feat(avm/public): user space PublicContext::get_args_hash

[fcarreiro]

This PR implements PublicContext::get_args_hash in user space. We are still passing the calldata length as a runtime variable until we can get it at compile time. This requires @Thunkar 's work on aztec(public) as a macro.

Once that is done, we'll pass the hasher as a closure when creating the PublicContext, i.e.:

struct PublicContext {
    hash_getter: fn[(Field,)]() -> Field,
    // ...
}

impl PublicContext {
    pub fn new(..., hash_getter) -> Self {
        // ...
    }

    fn get_args_hash(self) -> Field {
        (self.hash_getter)()
    }
}

// In the aztec(public) macro
comptime let N = get_calldata_length();
let hash_getter = || {
    let mut hasher = ArgsHasher::new();
    let mut fields = std::meta::unquote!(quote { [0; $N] });
    fields = calldata_copy(2 /*or 1*/, N);
    hasher.add_many(fields);
    hasher.hash()
};
let context = PublicContext::new(..., hash_getter);

[fcarreiro]

• feat(avm/public): user space PublicContext::get_args_hash #8292 👈
• master

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

Join @fcarreiro and the rest of your teammates on Graphite

[AztecBot]

Benchmark results

Metrics with a significant change:

• avm_simulation_time_ms (FeeJuice:_increase_public_balance): 159 (+231%)
• avm_simulation_time_ms (Token:constructor): 279 (+281%)
• avm_simulation_time_ms (FPC:constructor): 81.5 (+80%)
• avm_simulation_time_ms (Token:mint_public): 307 (+519%)
• avm_simulation_time_ms (Token:assert_minter_and_mint): 50.9 (-92%)
• avm_simulation_time_ms (FPC:prepare_fee): 118 (+80%)
• avm_simulation_time_ms (Token:transfer_public): 36.3 (+181%)
• avm_simulation_bytecode_size_in_bytes (Token:constructor): 19,994 (+117%)
• avm_simulation_bytecode_size_in_bytes (FPC:constructor): 16,141 (+200%)
• avm_simulation_bytecode_size_in_bytes (Token:transfer_public): 17,019 (+170%)

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	5,745	1,541	707	767 (-1%)	768 (-1%)
proof_construction_time_sha256_30_ms	11,785	3,147	1,405 (+1%)	1,448	1,472
proof_construction_time_sha256_100_ms	43,998	11,770	5,399	5,703	5,506 (+2%)
proof_construction_time_poseidon_hash_ms	79.0 (+1%)	34.0	34.0	58.0 (-3%)	89.0 (+1%)
proof_construction_time_poseidon_hash_30_ms	1,527	414	198	227 (-1%)	268
proof_construction_time_poseidon_hash_100_ms	5,643	1,515	670	737	754 (+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	4,420	7,940	14,948
l1_rollup_calldata_gas	50,500	93,252	178,472
l1_rollup_execution_gas	843,633	1,577,281	3,362,558
l2_block_processing_time_in_ms	254 (-2%)	452 (+1%)	820 (+1%)
l2_block_building_time_in_ms	11,449 (-1%)	22,457	44,639 (-1%)
l2_block_rollup_simulation_time_in_ms	11,449 (-1%)	22,457	44,638 (-1%)
l2_block_public_tx_process_time_in_ms	9,785 (-1%)	20,762	42,919 (-1%)

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	3,409 (+3%)	4,389 (+3%)
node_database_size_in_bytes	14,205,168	18,174,192
pxe_database_size_in_bytes	16,258	26,818

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	74.9 (-2%)	398 (+6%)	21,027	44,933	N/A
private-kernel-inner	141 (+1%)	673 (-2%)	71,803	45,067	N/A
private-kernel-reset-tiny	287	664	64,669	44,922	N/A
private-kernel-tail	150	127 (-2%)	50,553	57,178	N/A
base-parity	5.55 (+1%)	N/A	160	96.0	N/A
root-parity	35.2	N/A	73,964	96.0	N/A
base-rollup	2,584 (-1%)	N/A	189,128	664	N/A
block-root-rollup	41.2 (+1%)	N/A	58,217	2,448	N/A
public-kernel-setup	85.1 (+1%)	N/A	114,433	79,670	N/A
public-kernel-app-logic	97.6 (+1%)	N/A	114,251	79,670	N/A
public-kernel-tail	838	N/A	487,098	16,414	N/A
private-kernel-reset-small	269	N/A	66,345	45,629	N/A
private-kernel-tail-to-public	1,418 (-6%)	611 (+4%)	449,631	1,825	N/A
public-kernel-teardown	86.1 (+3%)	N/A	114,697	79,670	N/A
merge-rollup	20.0	N/A	38,182	664	N/A
undefined	N/A	N/A	N/A	N/A	83,960 (+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	340
ContractInstanceDeployer:deploy	1,408	11,731	18.0
MultiCallEntrypoint:entrypoint	1,920	11,731	418
FeeJuice:deploy	1,376	11,731	384 (-1%)
SchnorrAccount:constructor	1,312	11,731	59.6
SchnorrAccount:entrypoint	2,336	11,731	368 (+1%)
FeeJuice:claim	1,344	11,731	37.1 (+8%)
Token:privately_mint_private_note	1,280	11,731	65.3 (-2%)
FPC:fee_entrypoint_public	1,344	11,731	27.8 (+17%)
Token:transfer	1,312	11,731	207
Benchmarking:create_note	1,344	11,731	73.2 (-1%)
SchnorrAccount:verify_private_authwit	1,280	11,731	26.4
Token:unshield	1,376	11,731	487 (+1%)
FPC:fee_entrypoint_private	1,376	11,731	646 (+1%)

AVM Simulation

Time to simulate various public functions in the AVM.

Function	time_ms	bytecode_size_in_bytes
FeeJuice:_increase_public_balance	⚠️ 159 (+231%)	1,788 (+3%)
FeeJuice:set_portal	9.02 (-15%)	1,807 (+3%)
Token:constructor	⚠️ 279 (+281%)	⚠️ 19,994 (+117%)
FPC:constructor	⚠️ 81.5 (+80%)	⚠️ 16,141 (+200%)
FeeJuice:check_balance	35.8 (-4%)	1,467
Token:mint_public	⚠️ 307 (+519%)	3,395 (+2%)
Token:assert_minter_and_mint	⚠️ 50.9 (-92%)	2,249 (+3%)
AuthRegistry:set_authorized	45.6 (+15%)	670
FPC:prepare_fee	⚠️ 118 (+80%)	2,621
Token:transfer_public	⚠️ 36.3 (+181%)	⚠️ 17,019 (+170%)
FPC:pay_refund	47.6 (+19%)	3,338 (+2%)
Benchmarking:increment_balance	1,223	2,180 (+3%)
Token:_increase_public_balance	37.3 (-16%)	1,738
FPC:pay_refund_with_shielded_rebate	49.0 (-3%)	3,338 (+2%)

Public DB Access

Time to access various public DBs.

Function	time_ms
get-nullifier-index	0.265 (+1%)

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	2.18	3.91	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_count	16.7	31.8	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_ms	0.114	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	11.0 (-1%)	17.6	31.2 (+3%)	62.6 (+5%)	113 (+1%)
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	0.106 (-1%)	0.102	0.101 (+2%)	0.108 (+5%)	0.0994 (-1%)
batch_insert_into_indexed_tree_20_depth_ms	N/A	N/A	14.1 (-2%)	25.9 (-1%)	44.4 (+1%)	82.8 (-4%)	161 (-1%)
batch_insert_into_indexed_tree_20_depth_hash_count	N/A	N/A	109	207	357	691	1,363
batch_insert_into_indexed_tree_20_depth_hash_ms	N/A	N/A	0.107 (-2%)	0.105 (-1%)	0.107 (+1%)	0.102 (-5%)	0.103 (+1%)
batch_insert_into_indexed_tree_40_depth_ms	N/A	N/A	16.1 (-1%)	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_count	N/A	N/A	129	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_ms	N/A	N/A	0.106 (-2%)	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	72,416	668,302

Transaction size based on fee payment method

| Metric | |
| - | |

[dbanks12]

Approved, but couple of questions

[dbanks12]

I thought we were making calldata size a runtime variable?

[fcarreiro]

Yes yes, the difference is a bit subtle. The opcode takes an address (so, a runtime variable)

• We need to know the size at compile time for different reasons: to be able to create the [Field; N] where it will be stored, and then be able to pass that to the hasher.

[dbanks12]

Okay, but I'm a bit confused. How can the size be known at compile time if it's a runtime variable?

[fcarreiro]

It's not known at compile time now, that's why the current code is different from the expected code (in the description).

[dbanks12]

Why do we need this oracle now, but didn't need it before? There is no longer just a simple 1-1 mapping between the calldatacopy brillig opcode and the avm opcode?

[dbanks12]

same question about compile-time vs runtime

[fcarreiro]

Merge activity

• Sep 10, 11:42 AM EDT: @fcarreiro merged this pull request with Graphite.