Home › Hard problems › Chips, compute and energy

Chips, compute and energy

AI for chip design

Solo: possibleOne person can test a narrow slice of it.

monthsmonths ofexpert workbillions of partsplace and wire every partAI"we did it better!"?published resultsare disputed
Placing and wiring billions of parts on a chip takes months of expert work, and the published claims that AI does it better are disputed.

Why it is still open

Placing and routing billions of transistors takes months of expert work. Published AI results are disputed.

An experiment one person can run in 8 weeks

On the open OpenROAD flow, compare a learned placement method with the default on one open chip design.

Verdict: Solo: possible

Free tools exist, but gains are hard to prove.

The solo test behind the verdict

  1. Can it be measured? There is a number that shows you did better, such as an error rate, a catch rate or a cost per answer.
  2. Is it cheap to try? It runs on a laptop, free open models and open data, or under about ₹50,000 of rented GPU time.
  3. Is there a public yardstick? A benchmark or dataset others use, so your result can be compared and believed.
  4. Would a small win get noticed? A lab, a funder or a customer would care about a 10% improvement, not only a full solution.

Four yeses means Solo: strong. Two or three means Solo: possible. Fewer means Follow, don't build.

PreviousPower and cooling for AI data centresNextComputing with light or analog circuits

All 79 hard problems · Working on this, or know of a result we missed? basant@cleandataengineering.ai