Fermilab accelerated development of fault-tolerant quantum architectures from five months to three weeks, a sevenfold speedup, using NVIDIA’s newly expanded CUDA-Q platform. The company has added CUDA-Q Logical, an orchestration layer designed to simplify the complex codesign process of building applications for fault-tolerant quantum computers, NVIDIA says.

This open, programmable system allows researchers to rapidly test different configurations for performance with logical qubits, helping to overcome errors in physical qubits and enabling practical applications like drug discovery and materials development. “Quantum computing is maturing into an era of logical qubits,” said Timothy Costa, vice president and general manager of quantum at NVIDIA, “and researchers need an open, customizable platform capable of representing all aspects of a fault-tolerant system.”

CUDA-Q Logical Accelerates Fault-Tolerance Architecture Development

Iceberg Quantum used CUDA-Q Logical to refine its fault-tolerant architecture for Diraq qubits, achieving a projected reduction in physical qubit requirements to 150,000 for 1,000 logical qubits, a ten-fold improvement over prior estimates. This optimization stems from the platform’s ability to rapidly assess diverse implementation possibilities, streamlining the path toward practical, error-corrected quantum computation.

The capability to model such complex systems efficiently addresses a critical bottleneck in quantum hardware development, allowing researchers to focus on minimizing resource demands. Fermilab’s exploration of fault-tolerant applications benefited significantly from the platform, transforming previously lengthy design processes into repeatable computational workflows.

Using CUDA-Q Logical, the team validated existing results and evaluated the interplay between physical qubits, runtimes, and various error-correction strategies across different quantum hardware, according to NVIDIA. “Using CUDA-Q Logical, our team explored combinations of these resources in just three weeks, compared with what would have typically taken about five months of building specialized infrastructure,” researchers reported, demonstrating a seven-fold speedup. This speedup allows for more iterative design cycles and faster validation of novel approaches.

Sandia National Laboratories contributed to the CUDA-Q ecosystem by making QUOPS, a benchmark for evaluating fault-tolerant quantum hardware, available within the NVIDIA platform, the firm reports. “Our mission right now is to bring that about sooner, by accelerating our industry partners’ progress,” stated a representative from Sandia, adding, “To do that, we, and other quantum computing stakeholders, have to be able to track and forecast the growth of quantum computer abilities.

We created QUOPS to do exactly that, and we’re excited to see it used by quantum computing vendors and customers.” The integration of QUOPS provides a standardized metric for assessing progress and comparing different hardware implementations, fostering greater transparency and collaboration within the field.

At Sandia, we can’t wait to see fault-tolerant quantum computers helping to solve problems of national importance for the Department of Energy and the United States.
Timothy Proctor, co-director of Sandia National Laboratories‘ Quantum Performance Laboratory

QUOPS Benchmark Measures Progress Toward Fault-Tolerant Quantum Systems

The newly available QUOPS benchmark provides a standardized method for evaluating fault-tolerant quantum hardware, moving beyond metrics focused solely on physical qubits. Developed by Sandia National Laboratories, QUOPS assesses readiness for practical applications, offering a hardware-agnostic approach to track progress toward scalable, error-corrected systems.

Initial benchmarks, detailed in a preprint ahead of IEEE Quantum Week, have already been run on quantum processing units from Google, IBM, and Quantinuum, with Sandia sharing these results, establishing a baseline for future comparisons. “Fault-tolerant quantum computing is the path to unlocking new scientific discovery, but getting there will require researchers to codesign algorithm, error correction, architectures and hardware together,” said Anna Grassellino, chief technology officer at Fermilab and director of the Superconducting Quantum Materials and Systems Center.

This integration allows for rapid testing of diverse configurations, a capability NVIDIA bolstered through partnerships with companies like IQM, whose Garnet and Emerald processors are now accessible via CUDA-Q, and QuEra Computing, a launch partner for NVQLink.

Fault-tolerant quantum computing is the path to unlocking new scientific discovery, but getting there will require researchers to codesign algorithm, error correction, architectures and hardware together.
Anna Grassellino, chief technology officer at Fermilab and director of the Superconducting Quantum Materials and Systems Center

Iceberg Quantum Reduces Logical Qubit Requirements with CUDA-Q

This advance stems from CUDA-Q Logical’s capacity to rapidly assess diverse implementation possibilities, enabling Iceberg Quantum to explore a wider range of configurations than previously feasible. The ability to model and optimize qubit requirements is critical as the field moves toward practical quantum computation, and this reduction in hardware demands represents a step toward scalability.

The platform’s orchestration layer provided a method for evaluating potential implementations of Iceberg’s architecture, accelerating the process of identifying efficient designs. Fermilab also benefited from CUDA-Q Logical, transforming complex fault-tolerant system designs into a repeatable computational workflow and accelerating algorithm development.

Researchers at Fermilab validated prior results and evaluated physical qubits, runtimes, and resource requirements across different error-correction approaches, achieving a seven-fold speedup, reducing a five-month process to just three weeks. “The addition of CUDA-Q Logical provides power and flexibility to explore fully integrated, co-optimized systems regardless of qubit type and architecture — drastically shortening the timeline to useful quantum-GPU supercomputing,” said Timothy Costa, VP and GM of Quantum at NVIDIA.

NVIDIA’s expansion of the CUDA-Q platform, with the inclusion of CUDA-Q Logical, is gaining traction within the quantum industry, with adoption by QPU makers and labs including Infleqtion and IQM Quantum Computers. NVIDIA, founded in 1993 and headquartered in Santa Clara, provides quantum computing services and infrastructure through its CUDA Quantum platform and DGX Quantum systems, and maintains partnerships with Quantinuum, QuEra Computing, and IonQ among others.

Quantum computing is maturing into an era of logical qubits, and researchers need an open, customizable platform capable of representing all aspects of a fault-tolerant system.
Timothy Costa, vice president and general manager of quantum at NVIDIA

NVIDIA CUDA-Q Integrates with Broad Quantum Computing Ecosystem

This rapid prototyping is critical for advancing research in areas like drug discovery, financial modeling, and materials science, where complex quantum simulations are essential. Beyond accelerating development cycles, NVIDIA is fostering broader adoption through open standards and integration with diverse hardware. Diraq used NVIDIA Ising, an open model family for AI-driven quantum computing, to calibrate its silicon-based qubit processor. Several companies, including Anyon Computing and Quandela, are integrating their quantum processors with NVIDIA’s NVQLink architecture, an open system designed for tightly coupling quantum processors with GPU supercomputers.

Quantum Machines demonstrated integration of its supercomputing resources with qubits at the Israeli Quantum Computing Center, highlighting the platform’s versatility, NVIDIA reports. The CUDA-Q ecosystem is also seeing growth in funding and collaborative research. BlueQubit launched the Quantum Flywheel grant program, providing researchers with access to NVIDIA accelerated computing through CUDA-Q.

Qedma Quantum Computing and QCentroid integrated their technologies with CUDA-Q to enhance quantum error correction and application deployment. IonQ reported progress on DQAOA-GPT, a quantum generative AI framework, utilizing NVIDIA’s accelerated computing capabilities, while MITRE published work on building GPU-accelerated digital twins of quantum sensors, demonstrating the platform’s expanding applications. CUDA-Q Logical is available through GitHub, and the QUOPS repository is now accessible for further exploration. Sandia shared early results in a preprint posted ahead of IEEE Quantum Week, reporting initial QUOPS benchmarks for QPUs from Google, IBM and Quantinuum.