The integrated system for quantum-accelerated computing.
Coming soon.
For quantum leaders looking to integrate the best of quantum and classical processing and build quantum accelerated supercomputers, NVIDIA DGX™ Quantum offers a solution. Combining NVIDIA Grace Hopper™ Superchips and the Quantum Machines OPX+ control system, DGX Quantum runs hybrid algorithms, quantum error correction, calibration, and control at max performance.
DGX Quantum uses a PCIe Gen5 interconnect to deliver submicrosecond latency between GPU and QPU, enabling GPU-accelerated hybrid workloads with no performance bottlenecks. This accelerates essential workloads like quantum error correction within the coherence time of the QPU.
DGX Quantum scales with both quantum and classical computing needs. As quantum computers scale up, more DGX Quantum nodes can be added to couple to more qubits. The DGX Quantum nodes can also be connected to more classical-only nodes up to a full supercomputer, paving the way for quantum-accelerated supercomputing.
DGX Quantum leverages NVIDIA® CUDA® Quantum, an open-source programming model and compiler toolchain that enables high-performance coprogramming of CPUs, GPUs, and quantum processing units (QPUs).
Integrating the Quantum Machines OPX+ control system with NVIDIA’s Grace Hopper Superchip, DGX Quantum accelerates any type of quantum processor.
DGX Quantum scales with both quantum and classical compute requirements, from a few to thousands of qubits and from a single GPU to a quantum-accelerated supercomputer.
Watch Session: Defining the Quantum-Accelerated Supercomputer
Read Press Release: NVIDIA Announces New System for Accelerated Quantum-Classical Computing
Web Page: NVIDIA CUDA Quantum
Sign up for quantum news and updates.
NVIDIA Privacy Policy