#sc25

Green House - Peace Piece 

The Dell AI Factory

Quantinuum Helios: Connecting HPC to Quantum

Quantinuum, the world's largest integrated quantum computing company, has led the next phase of high-performance computing (HPC) by demonstrating quantum-classical integration and proving quantum advantage in a difficult, practical scientific experiment. As the vital link between the long-standing HPC community and the emerging quantum era, the company’s extensive presence at the International Conference for High Performance Computing, Networking, Storage, and Analysis (SC25), which took place in St. Louis, Missouri, from November 16–21, was confirmed.

The SC25 Integration Frontier

The industry’s rapid transition to hybrid quantum-classical computing was underlined by Quantinuum’s involvement at SC25. At booth #4432, the company showed a live trapped-ion quantum hardware demonstration unit and a mockup of its next-generation Helios system and chip architecture.

Quantinuum promoted easy absorption. The Helios hardware; Nexus, the all-in-one quantum computing platform; and Hybrid Workflows, which highlighted the integration of NVIDIA CUDA-Q with Quantinuum Systems, were the major subjects of the experts’ daily tutorials on the company’s full-stack platform. These talks showed how quantum resources are being made accessible in HPC settings.

Grahame Vittorini of Quantinuum co-led a Birds of a Feather (BoF) session titled “Bridging the Gap: Making Quantum-Classical Hybridisation Work in HPC,” which addressed the need for a clear roadmap to integrate Quantum Processing Units (QPUs) into HPC infrastructures. This workshop brought together system architects, HPC professionals, and quantum scientists to set scaling milestones.

Helios Gives Materials Science Quantum Edge The most notable Quantinuum revelation focused on the Helios quantum computer's scientific capabilities, which exhibited a quantum advantage never previously attainable. In particular, the large-scale simulation of the non-equilibrium Fermi-Hubbard model was a condensed matter physics milestone. This model is necessary to explain light-induced superconductivity by explaining electron mobility and interaction in a crystal.

The search for a material that can superconduct at room temperature is still considered a scientific “holy grail,” with the potential to alter power grids and medical technology. However, the computing limits of typical supercomputers, which can only handle extremely small instances of the Fermi-Hubbard model, have long hindered this pursuit.

Helios used up to 90 qubits (72 system qubits plus 18 ancilla) to imitate the dynamics of a 6×6 lattice using a new fermionic encoding. This system's quantum state is complete at a previously impossible 2^72 dimensions. Helios is a “qubit-based laboratory” that helps researchers study phenomena better than other machines. This features flexible measurements, long dynamical simulation, unrestricted state preparation, and significant “off-diagonal” observables that suggest superconductivity.

Helios reported the first non-zero superconducting pairing correlations, or “eta” pairing correlations, on any quantum computer platform. This study demonstrates that the computer can help researchers at the Max Planck Institute in Hamburg understand microscopic processes like light-induced superconductivity. In contrast to analogue simulators or physical experiments, Helios allows users unlimited control over simulation parameters, allowing for the examination of scenarios that are not conceivable with genuine materials.

It is expected to revolutionise materials science by allowing researchers to digitally test novel superconductors, reducing costly laboratory trial-and-error.

Architectural Triumphs and Unrivaled Fidelity

The extraordinary engineering of the Helios system, which consists of 98 physically connected qubits, provides the cornerstone of these accomplishments. Helios is the most accurate commercial quantum computer in the world with a single-qubit gate fidelity of 99.9975% and a two-qubit gate fidelity of 99.921% over all qubit pairs.

Benchmarks such as Random Circuit Sampling (RCS) provide a clear indication of accuracy. Quantinuum shows that the power needed to execute the same RCS computation classically in the same period would be higher than all the visible universe stars. Helios achieved this with the power of a data centre rack, demonstrating quantum efficiency.

Helios architecture is much improved over QCCD. An engineering marvel called a “traffic intersection for qubits,” it is the first commercial ion junction, enabling routing and scaling. Helios QPU now behaves as a processor with separate ring storage (memory), cache, and computational (logic) zones.

This design allows concurrent activities like cooling and sorting qubits, speeding up the processor and reducing errors. Complete all-to-all connection is a major advantage of QCCD over fixed-qubit systems. Helios has switched its qubits from ytterbium to barium, which allows atomic-level leakage error detection and manipulation using cheaper, more reliable visible spectrum lasers.

The Practical Way to Fault Tolerance

Although most of the industry promises fault tolerance, Quantinuum is already supplying it. Only this company has established a worldwide fault-tolerant gate set and has best-in-class logical fidelities.

Helios' 98 physical qubits were used to create 94 logical qubits entangled in one of the largest GHZ states. Logical qubits based on the effective Iceberg code outperformed physical qubits with the same algorithm and had better-than-break-even fidelity.

Iceberg code, which demands a 1:1 physical-to-logical qubit ratio, has the best industry encoding efficiency. Helios also produced 48 error-corrected logical qubits at a hitherto unthinkable 2:1 encoding rate.

Enhanced real-time control systems that execute dynamic quantum algorithms that react quickly to measurement data help this breakthrough. For the first time on a quantum computer, our engine allows GPU-accelerated classical and quantum calculations in a single application. Guppy, a new quantum programming language, lets programmers design dynamic circuits with loops, arbitrary control flow, and dynamic qubit allocation, which are essential for large-scale fault-tolerant quantum computing.

Scarfolk Council