Keyboard shortcuts

Press or to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Evidence ladder

Current completion boundary: generated RTL is the highest passed signer-performance tier. The U280 250 MHz and 300 MHz synthesis and route records, plus programmed-card execution, are pending. Until those exact artifact chains pass, do not describe the project as synthesized, routed, timed at either target clock, card-validated, or hardware-complete.

TierWhat it can establishWhat remains outside the tier
0. ModelDependency arithmetic, bounds, or a software oracleRHDL behavior and all implementation claims
1. Source simulationRust/RHDL behavior and modeled cycle eventsEmitted RTL, mapped resources, timing, route, hardware
2. Generated RTLLowering equivalence for an executed traceFormal proof, synthesis, clock, mapping, route, card
3. U280 synthesisMapping and estimated timing for an exact top and constraintPlacement, route, shell, physical clock, card rate
4. U280 routeResource and timing of a completely routed exact designShell/card behavior unless included and executed
5. Card executionValidated runtime behavior and throughput for the tested setupUntested workloads, clocks, shells, or revisions

Promotion moves one exact source and artifact chain upward. It cannot combine a cycle interval from one commit with timing from another, or a routed core from one top with a shell from another.

Negative results

A failed synthesis, placement, or route attempt remains useful when it names the exact source, tool, constraint, failure, and reports. Retaining it prevents future readers from treating “Vivado exited” as “timing closed.”

Dirty source

Dirty-source evidence is permitted only when every differing source file and checksum is recorded. A clean source commit is preferred for promotion and deployment. The manifest displays dirty state rather than hiding it.

Four-state and formal boundaries

Verilator’s ordinary two-state simulation does not prove unknown propagation. A deterministic testbench is not a formal proof. These limitations remain nonclaims even when every expected output in that trace matches.