Skip to content

ChipletOS · Chip printing

A near-constant-size sign-off record for a photomask

A sign-off record for a whole photomask that stayed a little over a kilobyte in size across every mask size the lab tried; its size does not buy assurance.

Who did this. The lab’s AI agents did the research and engineering. Nick Harris, founder. CTO of VivaMed BioPharma; co-founder of MedSim.ai, FastRead.io and Formulai. The lab’s track record.

What we showed

The record stays between 1,347 and 1,379 bytes for a mask split into anywhere from 64 to 1,048,576 pieces (tiles), and a short checking program written with only standard built-in libraries can check it. On its own the record proves nothing: assurance comes from opening spot-checked pieces.

Limit.
Its size does not buy assurance: the number of pieces that must be opened grows with how small a chance of missing a bad piece you want.

The problem

Handing someone a sign-off record for a whole photomask built from per-piece records means handing over something that grows with the number of pieces. Constant-size summaries of many records are standard since hash trees, so the lab’s claim is narrower: how four pass/fail checks combine into one record.

What it means for a buyer

If you exchange mask sign-off records: a record built from per-piece records normally grows with the number of pieces. This one stayed within a narrow size range in the lab’s tests, with a short checker that uses only standard libraries. Its size does not buy assurance: that comes from spot-checking pieces.

Who we expect would buy

Teams we expect would care (no customer or pilot yet): mask-synthesis and mask-data-preparation tool vendors and foundry sign-off teams that exchange sign-off records.

Why now

Simulating how masks will print already costs the industry tens of billions of processor hours a year, by NVIDIA’s account when it launched its GPU lithography library with ASML, TSMC and Synopsys (press release). This result does not reduce that compute: it concerns the size of the sign-off record a team hands over.

Why you can trust the check

The lab’s check confirms that the two files it cites carry the claimed figures; it does not rebuild the record, and no tool outside the lab is involved.

No outside firm has audited it. How this result’s check works, step by step.

What this does not show yet

  • The record on its own proves nothing: assurance comes only from opening spot-checked pieces, and the number needed grows with how small a chance of missing a bad piece you want. Opening a few pieces gives little assurance against a single corrupted one; the exact bound is on the formal statements page.
  • Constant-size summaries over many records are standard since hash trees. The lab’s own record names a published design for authenticated data structures (Miller, Hicks, Katz and Shi, POPL 2014) as a risk to the novelty of the idea; what is left as new, the lab says, is how its four pass/fail checks combine.
  • Three fixed defects travel with the record; they are listed in the lab’s exact wording.
  • The lab’s chiplet command-line tool separately prints a compression ratio that is a fixed number in its code, not a measurement. The measured ratio is in the lab’s exact wording.

Prior work

Named in the lab’s prior-art search for this result, and credited here.

The exact wording, for a technical reader

The lab’s own sentences and figures for this result, word for word, its limits in full and its formal statement: A near-constant-size sign-off record for a photomask, exact wording. The formal statements of all the results are on one page.

Check it yourself

  • This result’s file: every sentence and figure on this page that is the lab’s own, copied from its current record at the commit the file names.
  • The lab’s result file, copied from its codebase at the commit it names.
  • The file this result’s statement is checked against (a proof, a certificate or a measurement record; the formal statements page says which).
  • The formal statement, set as a formula.
  • ChipletOS, the company that carries this result.

All resultsContact / M&A

How we show numbers

Every number on this site links to the file it comes from. How each result is checked

  • We never show a number before its file has loaded.
  • A question we have not checked yet is marked as unchecked.
  • A check that found nothing says so.
  • A file with no value for a question says so.
  • A number whose file is missing or has changed is not shown.
  • Two files that disagree about what a number describes are both flagged.
  • A number from too few samples shows its sample size.
  • Two files that give different values are both shown.
  • A file we cannot publish is listed by its fingerprint only.
  • A measurement more than a week old shows its age.
  • A question that does not apply to a page is left off it.
  • A measurement whose program failed is shown as failed.