Performance Benchmark Comparison versus QuTip for Five Different Cases on both CPU and GPU based Systems. Credit: Alice & Bob

Alice & Bob has introduced Dynamiqs, an open-source quantum simulation Python library that integrates NVIDIA-accelerated computing to achieve up to 60x faster performance in quantum simulations. Developed in collaboration with researchers from institutions including the University of Sherbrooke, Yale, INRIA, Ecole de Mines, and ENS, Dynamiqs addresses the computational challenges of simulating complex quantum systems, enabling breakthroughs in quantum research and development.

The library leverages NVIDIA GPU acceleration and the differentiability of JAX and Diffrax frameworks to enhance simulation capabilities. GPU acceleration enables rapid large-scale simulations, parameter sweeps, and scalability for studying systems involving qubits and hardware components. Differentiability allows for computing gradients essential for tasks such as quantum optimal control, parameter estimation, and quantum state tomography.

Dynamiqs positions Alice & Bob as a leader in quantum processing unit (QPU) development, facilitating simulations of open systems, fast time-dependent dynamics, and interactions within large Hilbert spaces—tasks previously infeasible with traditional methods. The collaboration with NVIDIA highlights the importance of accelerated computing in advancing quantum research.

Dynamiqs is available for researchers via dynamiqs.org or GitHub, marking a significant step forward in accessible, high-speed quantum simulation tools. For more details, visit Alice & Bob’s news release here as well as a blog article on Dynamiqs here.

November 20, 2024