fix: tree heights that last past 3 days #9760
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The L1-L2 message tree height was a bottleneck running ``` post-mortem of 1-validator network (bot set to 0.05 TPS, 1 private / 2 public transfers per tx) Lasted long, got to block 4091, last tried to propose block 4097 Hit issues and did not reorg past them Root issue (guess): 2024-11-05 08:32:53.148 Error assembling block: 'Error: Failed to append leaves: Tree is full' ``` Automated test is a bit awkward here or I'd write one. It'd either trivially pass or have to go through 3-days worth of transactions.
ludamad
added
A-installation
Changes to circuit sizes
🧾 Summary (100% most significant diffs)
Full diff report 👇
|
ludamad
requested review from
jeanmon,
IlyasRidhuan and
fcarreiro
as code owners
…' into ad/fix-tree-heights-crash-3-days
fcarreiro
requested review from
dbanks12
and removed request for
fcarreiro,
IlyasRidhuan and
jeanmon
dbanks12
approved these changes
Nov 7, 2024
There was a problem hiding this comment.
The resultant bumps in avm gas are fine with me! Those opcodes trigger merkle checks, so I made them scale with tree heights. The scaling factors are somewhat arbitrary at the moment.
I don't know if someone else should review who has a better understanding of exactly what our tree height requirements are, but I'm happy to approve
|
The kind smoke test failure I think was fixed by #9790 Maybe changing all the constants at once was too ambitious? Maybe just start with the l1-l2 tree? I'm just concerned about the time needed to track down the places where the other values are hardcoded/assumed. |
|
I systematically went through the hardcoded ones now that had a comment above them with the formula and hopefully that's enough (there was 1, alerted to by the e2e-authwit test). If not I can redo this - but note a redo is a surprising amount of faff unless I'm doing it wrong |
ludamad
added a commit
that referenced
this pull request
Nov 11, 2024
The L1-L2 message tree height was a bottleneck running ``` post-mortem of 1-validator network (bot set to 0.05 TPS, 1 private / 2 public transfers per tx) Lasted long, got to block 4091, last tried to propose block 4097 Hit issues and did not reorg past them Root issue (guess): 2024-11-05 08:32:53.148 Error assembling block: 'Error: Failed to append leaves: Tree is full' ``` Also updated tree heights with constants proposed by @iAmMichaelConnor here (#9451) (thanks for the thoughtful analysis I could lazily steal! Automated test is a bit awkward here or I'd write one. It'd either trivially pass or have to go through 3-days worth of transactions.
TomAFrench
added a commit
that referenced
this pull request
Nov 11, 2024
* master: (182 commits) feat(avm): mem specific range check (#9828) refactor: remove public kernel inner (#9865) chore: Revert "chore: Validate RPC inputs" (#9875) Revert "fix: deploy l2 contracts fails on 48 validator" (#9871) fix: deploy l2 contracts fails on 48 validator (#9860) chore: Validate RPC inputs (#9672) fix: fixing devcontainers to use the sandbox docker-compose file (#9782) fix: Revert changes to ci.yml (#9863) chore: Move epoch and slot durations to config (#9861) git subrepo push --branch=master noir-projects/aztec-nr git_subrepo.sh: Fix parent in .gitrepo file. [skip ci] chore: replace relative paths to noir-protocol-circuits git subrepo push --branch=master barretenberg fix: tree heights that last past 3 days (#9760) fix(build): l1-contracts .rebuild_patterns did not cover test files (#9862) fix: bench prover test (#9856) fix: Fix mac build by calling `count` on durations (#9855) feat: zk shplemini (#9830) feat: domain separate block proposals and attestations (#9842) chore: bump runner cache disk size (#9849) ...
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The L1-L2 message tree height was a bottleneck running
Also updated tree heights with constants proposed by @iAmMichaelConnor here (#9451) (thanks for the thoughtful analysis I could lazily steal!
Automated test is a bit awkward here or I'd write one. It'd either trivially pass or have to go through 3-days worth of transactions.