fix: lay plumbing for having simulations throw an error if they cannot be added in a block

[sklppy88]

This PR gives an option to make a simulation of a tx fail, if the tx cannot be included in a block and added to the state.

To minimize impact on callsites / disruption, this update does not make this validation necessary, and defaults to skipping tx validation, exhibiting previous non-validated behavior. This will be changed to be default behavior in the future.

e.g. If a simulation produces duplicate nullifiers, or nullifiers that already exist in a state tree, the results of this simulation should not be returned, but should warn users that the transaction simulated is impossible to actually be added to a block due to being an invalid transaction.

The method for achieving the above is that a new API on the node was created, used for validating the correctness of the metadata and side-effects produced by a transaction. A transaction is deemed valid if and only if the transaction can be added to a block that can be used to advance state.

Another note just for reference: there is weirdness in e2e_prover, that fails the proof validation.

[sklppy88]

• fix: lay plumbing for having simulations throw an error if they cannot be added in a block #7839 : 2 dependent PRs (#7840 , #7951 ) 👈
• master

This stack of pull requests is managed by Graphite. Learn more about stacking.

Join @sklppy88 and the rest of your teammates on Graphite

[AztecBot]

Benchmark results

Metrics with a significant change:

• avm_simulation_time_ms (Token:mint_public): 43.5 (-42%)
• avm_simulation_time_ms (Token:assert_minter_and_mint): 73.2 (+88%)

Detailed results

All benchmarks are run on txs on the Benchmarking contract on the repository. Each tx consists of a batch call to create_note and increment_balance, which guarantees that each tx has a private call, a nested private call, a public call, and a nested public call, as well as an emitted private note, an unencrypted log, and public storage read and write.

This benchmark source data is available in JSON format on S3 here.

Proof generation

Each column represents the number of threads used in proof generation.

Metric	1 threads	4 threads	16 threads	32 threads	64 threads
proof_construction_time_sha256_ms	5,766 (-1%)	1,551	710 (+2%)	754 (-1%)	770 (+1%)
proof_construction_time_sha256_30_ms	11,838	3,168 (-1%)	1,410	1,424 (-3%)	1,472
proof_construction_time_sha256_100_ms	44,262	12,537 (+2%)	5,645 (-1%)	5,558	5,498
proof_construction_time_poseidon_hash_ms	78.0	34.0 (-29%)	34.0	59.0 (+2%)	87.0 (-1%)
proof_construction_time_poseidon_hash_30_ms	1,531 (-1%)	420	201	238 (+3%)	272 (+1%)
proof_construction_time_poseidon_hash_100_ms	5,661	1,531	683 (+1%)	746 (+3%)	748 (-2%)

L2 block published to L1

Each column represents the number of txs on an L2 block published to L1.

Metric	4 txs	8 txs	16 txs
l1_rollup_calldata_size_in_bytes	4,324	7,844	14,852
l1_rollup_calldata_gas	49,720	92,472	177,596
l1_rollup_execution_gas	1,294,383	2,042,103	3,869,266
l2_block_processing_time_in_ms	255 (-2%)	449 (+1%)	799 (-4%)
l2_block_building_time_in_ms	9,351	18,088 (-1%)	36,349
l2_block_rollup_simulation_time_in_ms	9,351	18,088 (-1%)	36,349
l2_block_public_tx_process_time_in_ms	7,881	16,545 (-1%)	34,761

L2 chain processing

Each column represents the number of blocks on the L2 chain where each block has 8 txs.

Metric	3 blocks	5 blocks
node_history_sync_time_in_ms	2,979 (-1%)	3,935 (+3%)
node_database_size_in_bytes	12,771,408 (+1%)	16,838,736
pxe_database_size_in_bytes	16,254	26,813

Circuits stats

Stats on running time and I/O sizes collected for every kernel circuit run across all benchmarks.

Circuit	simulation_time_in_ms	witness_generation_time_in_ms	input_size_in_bytes	output_size_in_bytes	proving_time_in_ms
private-kernel-init	101 (+6%)	392 (-2%)	21,846	44,858	N/A
private-kernel-inner	169 (-3%)	694 (-2%)	72,545	45,005	N/A
private-kernel-reset-tiny	513 (+2%)	1,193 (-2%)	69,614	44,844	N/A
private-kernel-tail	273 (+1%)	194 (-1%)	50,760	52,256	N/A
base-parity	5.49 (-1%)	N/A	160	96.0	N/A
root-parity	33.4	N/A	69,084	96.0	N/A
base-rollup	2,853 (-1%)	N/A	187,817	664	N/A
root-rollup	38.4	N/A	54,525	716	N/A
public-kernel-setup	98.5 (-1%)	N/A	103,760	71,222	N/A
public-kernel-app-logic	106	N/A	103,599	71,222	N/A
public-kernel-tail	569 (-1%)	N/A	409,190	16,414	N/A
private-kernel-reset-small	493 (+2%)	N/A	66,533	45,629	N/A
private-kernel-tail-to-public	1,487 (+11%)	704 (-1%)	507,554	1,697	N/A
public-kernel-teardown	94.2 (-1%)	N/A	104,005	71,222	N/A
merge-rollup	19.1	N/A	35,742	664	N/A
undefined	N/A	N/A	N/A	N/A	69,194 (+3%)

Stats on running time collected for app circuits

Function	input_size_in_bytes	output_size_in_bytes	witness_generation_time_in_ms
ContractClassRegisterer:register	1,344	11,731	347 (+1%)
ContractInstanceDeployer:deploy	1,408	11,731	18.1 (-1%)
MultiCallEntrypoint:entrypoint	1,920	11,731	460
FeeJuice:deploy	1,376	11,731	393
SchnorrAccount:constructor	1,312	11,731	169 (-2%)
SchnorrAccount:entrypoint	2,304	11,731	504 (-1%)
Token:privately_mint_private_note	1,280	11,731	215
FPC:fee_entrypoint_public	1,344	11,731	25.4 (+2%)
Token:transfer	1,312	11,731	403 (-1%)
Benchmarking:create_note	1,344	11,731	166
SchnorrAccount:verify_private_authwit	1,280	11,731	27.3
Token:unshield	1,376	11,731	626
FPC:fee_entrypoint_private	1,376	11,731	847

AVM Simulation

Time to simulate various public functions in the AVM.

Function	time_ms	bytecode_size_in_bytes
FeeJuice:_increase_public_balance	55.0 (+2%)	8,139
FeeJuice:set_portal	9.98 (+2%)	2,362
Token:constructor	80.1 (-3%)	31,107
FPC:constructor	52.6 (+2%)	22,380
FeeJuice:mint_public	38.8 (+3%)	6,150
Token:mint_public	⚠️ 43.5 (-42%)	11,720
Token:assert_minter_and_mint	⚠️ 73.2 (+88%)	8,028
AuthRegistry:set_authorized	37.3 (+4%)	4,537
FPC:prepare_fee	234 (+2%)	8,812
Token:transfer_public	29.6 (-9%)	47,374
FPC:pay_refund	56.3 (-11%)	12,114
Benchmarking:increment_balance	970 (-1%)	7,450
Token:_increase_public_balance	38.1 (-4%)	8,960
FPC:pay_refund_with_shielded_rebate	68.0 (+2%)	12,663

Public DB Access

Time to access various public DBs.

Function	time_ms
get-nullifier-index	0.161 (-2%)

Tree insertion stats

The duration to insert a fixed batch of leaves into each tree type.

Metric	1 leaves	16 leaves	64 leaves	128 leaves	256 leaves	512 leaves	1024 leaves
batch_insert_into_append_only_tree_16_depth_ms	2.21 (+1%)	3.85 (-1%)	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_count	16.8	31.7	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_ms	0.114 (+1%)	0.109 (-1%)	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_32_depth_ms	N/A	N/A	11.4 (-1%)	18.0 (+1%)	30.9 (-2%)	60.4	112 (-3%)
batch_insert_into_append_only_tree_32_depth_hash_count	N/A	N/A	95.9	159	287	543	1,055
batch_insert_into_append_only_tree_32_depth_hash_ms	N/A	N/A	0.109 (-1%)	0.105	0.100 (-3%)	0.104	0.101 (-3%)
batch_insert_into_indexed_tree_20_depth_ms	N/A	N/A	14.3	25.9 (-1%)	43.6 (-1%)	83.6 (-1%)	161 (-2%)
batch_insert_into_indexed_tree_20_depth_hash_count	N/A	N/A	109	207	355	691	1,363
batch_insert_into_indexed_tree_20_depth_hash_ms	N/A	N/A	0.109 (+1%)	0.104 (-2%)	0.105 (+1%)	0.103 (-1%)	0.102 (-3%)
batch_insert_into_indexed_tree_40_depth_ms	N/A	N/A	16.4 (-1%)	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_count	N/A	N/A	132	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_ms	N/A	N/A	0.105 (-1%)	N/A	N/A	N/A	N/A

Miscellaneous

Transaction sizes based on how many contract classes are registered in the tx.

Metric	0 registered classes	1 registered classes
tx_size_in_bytes	64,779	668,997

Transaction size based on fee payment method

| Metric | |
| - | |

[nventuro]

Why go with an offline default? To minimize disruption?

[sklppy88]

Yep, exactly that.

[nventuro]

Looks good! Included a minor naming suggestion, as I was left wondering for a bit what the method actually did. I'm now thinking too that perhaps offline is not the best term, since we are accesing the node to e.g. fetch public data, merkle paths, etc.

[sklppy88]

Looks good! Included a minor naming suggestion, as I was left wondering for a bit what the method actually did. I'm now thinking too that perhaps offline is not the best term, since we are accesing the node to e.g. fetch public data, merkle paths, etc.

Hmmm, very fair. I have changed it to ignoreInvalidTx. Wdyt about that ?

[nventuro]

Looks good!