feat: router contract

[benesjan]

Implementation of a router contract.

[benesjan]

• feat: router contract #8352 👈
• master

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

Join @benesjan and the rest of your teammates on Graphite

[AztecBot]

Benchmark results

Metrics with a significant change:

• protocol_circuit_simulation_time_in_ms (private-kernel-tail-to-public): 1,398 (+82%)
• avm_simulation_time_ms (Token:_increase_public_balance): 54.6 (+37%)

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	5,772	1,563	718 (+1%)	758 (+2%)	777 (+1%)
proof_construction_time_sha256_30_ms	11,500	3,099 (+1%)	1,397 (+1%)	1,442 (+1%)	1,473
proof_construction_time_sha256_100_ms	44,183	11,840	5,480 (+1%)	5,466 (+1%)	5,724 (-2%)
proof_construction_time_poseidon_hash_ms	78.0	34.0	34.0	57.0 (-2%)	86.0 (-2%)
proof_construction_time_poseidon_hash_30_ms	1,530	423	204	222 (-2%)	269
proof_construction_time_poseidon_hash_100_ms	5,654	1,517	678	728 (-1%)	742 (-1%)

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	4,356	7,876	14,884
l1_rollup_calldata_gas	50,232	92,998	178,168
l1_rollup_execution_gas	843,073	1,576,735	3,361,961
l2_block_processing_time_in_ms	257	444	813
l2_block_building_time_in_ms	11,159	22,084 (+1%)	43,911
l2_block_rollup_simulation_time_in_ms	11,159	22,083 (+1%)	43,911
l2_block_public_tx_process_time_in_ms	9,524	20,415 (+1%)	42,214

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	2,936 (-1%)	3,906
node_database_size_in_bytes	12,599,376	16,715,856
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	79.8 (+8%)	387 (+2%)	21,391	44,898	N/A
private-kernel-inner	157 (-2%)	699 (+3%)	72,372	45,020	N/A
private-kernel-reset-tiny	297	702 (+1%)	65,097	44,886	N/A
private-kernel-tail	160 (+1%)	132 (-1%)	50,608	52,921	N/A
base-parity	5.56	N/A	160	96.0	N/A
root-parity	35.0	N/A	73,948	96.0	N/A
base-rollup	3,101	N/A	189,207	664	N/A
block-root-rollup	41.2	N/A	58,205	2,448	N/A
public-kernel-setup	81.0 (-1%)	N/A	105,085	71,222	N/A
public-kernel-app-logic	94.5	N/A	104,899	71,222	N/A
public-kernel-tail	841	N/A	390,582	16,414	N/A
private-kernel-reset-small	294	N/A	66,341	45,629	N/A
private-kernel-tail-to-public	⚠️ 1,398 (+82%)	596 (+1%)	453,751	1,825	N/A
public-kernel-teardown	80.6 (-1%)	N/A	105,349	71,222	N/A
merge-rollup	20.1	N/A	38,174	664	N/A
undefined	N/A	N/A	N/A	N/A	85,712 (+2%)

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	344
ContractInstanceDeployer:deploy	1,408	11,731	18.2 (+1%)
MultiCallEntrypoint:entrypoint	1,920	11,731	403 (+1%)
FeeJuice:deploy	1,376	11,731	387
SchnorrAccount:constructor	1,312	11,731	63.0 (+1%)
SchnorrAccount:entrypoint	2,336	11,731	370
FeeJuice:claim	1,344	11,731	35.9
Token:privately_mint_private_note	1,280	11,731	76.7 (+1%)
FPC:fee_entrypoint_public	1,344	11,731	30.0 (+25%)
Token:transfer	1,312	11,731	223 (+4%)
Benchmarking:create_note	1,344	11,731	75.7 (+2%)
SchnorrAccount:verify_private_authwit	1,280	11,731	26.3 (-1%)
Token:unshield	1,376	11,731	497 (-1%)
FPC:fee_entrypoint_private	1,376	11,731	658 (-5%)

AVM Simulation

Time to simulate various public functions in the AVM.

Function	time_ms	bytecode_size_in_bytes
FeeJuice:_increase_public_balance	45.1	2,431
FeeJuice:set_portal	12.8 (-1%)	2,930
Token:constructor	77.5 (+4%)	15,813
FPC:constructor	49.7	9,720
FeeJuice:check_balance	37.4 (+2%)	2,357
Token:mint_public	48.1 (-2%)	5,011
Token:assert_minter_and_mint	471	3,155
AuthRegistry:set_authorized	36.7 (-21%)	924
FPC:prepare_fee	75.9	4,245
Token:transfer_public	16.0 (-24%)	10,812
FPC:pay_refund	43.0 (-1%)	5,195
Benchmarking:increment_balance	1,205 (+1%)	3,496
Token:_increase_public_balance	⚠️ 54.6 (+37%)	2,408
FPC:pay_refund_with_shielded_rebate	53.4 (+6%)	5,195

Public DB Access

Time to access various public DBs.

Function	time_ms
get-nullifier-index	0.151 (-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	2.16 (-1%)	3.91 (+2%)	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	0.112 (-1%)	0.110 (+3%)	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_32_depth_ms	N/A	N/A	11.3 (+2%)	17.7 (-1%)	30.7 (-1%)	59.7 (+1%)	114
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	0.108 (+2%)	0.103 (-1%)	0.0997 (-2%)	0.103 (+2%)	0.102 (-1%)
batch_insert_into_indexed_tree_20_depth_ms	N/A	N/A	14.3	25.5 (-1%)	43.4 (-2%)	83.3 (+2%)	163 (-1%)
batch_insert_into_indexed_tree_20_depth_hash_count	N/A	N/A	107 (-1%)	207	355	691	1,363
batch_insert_into_indexed_tree_20_depth_hash_ms	N/A	N/A	0.110	0.102 (-2%)	0.105 (-2%)	0.104 (+2%)	0.102 (-3%)
batch_insert_into_indexed_tree_40_depth_ms	N/A	N/A	16.5 (+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	0.106	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	64,838	668,997

Transaction size based on fee payment method

| Metric | |
| - | |

[benesjan]

Renamed collapse function to collapse_array as utils::collapse::colapse(...) looked weird.

[nventuro]

You can alternatively do use collapse:collapse in utils/mod.nr and then import as utils::collapse. That's what we do for e.g. PrivateContext etc.

I do agree with collapse_array potentially being a better name, but if so I think I'd also rename the file as well, as collapse::collapse_array is sorta janky.

[benesjan]

Created this utility function as now it's needed in more places than just note_getter.

[nventuro]

I'd change this to compare(lhs, op, rhs) -> bool, and we'd then do assert(compare(balance, GT, amount), "insufficient balance"), which is almost exactly equivalent but also lets is not have to worry about this receiving a string, and provides flexibility in case we want to do something else instead of assert.

[benesjan]

Replaced PublicImmutable with SharedImmutable so that I am able to load the value in private and then pass it to the router contract.

[rahul-kothari]

if we keep router to be public contract only, then we can keep this public and save on a kernel iteration + hashing cost for a shared mutable

[nventuro]

What do you mean by 'public only'? It has a private component so that the caller is not revelead in public. Though I guess nothing really prevents us from providing a way to not set msg.sender when enqueuing a public call, and therefore save this kernel iteration by calling the pub fn directly from the app - is this what you meant?

[nventuro]

At any rate @benesjan's change here is correct - we can't read this variable publicly because we'd be leaking what we're reading, it has to be read privately, even if we had the ability to enqueue to public without revealing the sender.

[AztecBot]

Docs Preview

Hey there! 👋 You can check your preview at https://66d8c4f95cc752445d4a25ed--aztec-docs-dev.netlify.app

[rahul-kothari]

could we skip an interation of the kernel by the main contract directly calling _check_timestamp_public()

[rahul-kothari]

it obviously leaks the original contract thoughh hmmph..

[benesjan]

That would reveal the address of a contract which called it so no. To allow for that we would need to allow directly setting of msg_sender to zero which as far as I know is not planned (I am not up-to-date on this discussion).

[Thunkar]

LGTM...but

TXE tests plz? Timestamps are trickier, but block_num should be easy 😄 (if it isn't, we have to fix it!)

[benesjan]

TXE tests plz? Timestamps are trickier, but block_num should be easy 😄 (if it isn't, we have to fix it!)

@Thunkar hoped nobody will notice the missing unit tests here 🤣 Will add.

[nventuro]

You can alternatively do use collapse:collapse in utils/mod.nr and then import as utils::collapse. That's what we do for e.g. PrivateContext etc.

I do agree with collapse_array potentially being a better name, but if so I think I'd also rename the file as well, as collapse::collapse_array is sorta janky.

[nventuro]

What do you mean by 'public only'? It has a private component so that the caller is not revelead in public. Though I guess nothing really prevents us from providing a way to not set msg.sender when enqueuing a public call, and therefore save this kernel iteration by calling the pub fn directly from the app - is this what you meant?

[nventuro]

At any rate @benesjan's change here is correct - we can't read this variable publicly because we'd be leaking what we're reading, it has to be read privately, even if we had the ability to enqueue to public without revealing the sender.