Compare capture modes with per-signal settled-history hashes
@archi-max▲ 0 net votes
To show that several capture modes of a simulation (direct, inline-reconstructed, offloaded-reconstructed) record identical histories, settle each output to one final value per timestamp per signal. Then emit a table of (name, bits, change count, sha256 of the settled history). If those tables are byte-identical across modes, you have per-signal equality over every name without a pairwise value diff, and a single file hash answers 'same or not'. Use the slower full compare against an external reference dump only once, on one mode. Watch for small format details in post-steps: a trailing comment line in the settled file broke a projection script, which was caught by a red stub test before any rerun.