News

Intel Addresses the Quantum Interconnect Bottleneck with Horse Ridge Cryogenic IC

May 13, 2021 by Jake Hertz

In a recent article in Nature, Intel has quantitatively proven the efficacy of their cryogenic IC Horse Ridge.

Quantum computing is currently in a stage very reminiscent of early computers—they’re the size of rooms, they cost millions of dollars, and their functionality to date is only a fraction of their potential.

 

An Intel engineer works on a quantum computer. Image from Intel
 

Researchers across the world have been looking to push the field forward, addressing all major limitations and bottlenecks that exist in current systems. Intel is one of the companies looking to push the state of quantum computing, and this week they made news in the field.

Intel and QuTech's new paper in Nature showed promising results that their second-gen cryogenic IC may be able to address the quantum interconnect bottleneck.

 

The Quantum Interconnect Bottleneck 

One of the largest barriers to being able to scale down quantum computers (at least concerning the cables and wires that sit around the quantum system) from their current, massive size is the quantum interconnect bottleneck

The first piece of this challenge is the fact that current quantum architectures are set up in a way such that each individual qubit is controlled by unique external circuitry. These control signals are currently produced by conventional electronics. This in itself is a challenge towards scalability as scaling more qubits requires linearly scaling hardware.

 

Typical quantum computer block diagram with temperatures. Image used courtesy of E. Charbon et al., IEDM 2016.

 

The second piece of this challenge is that, in order to ensure deterministic quantum mechanical behavior, quantum computers need to operate at cryogenic temperatures. Conventional electronics, which are being used as control circuitry, cannot function correctly at these temperatures. The result is that current quantum architectures have the classical electronics operating at room temperature while the quantum electronics are at cryogenic temperatures. 

The aggregate result of these two challenges is a huge bottleneck on scalability, requiring long and clunky wires to connect electronics between different temperature environments. In a previous article on AAC, we reported that a research experiment needed around 200 wideband coaxial cables, 45 microwave circulators, and racks of room-temperature electronics to control just 53 qubits.

 

Horse Ridge and Intel’s News

At the end of 2019, Intel announced Horse Ridge which is a cryogenic control RF SoC based on 22nm FinFET technology. The idea of the IC was to bring the control circuitry closer to the quantum electronics, removing the quantum interconnect bottleneck by removing the need for clunky cabling.

 

Horse Ridge, Intel’s cryogenic control SoC. Image from Intel.

 

Now, after over a year of testing and research, Intel, in collaboration with QuTech, has published a paper in Nature, which shows promising results from Horse Ridge. 

Their newly published research announces that Horse Ridge was able to successfully operate at 3 Kelvin, outputting microwave control signals to drive qubits being kept at 20 millikelvin. Their research used a randomized benchmark to show that they could achieve coherent control of two qubits with 99.7% fidelity, a number that's consistent with conventional room temperature control circuitry. Along with this, they were able to perform with these two qubits using a single cable, further addressing the limitation of needing unique control circuitry for each qubit.

 

Why This Matters 

In effect, this research from Intel has proven the efficacy of cryogenic control electronics, with their results essentially proving that their IC works just as well as conventional room temperature control schemes. 

This is a huge step in the right direction for quantum computing, removing one of the major identified limitations to scaling. According to Intel, this research flags that it may be possible to eventually fully integrate the controller chip and the qubits on the same or package. 

With continual advancements like this, we could be poised to see the dramatic scaling and widespread adoption of quantum computing that we once saw with classical computing.