Quantum computers save not just time, but also energy ⚡ экспресс
The energy consumption of a realistic photonic quantum computing architecture for boson sampling—a paradigmatic task of quantum supremacy—is analyzed. Using the Metric-Noise-Resource methodology, a quantitative link is established between control parameters, dominant noise processes, and energy resources via a specialized performance metric. The energy per computational sample is estimated, and operating regimes with optimal energy efficiency are identified. A comparison with the best classical implementations reveals the existence of a quantum energy advantage—defined as a lower energy cost per sample—that emerges before computational advantage does, even when classical algorithms remain faster. An experimentally viable boson sampling architecture is proposed, incorporating a full noise and loss budget, which could enable observing quantum energy advantage in the near term.
When making popcorn, a microwave uses electricity only to heat the kernels. A gas stove heats the entire burner and the air around it. The same difference applies between a photonic quantum chip and a classical supercomputer when solving a problem devised by Scott Aaronson. This problem — essentially calculating how light entangles in a maze of mirrors — is an ideal test for a quantum machine.
Secret of the savings is that classical computations generate a lot of entropy, i.e., disorder and heat. The photonic chip, where particles of light travel at speed of light and barely interact, creates far less such "waste". Precise counting of individual photons at the output confirms: energy is spent on the result, not on fighting chaos.
Surprisingly, the quantum processor wins in energy before it wins in time. This means that future quantum accelerators will take on the most energy-hungry tasks and drastically reduce the carbon footprint of data centers.
🎯 The essence of the experiment is simulating light passing through a labyrinth of mirrors and beam splitters. For large setups, even the best supercomputers cannot predict the outcome, but the quantum chip does it naturally, "computing" by the laws of nature itself.
🎬 Sci-fi often portrays quantum computers as all-powerful. Reality adds a pleasant surprise: they also save energy.