SEEQC builds first quantum GPU interface
Cette publication existe aussi en Français
US startup SEEQC has built the first digital interface to connect quantum chips to GPUs for real-time error correction.
SEEQC has demonstrated an end-to-end interface as a key step to a fully digital chip-to-chip interface between a quantum processor and Nvidia GPU.
The company uses superconducting Single Flux Quantum (SFQ) technology to build logic that works with quantum processors in a cryostat. This provides microsecond latency and reduce bandwidth bottlenecks.
The idea is that this interface can link effectively to GPUs running at room temperature to handle the error correction algorithms to control the quantum processor in real time. The interface was built using PCIe, the current high-speed data transfer standard. This will act as a reference for a future generation interface which will introduce custom on-GPU protocol integration to further enhance efficiency, ultimately scaling to million-qubit systems.
This universal chip-based interface, when integrated with any GPU, forms a fully integrated full-stack quantum/classical processor that operates at the same temperatures as qubits to offer real-time digital readout and control via SEEQC’s PRISM firmware and software.
This offers an alternative to the common approaches to control systems for quantum error correction which send control signals through long cables from cryogenically-cooled qubits to room-temperature electronics.
“Quantum and classical computing have often been seen as competing forces. It’s even shaped how these systems are designed – built separately, connected inefficiently, and unable to fully lean on each other’s strengths. Our technology and our latest breakthrough with Nvidia changes this. By creating a direct, fully digital link between quantum processors and GPUs, we’re unlocking the true power of both in a way that makes the whole greater than the sum of its parts,” said John Levy, SEEQC co-founder and CEO during a talk at the Quantum Day at the Nvidia GTC conference in San Jose, California today
“Tightly integrating quantum hardware and AI supercomputing is critical for the quantum error correction that enables useful large-scale quantum computing. Our work with SEEQC is helping to close the gap between these technologies and bring reliable quantum applications closer to reality,” said Sam Stanwyck, Group Product Manager for Quantum Computing at Nvidia.
The SEEQC proprietary Single Flux Quantum (SFQ) chips produced at the company’s multi-layer superconductive electronics chip foundry located in Elmsford, New York and has teams in London, UK, and Naples, Italy, where the team has a full stack quantum computer.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
