Title:	Numerical modeling of scaled-up superconducting qubit systems with GPUs
DNr:	Berzelius-2026-38
Project Type:	LiU Berzelius
Principal Investigator:	Jorge Fernández-Pendás <jorgefe@chalmers.se>
Affiliation:	Chalmers tekniska högskola
Duration:	2026-03-03 – 2026-10-01
Classification:	10302
Homepage:	https://www.chalmers.se/en/centres/wacqt/
Keywords:

Abstract

At Chalmers University of Technology, there's an important effort on the construction of a quantum computer using superconducting qubits, one of the most promising platforms for quantum technology. This represents the core project at the Wallenberg Centre for Quantum Technology (WACQT), a 12-year, billion-SEK initiative started in 2018 with the purpose of advancing Swedish academia and industry to the forefront of quantum technology. Our work inside this project involves theoretical modeling of the superconducting qubit systems, predicting their behavior and helping explain the experimental results. After demonstrating good control of few-qubit devices in several labs across the world, the field is now transitioning towards scaling up systems, meaning hundreds or even thousands of qubits. At Chalmers we are also focusing on increasing our qubit count. This presents challenges of varied nature. A particular one is that the computational cost of our simulations becomes problematic. In particular, memory becomes the most limiting factor due to the exponential growth of the Hilbert space as the number of qubits goes up. To address this, we employ so-called continuous unitary transformations. This mathematical approach transforms complex Hamiltonian models into more manageable forms, enabling easier computation of relevant observables. The transformations convert the original problem of diagonalizing matrices that are too large into that of solving systems of coupled ordinary differential equations (ODEs), which is a computationally intensive task but well-suited for parallel processing on GPUs. We have developed our own software solution. This tool allows us to efficiently handle the demanding computations required for scaled-up qubit systems. However, we are reaching its limit with the GPUs that we have currently access to. Since we are researchers funded by the Wallenberg foundation, we expect to be granted access to Berzelius, which will allow us to push the sizes of treatable systems up.