Alice & Bob has integrated NVIDIA’s CUDA-Q quantum development platform into its open-source Dynamiqs simulation library. CUDA-Q provides hybrid quantum–classical computing capabilities, while Dynamiqs is optimized for GPU acceleration to simulate time-dependent quantum systems. Early benchmarks show the combined toolchain can achieve up to 75x speed improvements over conventional simulation frameworks when modeling large-scale quantum dynamics.
The integration allows Dynamiqs to simulate complex, rapidly evolving quantum processors more efficiently. By offloading tasks to GPUs, researchers can perform large-scale simulations and parameter sweeps across diverse conditions, significantly reducing compute time. Alice & Bob’s development team collaborated with NVIDIA to create APIs that translate high-level simulation code into CUDA instructions optimized for NVIDIA hardware.
In addition to speed improvements, Dynamiqs introduces features such as automatic differentiation, enabling users to compute gradients of simulation outputs. These capabilities are critical for quantum optimal control, parameter estimation, and quantum state tomography—core techniques in quantum hardware development and characterization.
The Dynamiqs–CUDA-Q integration will continue to evolve over the coming months, with further performance gains expected. The open-source code and documentation are available at dynamiqs.org to support adoption and community development.
Read the full announcement from Alice & Bob here.
June 11, 2025
Leave A Comment