The orchestration layer helps researchers evaluate fault-tolerant quantum system configurations and resource requirements.
NVIDIA announced an expansion of the NVIDIA CUDA-Q open source platform with CUDA-Q Logical, an orchestration layer for developing and evaluating applications for fault-tolerant quantum computers. The platform provides tools for programming and verifying the components involved in quantum computing systems with logical qubits.
Fault-tolerant quantum processors use logical qubits to reduce the effects of errors in physical qubits. This approach is being explored for applications that require larger quantum computations, including drug discovery, financial modeling and materials development.
Developing applications for fault-tolerant systems involves coordinating algorithms, error-correction methods, hardware architectures and other QPU components. Changes to any of these elements can affect the resources required to run an application.
CUDA-Q Logical allows researchers to design and coordinate these components and switch between different configurations to evaluate their effects on systems using logical qubits.
CUDA-Q Logical is being used by QPU manufacturers and research organizations including Fermi National Accelerator Laboratory, Infleqtion, IQM Quantum Computers, QCDesign, Quantum Motion and Sandia National Laboratories.
Iceberg Quantum used CUDA-Q Logical to model a fault-tolerant architecture for Diraq’s qubits. The company reported that its modeling showed how 1,000 logical qubits could be created using 150,000 physical qubits, compared with previous estimates from Diraq. CUDA-Q Logical was used to evaluate potential implementations of the architecture and assess the associated hardware requirements.
Fermilab evaluates fault-tolerant applications
Fermilab researchers have used CUDA-Q Logical to validate previous results and evaluate physical qubits, runtimes and other resource requirements across different error-correction methods and quantum hardware configurations.
The researchers converted fault-tolerant system designs into a repeatable computational workflow. According to NVIDIA, this reduced the development time for fault-tolerant algorithms from five months to three weeks.
Sandia develops benchmark for quantum computing systems
Sandia National Laboratories developed QUOPS, an independent cross-platform benchmark for measuring the progress of quantum computing systems toward applications that require fault-tolerant quantum computers.
QUOPS provides a hardware-agnostic and open approach for benchmarking quantum computing systems against common performance goals. Traditional measurements of quantum computing progress have focused on physical qubit counts, fidelity and coherence times.
Sandia reported initial QUOPS results in a preprint ahead of IEEE Quantum Week, including benchmarks for QPUs from Google, IBM and Quantinuum. A QUOPS reference implementation is available through NVIDIA CUDA-Q.
Expanding the quantum-GPU computing platform
The quantum computing ecosystem is also using NVIDIA technologies to integrate quantum processors with GPU-accelerated computing systems.
Diraq used NVIDIA Ising models to calibrate its silicon-based qubit processor. NVIDIA Ising is a collection of open models intended for developing AI applications for quantum computing.
Companies are also integrating quantum processors with GPU supercomputing systems using NVIDIA NVQLink, an open system architecture for connecting QPUs with GPU-based supercomputers. Examples include:
- Anyon Computing: Developed a quantum control system using NVQLink.
- Quandela: Developed an architecture connecting a QPU with GPUs.
- Quantum Machines: Demonstrated an integration of quantum processors with supercomputing resources at the Israeli Quantum Computing Center.
The NVIDIA CUDA-Q open development platform is also being used by organizations working on quantum computing applications:
- BlueQubit: Launched the Quantum Flywheel grant program to provide researchers with access to NVIDIA accelerated computing through CUDA-Q.
- Qedma Quantum Computing and QCentroid: Integrated their technologies with CUDA-Q to support quantum error correction, error mitigation and quantum application deployment.
Other organizations are working with NVIDIA on quantum application development:
- IonQ: Reported progress on DQAOA-GPT, a quantum generative AI framework that uses NVIDIA accelerated computing.
- MITRE: Published work on developing GPU-accelerated digital twins for quantum sensors.
- Phasecraft: Uses NVIDIA cuQuantum to develop a database of molecular simulations for variational quantum eigensolver research.
- UCLA and Caltech: Are working on control sequences used to operate quantum applications.
Availability
CUDA-Q Logical is now available through GitHub. Explore the QUOPS repository.
NVIDIA
www.nvidia.com
