Scientists used 7,000 GPUs to simulate a tiny quantum chip in extreme detail

2026-03-18 · Source: Neural Interfaces News -- ScienceDaily · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Expert, medium

Summary

Researchers at Berkeley Lab's Quantum Systems Accelerator (QSA) have developed a novel method for simulating quantum chips in extreme detail before fabrication. Utilizing nearly 7,000 NVIDIA GPUs on the Perlmutter supercomputer, they modeled a 10mm x 0.3mm multilayer chip with micron-scale features, discretizing it into 11 billion grid cells. This approach, powered by the ARTEMIS exascale modeling tool, captures the physical layout, material properties, and real-time electromagnetic wave propagation, including nonlinear effects and qubit interactions. Unlike simplified "black box" models, this full-wave physical-level simulation allows for early identification of design flaws and optimization of signal coupling, significantly accelerating the development of next-generation quantum hardware.

Key takeaway

For AI Scientists designing quantum hardware, this detailed simulation approach changes how you validate chip designs. You should consider integrating full-wave physical-level simulations, especially those accounting for material properties and real-time electromagnetic interactions, to predict chip behavior accurately and avoid costly fabrication errors. This method allows for rapid iteration and optimization of quantum chip layouts and materials.

Key insights

Detailed electromagnetic simulation of quantum chips using exascale computing accelerates hardware design and problem detection.

Principles

• Full-wave physical modeling improves quantum chip design.
• Time-domain Maxwell's equations capture nonlinear chip behavior.

Method

The ARTEMIS exascale modeling tool simulates quantum chip electromagnetic wave propagation by discretizing the chip into billions of grid cells and running millions of time steps on thousands of GPUs.

In practice

• Simulate chip designs before fabrication to catch issues early.
• Use time-domain simulations to account for nonlinear effects.

Topics

• Quantum Chip Simulation
• High-Performance Computing
• GPU Computing
• Electromagnetic Simulation
• Quantum Hardware Design

Code references

• amrex-microelectronics/artemis

Best for: AI Scientist, AI Researcher, Research Scientist, AI Hardware Engineer

Related on AIssential

• New quantum algorithm solves “impossible” materials problem in seconds — A quantum-inspired algorithm using tensor networks can simulate complex quasicrystals, accelerating quantum materials and computing.
• Physics Simulation Just Crossed A Line — A CPU-based physics simulation uses domain decomposition to achieve superior speed and realism for complex deformable objects.
• Introducing NVIDIA Ising — NVIDIA's Ising AI models automate quantum processor calibration and error correction, accelerating quantum computing development.
• New light-powered chip could accelerate AI and quantum computing — A new integrated photonic valleytronic chip processes light-based information at room temperature, enabling faster, energy-efficient computing.
• QMatter secures $1.2M to tackle quantum scaling challenges — QMatter's quantum compression reduces complex quantum problems for more efficient processing on diverse hardware.
• Exploring quantum utility for aerospace — Quantum-centric supercomputing offers exponential speedups for specific aerospace workloads like materials simulation and CFD.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Neural Interfaces News -- ScienceDaily.