Library boosts quantum simulation 75-fold with GPUs
Cette publication existe aussi en Français
Announced today at GTC Paris, the ongoing integration of the NVIDIA CUDA-Q platform with the Dynamiqs outperforms the most widely used libraries today, accelerating the simulation of complex quantum dynamics by up to 75x on early benchmarks. The NVIDIA CUDA-Q is an open-source quantum development platform for hybrid quantum-classical supercomputing.
The Dynamiqs open-source, high-performance quantum simulation library from Alice & Bob enables rapid simulation of both open and closed quantum systems by utilising GPU-accelerated computing. Researchers can now significantly speed up the simulation of time-dependent quantum systems that change and evolve rapidly, such as quantum processing units (QPUs). Dynamiqs increases the size of systems that can be practically simulated and conducts parameter sweeps across a broad range of conditions in a fraction of the time previously required.
The long-lasting collaborative project with NVIDIA developed APIs that translate high-level programming instructions to low-level CUDA, providing the critical ability to interact with GPUs specifically optimised for quantum applications. The integration will continue over the following months, with further performance improvements expected.
“Simulation is a critical step in the development of useful quantum processors, allowing us to understand how these complex quantum systems behave,” said Théau Peronnin, CEO of Alice & Bob. “Thanks to the integration with NVIDIA CUDA-Q, Dynamiqs can now run these simulations even faster, speeding up the development of our QPUs.”
“With Dynamiqs, our goal was to make time-dependent quantum systems simulations faster, and from the beginning, we decided to fully run them on GPUs, a new approach in the field,” said Ronan Gautier, a member of the core Dynamiqs development team and a Theoretical Physicist at Alice & Bob.
In addition to speed, Dynamiqs unlocks new capabilities for researchers, who can now use automatic differentiation to compute gradients of simulation outputs for various input parameters. This tool is essential for paramount tasks, such as quantum optimal control, which guides the system toward the target state with high accuracy; parameter estimation, which infers unknown properties from data; and quantum state tomography, which reconstructs the quantum state from measurements.
“CUDA-Q’s GPU acceleration opens up new possibilities for fast, scalable, and intelligent quantum system design and analysis,” said Tim Costa, Senior Director of Quantum and CUDA-X at NVIDIA. “By combining this with its other capabilities, Dynamiqs is helping provide researchers with access to state-of-the-art accelerated computing that they need for practical breakthroughs in quantum research.”
Alongside a $50 million quantum lab and a €100M funding round, the work on speeding up quantum simulation represents another milestone toward the development of the first useful quantum computer by 2030.
www.dynamiqs.org
https://github.com/dynamiqs/dynamiqs
https://the-cat-tree.slack.com
www.alice-bob.com
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
