advantage over HPC systems. The consortium will realise the first fully integrated, hybrid QC/HPC infrastructure in Europe. The developed interfaces will facilitate and increase the speed of development of quantum computers, while the connection towards HPC systems will allow to train computer scientists for the first time directly on quantum-aware algorithms and use cases using quantum hardware. To the best of our knowledge, we will develop first fundamental methods for the systematic performance benchmarking of traditional and quantum-accelerated HPC code. The knowledge gained can be used later as a building block for the development of high-level languages for the hybrid Quantum/HPC programming. We aim to provide a bridge for the connection of quantum research to computer science, and subsequently strengthen the Austrian and European quantum eco-system. In this context I am leading the efforts in developing reinforcement-learning algorithms [building on prior work in Quantum 3, 215 (2019)] that can make use of the HPC infrastructure to provide real-time decoding and code switching [based on previous theoretical and experimental work in Nat. Commun. 8, 1321 (2017) and Nature 589, 220–224 (2021), repectively] for dynamical adaption to noise in the quantum hardware.
Consortium
The project is being carried out by a consortium led by Thomas Monz and Martin Ringbauer at the Institute for Experimental Physics of the University of Innsbruck; and, in addition to my team at TU Wien, further includes the academic partners Alexander Ostermann (Research Area Scientific Computing, University of Innsbruck), Richard Küng (Institute for Integrated Circuits, Johannes Kepler University Linz), and Ivona Brandic and Vincenzo De Maio (Institute for Information Systems Engineering, TU Wien), as well as industry partners such as Alpine Quantum Technologies (AQT) and Math. Tec.
The project is being carried out by a consortium led by Thomas Monz and Martin Ringbauer at the Institute for Experimental Physics of the University of Innsbruck; and, in addition to my team at TU Wien, further includes the academic partners Alexander Ostermann (Research Area Scientific Computing, University of Innsbruck), Richard Küng (Institute for Integrated Circuits, Johannes Kepler University Linz), and Ivona Brandic and Vincenzo De Maio (Institute for Information Systems Engineering, TU Wien), as well as industry partners such as Alpine Quantum Technologies (AQT) and Math. Tec.
This project is running from December 31, 2022 until December 30, 2025.
Team at Atominstitut - TU Wien
- Nicolai Friis (PI/Project Lead)
- Phila Rembold (Postdoc)
- Tomasz Andrzejewski (PhD student)