Dynamiqs

Dynamiqs is an open-source Python library sponsored by Alice & Bob that enables high-speed simulation of open and closed quantum systems leveraging NVIDIA accelerated computing. The NVIDIA-powered GPU acceleration increases efficiency of matrix operations by up to 60x, allowing simulation of large, complex systems involving multiple qubits and physical hardware components. Developed based on Google AI's JAX machine learning library and Diffrax state-of-the-art differential equations library. The team includes researchers from Alice & Bob, University of Sherbrooke, Yale University, INRIA, Ecole de Mines, and ENS. In 2025, integration with NVIDIA CUDA-Q platform accelerates complex quantum dynamics simulation by up to 75x on early benchmarks. Differentiability and GPU acceleration provide researchers with improved and completely new functionalities for quantum device calibration.