⚙️ Hardware architecture and transition to barium ions
In the design of Helios, there has been a strategic change in the ion species used: replacing ytterbium ions with barium ions.
This change has made it possible to control qubits using visible light lasers, which leads to cost reduction, increased optical stability, and most importantly, the ability to instantly detect leaks at the atomic level.
The basic architecture of the system is still based on the QCCD (Quantum Charge Coupler Device) technology, which by embedding a “spinning storage ring” and cross-connections, enables the physical movement of ions and the realization of all-to-all connectivity (All-to-All Connectivity) without the need for costly SWAP operations.
📊 performance indicators and benchmarks
With 98 physical qubits, the Helios system has achieved a level of precision unmatched by any current commercial system:
– One-qubit gate fidelity: 99.9975%
– Twobit Gate fidelity: 99.921%
This level of accuracy makes it possible to implement high-performance error correction codes. In initial tests, the system managed to create 94 logical qubits using the Iceberg code, which showed performance beyond the break-even point; This means that the coded logical qubits had a lower error rate than their physical components.
Software ecosystem and hybrid computing
Along with the hardware, Quantinuum unveiled a new software stack that uses the Guppy (Python-based) programming language. This ecosystem is designed for real-time hybrid computing, where classical (GPU-based) and quantum processing dynamically interact during circuit execution.
Industrial applicability
In the test phase, institutions such as JPMorgan Chase and SoftBank have used this system to simulate complex financial models and material physics (high temperature superconductivity). According to Quantinuum, simulating the results obtained from Helios on Google’s RCS benchmark would be thermodynamically impossible for the most powerful classical supercomputers.
🔗 Link to the source