Layer1Labs is raising seed funding — $1–1.5M on a SAFE, seeking a lead investor. Talk to us →
Document · ChronoFabric

ChronoFabric Measured Results

What ChronoFabric measured on Kria FPGA boards — every figure with the conditions it was measured under, the equivalence runs labeled as equivalence, and the limits stated plainly.

Version 1.0 · October 26, 2026 · PDF · 12 pages · Layer1Labs Silicon Inc.

Key figures

21,000+

scheduled packets delivered with zero missed departures and zero dropped — 100% delivery, including a continuous 10,000-packet endurance run

Measured on hardware — two Kria boards, 10-gigabit link

1,000,000 ns

the median and 99th-percentile departure interval — exactly the programmed 1 ms tick, across 2,000-packet runs

Measured on hardware

99.82%

of scheduled packets on target under a flood of 50,000 competing packets, still departing on the programmed tick

Measured on hardware

620 ns

the tightest timing budget validated in the equivalence runs — the 1.6-terabit cadence; the schedule held exactly, and no transport at that rate is claimed

Equivalence-validated on hardware

What it covers

  1. How to read these results
  2. The headline result
  3. Delivery, run by run
  4. Departure precision
  5. Scheduling under load
  6. Timing-budget equivalence at higher-rate cadences
  7. The 24-hour runs
  8. Equivalence at the 400G budget, on the two-board link
  9. The results at a glance
  10. Limits, stated plainly
  11. What these results do not establish

Basis

This document collects ChronoFabric’s results as measured on Kria FPGA boards, each printed with its basis: the setup, the run, and the label the result carries. Equivalence runs are labeled as equivalence — a timing budget exercised on the hardware that exists, never a measurement at the faster line rate.

Modeled scale figures are not results and do not appear. Figures from different runs are never stacked, and the closing sections state what these results do not establish.