feat: merge-train/avm#18867
Merged
Merged
Conversation
…18747) [Engineering Design Document](https://www.notion.so/aztecnetwork/E2E-AVM-Circuit-Inputs-Collection-Testing-2bea1f6b0e35809384e1fad21691c59d) To gain confidence in the AVM circuit (its completeness in particular), we would like to prove all end-to-end tests. Changing all of those tests to run provers is a huge task and would lead to very slow tests. Instead, we can take a much simpler approach of optionally dumping all AvmCircuitInputs (basically the proving requests for the AVM) to a tarball that gets cached. This can be wrapped in a CI job that runs on `next` commits, and paired with another job that iterates over all such AvmCircuitInputs and proves (or just check-circuits) them. This will help us gain confidence in the AVM circuit, add a layer of regression testing, and provide more transactions to feed our fuzzing corpus. **Proposed solution**: Add a `DumpingCppPublicTxSimulator` that extends the base simulator and writes circuit inputs to disk after each simulation. Control via an environment variable `DUMP_AVM_INPUTS_TO_DIR`. Run nightly in CI, package results, and upload to S3 cache. Run another nightly CI job to download from cache and prove (or probably just check-circuit) all collected inputs.
AztecBot
requested review from
charlielye,
dbanks12,
fcarreiro and
sirasistant
as code owners
In fast simulation, don't hash bytecode and don't update check unless asking for hints. Solves #18161 and https://linear.app/aztec-labs/issue/AVM-12/disable-bytecode-hashing-during-fast-simulation. The following bench was just running before and after once, so don't trust it a lot. We'll see the real change when this lands on next. # Performance Comparison: Before vs After | TX | Before (ms) | After (ms) | Difference % | | ------ | ----------- | ---------- | ------------ | | TX 1 | 45.6205 | 32.5902 | -28.56% | | TX 2 | 45.0508 | 14.1821 | -68.52% | | TX 3 | 43.4434 | 15.7217 | -63.81% | | TX 4 | 27.1896 | 16.8379 | -38.07% | | TX 5 | 14.2418 | 15.0181 | 5.45% | | TX 6 | 35.6377 | 15.7467 | -55.81% | | TX 7 | 25.9184 | 15.5752 | -39.91% | | TX 8 | 39.0134 | 15.1753 | -61.10% | | TX 9 | 27.8695 | 14.4692 | -48.08% | | TX 10 | 59.7841 | 14.3228 | -76.04% | | TX 11 | 57.9308 | 14.8002 | -74.45% | | TX 12 | 35.5910 | 15.8639 | -55.43% | | TX 13 | 15.5797 | 13.0764 | -16.07% | | TX 14 | 12.9283 | 61.8159 | 378.14% | | TX 15 | 13.6121 | 50.7118 | 272.55% | | TX 16 | 12.9670 | 34.7158 | 167.72% | | TX 17 | 14.1347 | 13.3011 | -5.90% | | TX 18 | 12.2613 | 12.6620 | 3.27% | | TX 19 | 13.7812 | 12.0825 | -12.33% | | TX 20 | 13.6484 | 13.0992 | -4.02% | | TX 21 | 13.4302 | 11.6249 | -13.44% | | TX 22 | 111.1149 | 10.3152 | -90.72% | | TX 23 | 40.7040 | 12.4170 | -69.50% | | TX 24 | 15.7238 | 12.3420 | -21.51% | | TX 25 | 11.2228 | 12.7231 | 13.37% | | TX 26 | 14.4293 | 33.2565 | 130.48% | | TX 27 | 13.1613 | 14.6755 | 11.51% | | TX 28 | 13.3343 | 11.8471 | -11.15% | | TX 29 | 15.0849 | 13.6830 | -9.29% | | TX 30 | 12.7343 | 11.2850 | -11.38% | | TX 31 | 45.3641 | 11.3602 | -74.96% | | TX 32 | 14.6793 | 13.8814 | -5.44% | | TX 33 | 15.1199 | 11.1683 | -26.13% | | TX 34 | 13.7063 | 18.0565 | 31.74% | | TX 35 | 14.3068 | 16.4247 | 14.80% | | TX 36 | 14.9269 | 15.1366 | 1.40% | | TX 37 | 15.4937 | 91.5274 | 490.74% | | TX 38 | 15.7002 | 32.1775 | 104.95% | | TX 39 | 15.1609 | 34.0217 | 124.40% | | TX 40 | 15.6692 | 32.1142 | 104.95% | | TX 41 | 16.5157 | 19.6163 | 18.77% | | TX 42 | 16.2124 | 13.6257 | -15.95% | | TX 43 | 17.9787 | 30.9630 | 72.22% | | TX 44 | 20.5974 | 23.4374 | 13.79% | | TX 45 | 17.3338 | 29.7317 | 71.52% | | TX 46 | 29.1956 | 13.2776 | -54.52% | | TX 47 | 30.5091 | 15.0963 | -50.52% | | TX 48 | 29.2279 | 15.7975 | -45.95% | | TX 49 | 29.5552 | 15.7246 | -46.80% | | TX 50 | 29.3517 | 15.9813 | -45.55% | | TX 51 | 30.3135 | 20.9737 | -30.81% | | TX 52 | 29.6178 | 16.3860 | -44.68% | | TX 53 | 31.7046 | 21.9473 | -30.78% | | TX 54 | 31.2743 | 19.2922 | -38.31% | | TX 55 | 31.4366 | 19.0294 | -39.47% | | TX 56 | 40.4035 | 27.2617 | -32.53% | | TX 57 | 43.0120 | 26.9692 | -37.30% | | TX 58 | 53.8648 | 51.7758 | -3.88% | | TX 59 | 56.3566 | 41.1194 | -27.04% | | TX 60 | 80.4779 | 68.8234 | -14.48% | | TX 61 | 108.5981 | 96.5275 | -11.11% | | TX 62 | 30.1644 | 14.7178 | -51.21% | | TX 63 | 116.4025 | 102.8303 | -11.66% | | TX 64 | 27.4064 | 14.6618 | -46.50% | | TX 65 | 26.8179 | 13.4844 | -49.72% | | TX 66 | 104.8157 | 96.1230 | -8.29% | | TX 67 | 87.5321 | 57.7488 | -34.02% | | TX 68 | 72.0782 | 58.4979 | -18.84% | | TX 69 | 44.7683 | 35.1704 | -21.44% | | TX 70 | 15.4529 | 13.9217 | -9.91% | | TX 71 | 12.0968 | 11.6362 | -3.81% | | TX 72 | 16.0817 | 13.9407 | -13.31% | | TX 73 | 12.6076 | 11.1379 | -11.66% | | TX 74 | 13.5392 | 10.6382 | -21.42% | | TX 75 | 16.1866 | 12.4156 | -23.28% | | TX 76 | 12.9790 | 10.6062 | -18.28% | | TX 77 | 41.2954 | 31.2786 | -24.25% | | TX 78 | 42.9297 | 33.5444 | -21.87% | | TX 79 | 41.3766 | 31.8146 | -23.11% | | TX 80 | 24.9970 | 18.8563 | -24.57% | | TX 81 | 15.0835 | 12.5931 | -16.52% | | TX 82 | 33.4667 | 29.3651 | -12.26% | | TX 83 | 26.6582 | 23.3770 | -12.31% | | TX 84 | 35.6922 | 30.4587 | -14.66% | | TX 85 | 106.9044 | 82.7466 | -22.60% | | TX 86 | 249.9548 | 242.4194 | -3.01% | | TX 87 | 58.4410 | 40.4346 | -30.81% | | TX 88 | 28.3157 | 14.2439 | -49.70% | | TX 89 | 27.2518 | 14.6348 | -46.30% | | TX 90 | 31.8427 | 15.0903 | -52.61% | | TX 91 | 27.6194 | 14.9010 | -46.05% | | TX 92 | 30.1784 | 15.1709 | -49.73% | | TX 93 | 30.1447 | 16.0628 | -46.71% | | TX 94 | 30.5910 | 17.4399 | -42.99% | | TX 95 | 31.9434 | 17.2256 | -46.07% | | TX 96 | 33.0917 | 17.7947 | -46.23% | | TX 97 | 32.9948 | 18.0287 | -45.36% | | TX 98 | 40.3504 | 27.3818 | -32.14% | | TX 99 | 40.4654 | 26.2407 | -35.15% | | TX 100 | 54.7813 | 40.5642 | -25.95% | | TX 101 | 58.3964 | 40.3257 | -30.95% | | TX 102 | 83.9649 | 67.0056 | -20.20% | | TX 103 | 107.8069 | 94.8606 | -12.01% | | TX 104 | 29.5196 | 15.4562 | -47.64% | | TX 105 | 113.7004 | 102.0835 | -10.22% | | TX 106 | 25.9757 | 12.7171 | -51.04% | | TX 107 | 26.7753 | 11.7045 | -56.29% | | TX 108 | 105.3427 | 91.6692 | -12.98% | | TX 109 | 73.3440 | 58.8163 | -19.81% | | TX 110 | 73.1386 | 58.2437 | -20.36% |
We have a better way to do this, so use it.
fcarreiro
approved these changes
Dec 8, 2025
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
BEGIN_COMMIT_OVERRIDE
test(avm): e2e tests can optionally dump avm circuit inputs to files (#18747)
feat(avm): don't hash bytecode in fast simulation (#18853)
chore(avm): use ContractDeploymentData.fromPlainObject (#18855)
chore(avm): earlier bytecode deduplication
END_COMMIT_OVERRIDE