photonic computing
In plain wordsUsing light instead of electricity to compute or move data. Light can carry huge amounts of information with little energy, but building full computers from it is still very hard.
Computation, or data movement, using photons rather than electrons. For AI the main goal is optical matrix multiplication: interferometer meshes, microring resonators or free-space optics encode inputs in the intensity or phase of light so that a matrix–vector productmatrix multiplication, n.: The rule for combining two grids of numbers into a new grid by repeatedly multiplying rows by columns. It is the single calculation AI chips spend most of their time doing. is computed as light propagates, potentially at very high speed and low energy. Obstacles include the weak interaction of light with light (which makes nonlinear activation functions hard), the energy and latency lost converting between electrical and optical signals, limited analog precision and the large size of optical components. Nearer-term success has come in communication: optical interconnectsinterconnect, n.: The wiring and networking that connect chips to each other — inside a server, across a data center, or within a single chip — so they can share data and work as one., optical circuit switches (used in Google's TPU v4 and later pods) and co-packaged optics in network switches, introduced from 2024 and shipping from both Nvidia and Broadcom by 2026.
Origin
Optical information processing has been studied since the 1960s, notably with optical correlators for pattern recognition. Yichen Shen, Nicholas Harris and colleagues at MIT demonstrated a programmable nanophotonic processor for neural-network inference in Nature Photonics (2017), prompting startups such as Lightmatter and Lightelligence.
Related
- analog in-memory computinganalog in-memory computing, n.: Doing AI math inside the memory itself, using electrical currents through stored values, instead of shuttling numbers back and forth to a separate processor. It could save lots of energy but is imprecise.
- interconnectinterconnect, n.: The wiring and networking that connect chips to each other — inside a server, across a data center, or within a single chip — so they can share data and work as one.
- neuromorphic computingneuromorphic computing, n.: Computer chips designed to work more like brains, with artificial neurons that send brief spikes only when needed, aiming to do AI tasks with far less energy.
- advanced packagingadvanced packaging, n.: Ways of joining several chips very closely inside one package — side by side on a silicon base or stacked on top of each other — so they act almost like one bigger chip.
- AI acceleratorAI accelerator, n.: Any chip built or tuned to speed up AI — graphics chips, Google's TPUs, phone 'neural engines' and many startup designs. Such chips do the heavy math of training and running models.
Sources
- Shen, Y., Harris, N. C. et al. (2017). Deep learning with coherent nanophotonic circuits. Nature Photonics 11, 441–446.
Mentioned in
interconnectinterconnect, n.: The wiring and networking that connect chips to each other — inside a server, across a data center, or within a single chip — so they can share data and work as one., advanced packagingadvanced packaging, n.: Ways of joining several chips very closely inside one package — side by side on a silicon base or stacked on top of each other — so they act almost like one bigger chip., analog in-memory computinganalog in-memory computing, n.: Doing AI math inside the memory itself, using electrical currents through stored values, instead of shuttling numbers back and forth to a separate processor. It could save lots of energy but is imprecise., quantum machine learningquantum machine learning, n.: Research into whether quantum computers could learn from data faster or better than ordinary computers, and into using AI to study quantum systems. Practical advantages have not yet been shown.