Supercomputer simulates quantum chip in unprecedented detail

A broad team of researchers from Lawrence Berkeley National Laboratory (Berkeley Lab) and the University of California, Berkeley have collaborated to perform an unprecedented simulation of a quantum microchip, a key step in perfecting the chips required for this next-generation technology. The simulation used more than 7,000 NVIDIA GPUs on the Perlmutter supercomputer at the National Energy Research Scientific Computing Center (NERSC), a U.S. Department of Energy (DOE) user facility.

Modeling quantum chips allows researchers to understand their function and performance before they are manufactured, ensure they work as expected, and detect any problems that might arise. Zhi Jackie Yao and Andy Nonaka, researchers in the Quantum Systems Accelerator (QSA), Division of Applied Mathematics and Computational Research (AMCR) at Berkeley Lab, are developing electromagnetic models to simulate these chips, a key step in the process of producing better quantum hardware.

“The computer model predicts how design decisions affect the propagation of electromagnetic waves within the chip,” Nonaka said, “to ensure proper signal coupling and avoid unwanted crosstalk.”

Here, they used their exascale modeling tool, ARTEMIS, to model and optimize a chip designed in collaboration with Irfan Siddiqi’s Quantum Nanoelectronics Lab at the University of California, Berkeley and Berkeley Lab’s Advanced Quantum Testbed (AQT). This work will be presented in a technical demonstration by Yao at the International Conference on High-Performance Computing, Networking, Storage and Analysis (SC25).

Quantum chip design integrates traditional microwave engineering in addition to advanced low-temperature physics. This makes a classic electromagnetic modeling tool like ARTEMIS, developed as part of DOE’s Exascale Computing Project initiative, a natural choice for this type of modeling.

A big simulation for a small chip

Not all quantum chip simulations require this much computing power, but modeling the smallest details of this tiny and extremely complex chip required almost all of Perlmutter’s power. The researchers used almost all of its 7,168 NVIDIA GPUs over a 24-hour period to capture the structure and function of a multilayer chip measuring just 10 square millimeters and 0.3 millimeters thick, with engravings just one micron wide.

“I don’t know of anyone who has ever done physical modeling of microelectronic circuits on the full scale of the Perlmutter system. We were using almost 7,000 GPUs,” Nonaka said.

“We discretized the chip into 11 billion grid cells. We were able to run more than a million time steps in seven hours, which allowed us to evaluate three circuit layouts in a single day on Perlmutter. These simulations would not have been possible in that time frame without the full system.”

It is this level of detail that makes this simulation unique. While other simulations tend to treat chips as “black boxes” due to modeling capacity constraints, using Perlmutter’s massively parallel GPUs gave Yao and Nonaka the computing power to delve into the physical details and show the chip mechanism at work.

“We do full-wave physics simulation, which means we care about what material you use on the chip, what layout the chip has, how you wire the metal (niobium or any other type of metal wires), how you build the resonators, what size it is, what shape it is, what material you use,” Yao said. “We care about these physical details and we include them in our model.”

In addition to its fine-grained view of the chip, the simulation mimics the experience of laboratory experiments: how qubits communicate with each other and with other parts of the quantum circuit.

The combination of these qualities – the emphasis on the physical design of the chip and the ability to simulate in real time – is part of what makes simulation unique, Yao said: “The combination is instrumental, because we use the partial differential equation, Maxwell’s equation, and we do it in the time domain so we can incorporate nonlinear behavior.

NERSC has supported numerous quantum information science projects through the Quantum Information Science @ Perlmutter program, which grants Director’s Discretionary Reserve hours on Perlmutter to promising quantum projects. Still, staff say tackling a simulation of this size was an exciting challenge.

“This effort stands out as one of the most ambitious quantum projects on Perlmutter to date, using the computing capabilities of ARTEMIS and NERSC to capture the details of quantum hardware over four orders of magnitude,” said Katie Klymko, a NERSC quantum computing engineer who worked on the project.

Model the next step

Next, the team plans to run more simulations to strengthen its quantitative understanding of the chip design and see how it works as part of a larger system.

“We would like to do a more quantitative simulation so that we can do post-processing and quantify the spectral behavior of the system,” Yao said. “We would like to see how the qubit resonates with the rest of the circuit. In the frequency domain, we would like to compare it with other simulations in the frequency domain to give us greater confidence that, quantitatively, the simulation is correct.”

Ultimately, the simulation will pass the ultimate test: comparison with the physical world. When the chip is made and put through its paces, Yao and Nonaka will see how their model measured up and make adjustments from there.

Nonaka and Yao emphasized that successful simulation of this technology at this level of detail would not have been possible without strong collaboration within the Berkeley community, from AMCR to QSA and AQT to NERSC, which supported the simulation with staff expertise in addition to computing power. The collaboration has yielded important results for the advancement of science, said Bert de Jong, director of the QSA.

“This unprecedented simulation, made possible by a broad partnership between scientists and engineers, is a crucial step in accelerating the design and development of quantum hardware,” he said. “More powerful and higher-performance quantum chips will unlock new capabilities for researchers and open new scientific avenues.”