NVIDIA CUDA-Q
The high-performance platform for hybrid quantum-classical computing
To do algorithm research and build applications for future quantum advantages, a bridging technology is needed to enable dynamic workflows across disparate system architectures. With a unified and open programming model, NVIDIA CUDA-Q is an open-source platform for integrating and programming quantum processing units (QPUs), GPUs, and CPUs in one system. CUDA-Q enables GPU-accelerated system scalability and performance across heterogeneous QPU, CPU, GPU, and emulated quantum system elements.
Get Started
Key Benefits
Productive
Streamlines hybrid quantum-classical development with a unified programming model, improving productivity and scalability in quantum algorithm research.
Flexible Platform
Connects to partner QPUs and GPU simulators, easy toolchain integration, and interoperates with modern GPU-accelerated applications.
High Performing
2500X simulation speedup on a four A100 GPU for up to 26 qubits, and scaling to 40 qubits by distributing the simulation across 128 GPU nodes.
A Range of Features
- Kernel-based programming model extending C++ and Python for hybrid quantum-classical systems
- Native support for GPU hybrid compute, enabling GPU pre- and post-processing and classical optimizations
- System-level compiler toolchain featuring split compilation with NVQ++ compiler for quantum kernels, lowering to Multi-Level Intermediate Representation (MLIR) and Quantum Intermediate Representation (QIR)
- Initial NVQ++ benchmark shows 287X improvement in end-to-end VQE performance with 20 qubits and dramatically improved scaling with system size compared to standard Pythonic implementation
- Standard library of quantum algorithmic primitives
- Interoperable with partner QPUs as well as simulated QPUs using the cuQuantum GPU simulators; partnering with QPU builders across different qubit types
- Interoperable with CUDA and the CUDA software ecosystem
Built for Performance
NVIDIA CUDA-Q enables straightforward execution of hybrid code on many different types of quantum processors, simulated or physical. Researchers can leverage the cuQuantum-accelerated simulation backends as well as QPUs from our partners or connect their own simulator or quantum processor.
NVIDIA CUDA-Q can significantly speed up quantum algorithms, compared to other quantum frameworks. Quantum algorithms can achieve a speedup of up to 2500X over CPU, scaling number of qubits using multiple GPUs.
Quantum Computing Partners
Explore More Resources
- CUDA-Q Press Release: NVIDIA Announces Hybrid Quantum-Classical Computing Platform
- NVIDIA Special Address at Q2B: Defining the Quantum Accelerated Supercomputing Platform
- Blog: CUDA-Q Introduces More Capabilities for Quantum Accelerated Supercomputing
- Blog: An Introduction to Quantum Accelerated Supercomputing
- Blog: Merge Ahead: Researcher Takes Software Bridge to Quantum Computing
- Blog: Introducing CUDA-Q: The Platform for Hybrid Quantum-Classical Computing
- Explainer: What is a QPU?