Skip to content

Forge-stress solo/chain-fragment profiles and beacon export tooling#6627

Open
Russoul wants to merge 5 commits into
masterfrom
mkarg/solochain-profiles
Open

Forge-stress solo/chain-fragment profiles and beacon export tooling#6627
Russoul wants to merge 5 commits into
masterfrom
mkarg/solochain-profiles

Conversation

@Russoul

@Russoul Russoul commented Jul 24, 2026

Copy link
Copy Markdown
Contributor

Summary

  • Add forge-stress chain-fragment-creation profiles (fschain-768k-xs, fschain-768k, fschain-8io) and forge-stress-pre-plutus-solo, and switch the forge-stress solo profiles to Voltaire genesis so they can double as chain-fragment sources.
  • Add forge-stress-plutusv3-solo, mirroring forge-stress-plutus-solo but using the PlutusV3 LoopV3 script, tuned to 0.30 tps so blocks reach their per-block PlutusV3 execution-unit budget cap right away instead of ramping up over the first third of the run.
  • Add a wb chain-fragment export op: exports a profile run's node db/, genesis files (untrimmed shelley genesis in place of the trimmed one) and a path-patched config.json as a chain fragment consumable by ouroboros-consensus-tools' beacon tool, registering a minimal entry in its chain-register.json.

mgmeier and others added 4 commits July 21, 2026 13:40
Adds forge-stress-plutusv3-solo, mirroring forge-stress-plutus-solo but
using the PlutusV3 "LoopV3" script, and regenerates all-profiles.json
(which was already stale relative to the prior solo-profiles genesis
switch to Voltaire, in addition to missing this new profile).
Exports a profile run's node db/, genesis files (untrimmed shelley
genesis in place of the trimmed one), and a config.json with genesis
paths rewritten, into an ouroboros-consensus-tools beacon-data/chain/
directory usable by the `beacon` tool, registering a minimal entry in
chain-register.json. Mirrors chaindb.sh's structure/conventions.
….30 tps

chain.json from a run showed blocks well under the per-block PlutusV3
execution-unit budget cap (4 tx/block) for roughly the first third of
the run: at 0.20 tps the mempool only just keeps pace with that cap on
average, so it takes many blocks to refill after starting empty.
Verified with a fresh run: at 0.30 tps every block past the initial
fund-splitting one hits the steady 4-tx/block plateau immediately.
@Russoul
Russoul requested review from a team as code owners July 24, 2026 08:12
Declares itself bash (fixing SC2148), splits declare-from-assign to
avoid masking $? (SC2155), and documents the SC2016 on the jq --arg
filter as a false positive rather than pulling in the exceptions list.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants