QC Design has announced a major performance breakthrough in its quantum fault-tolerance design software, Plaquette, through integration with NVIDIA’s cuQuantum SDK. The update leverages cuTensorNet to enable high-performance, full-state simulations of fault-tolerant quantum circuits. As a result, Plaquette can now simulate over 400 qubits on a single NVIDIA RTX 4000 Ada GPU with 20GB memory—an improvement over current CPU-based Monte Carlo simulators, which typically simulate only ~60 qubits using 120GB of RAM.
The GPU-accelerated simulations also deliver significant performance gains. Plaquette achieves at least a 180x speedup in sampling throughput for 60-qubit circuits compared to CPU-based alternatives, making it viable to simulate complex error models such as coherent errors and amplitude damping in realistic fault-tolerant architectures. This enables hardware teams to precisely model, analyze, and optimize quantum error correction protocols with full-state fidelity, a capability critical for designing practical fault-tolerant quantum systems.
Plaquette currently models over 20 distinct hardware imperfections and supports the widest range of fault-tolerance protocols in the literature. With GPU acceleration now integrated as a core feature, QC Design aims to reduce development timelines and cost for quantum hardware teams.
More details are available in the official announcement here.
March 21, 2025
Leave A Comment