chore: refactor AVM simulator's side-effect tracing

[dbanks12]

No description provided.

[dbanks12]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• chore: refactor AVM simulator's side-effect tracing #7091 👈
• chore: avm simulator uses regular ContractStorageRead type #7090
• master

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

Join @dbanks12 and the rest of your teammates on Graphite

[fcarreiro]

This one doesn't seem to logically be part of a "side effect trace".

For example, for the kernel, the hints are computed by a different class, based on the requests:

aztec-packages/yarn-project/simulator/src/public/tail_phase_manager.ts

Lines 77 to 85 in a0b9c4b

	const nullifierReadRequestHints = await this.hintsBuilder.getNullifierReadRequestHints(
	validationRequests.nullifierReadRequests,
	pendingNullifiers,
	);
	const nullifierNonExistentReadRequestHints = await this.hintsBuilder.getNullifierNonExistentReadRequestHints(
	validationRequests.nullifierNonExistentReadRequests,
	pendingNullifiers,
	);

I wonder if all of our hints should (or not) be derivable from the data we'll send to the kernel. E.g., note hash existence is not derivable now, but nullifier existence is (we know this has to be changed). On the other hand, for a public storage read, I guess the kernel doesn't need the value, but we do ?

It feels like we are trying to reuse the kernel structures here (and in the proving request) but then we have to add extra hints for the AVM. That didn't bother me much, but if we are polluting the simulator (which could be just thought as an interpreter that tracks things) then relying on kernel structures feels too stretched.

(not my final words, just thoughts, still thinking about it)

[fcarreiro]

Looking nice!

[fcarreiro]

Consider making gathering this in a separate tracer (and passing a list of tracers to the sim). It might be useful for tests where you might want a mock tracer for side effects but want the real hints.

In any case, optional, the above case might not exist.

[dbanks12]

I'll give this some thought. That might be nice....

[fcarreiro]

Just realized that this might not be possible if you need to share the side effect counter between the tracers (unless you track it in both which is error prone)

[fcarreiro]

What a work, thanks! Left many comments about particular things but I'll approve since the general idea LGTM.

[fcarreiro]

I see you are changing the "address" to "storage address". I think this is correct BUT the only reason there is address and storage address is apparently because of delegate calls (asked Palla some weeks ago). So right now because they are unimplemented I expect the addresses to be the same. And if we remove delegate call, we'd use this_address everywhere when we do the switch.

[dbanks12]

Good point. I think until we decide to fully/finally remove all delegate call concepts from the protocol it's a good idea to separate the concept of address & storageAddress.

[fcarreiro]

This is correct, but just note that IIRC there is no getter for this so we might remove it ?

[dbanks12]

I think we need it to eventually enforce a max call depth

[fcarreiro]

I don't remember very well but we might even be able to remove this. IIRC it was related to parity with ACVM. (probably not in this PR).

[dbanks12]

added item to cleanup tidcket #5818

[AztecBot]

Benchmark results

Metrics with a significant change:

• proof_construction_time_sha256_30_ms (16): 1,720 (+19%)
• app_circuit_proving_time_in_ms (Token:transfer_public (avm)): 3,038 (-37%)
• protocol_circuit_witness_generation_time_in_ms (private-kernel-inner): 1,306 (+19%)
• protocol_circuit_witness_generation_time_in_ms (root-rollup): 173 (+95%)

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_30_ms	11,998 (+3%)	3,229 (+3%)	⚠️ 1,720 (+19%)	1,602 (+5%)	1,478 (-3%)
proof_construction_time_sha256_100_ms	44,644 (+2%)	11,898 (+1%)	5,431	5,441 (+1%)	5,373
proof_construction_time_poseidon_hash_ms	79.0 (+1%)	34.0	34.0	58.0	88.0
proof_construction_time_poseidon_hash_30_ms	1,551 (+2%)	423 (+1%)	200	221 (-1%)	263 (-2%)
proof_construction_time_poseidon_hash_100_ms	5,862 (+2%)	1,605 (+2%)	725	791 (+1%)	800 (+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	1,412	1,412	1,412
l1_rollup_calldata_gas	9,452	9,458	9,452
l1_rollup_execution_gas	610,273	610,279	610,273
l2_block_processing_time_in_ms	758 (-2%)	1,408	2,695
l2_block_building_time_in_ms	25,445 (+1%)	51,064 (+1%)	100,210 (+1%)
l2_block_rollup_simulation_time_in_ms	25,445 (+1%)	51,063 (+1%)	100,210 (+1%)
l2_block_public_tx_process_time_in_ms	21,865 (+1%)	47,279 (+1%)	96,455 (+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	7,006	9,787 (-1%)
node_database_size_in_bytes	12,152,912	16,068,688
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	proving_time_in_ms	input_size_in_bytes	output_size_in_bytes	proof_size_in_bytes	num_public_inputs	size_in_gates
private-kernel-init	129 (-1%)	470	12,913 (-1%)	20,634	67,190	92,352	2,819	524,288
private-kernel-inner	391	⚠️ 1,306 (+19%)	48,806	94,902	67,190	92,352	2,819	2,097,152
private-kernel-tail	371 (+1%)	1,769 (-2%)	49,423 (-12%)	99,121	71,733	14,912	399	2,097,152
base-parity	6.15 (-4%)	1,850 (-1%)	2,760	128	64.0	2,208	2.00	131,072
root-parity	48.8	76.6 (+5%)	40,443 (-2%)	27,100	64.0	2,720	18.0	2,097,152
base-rollup	7,934 (+1%)	4,823 (-3%)	85,767	170,330	756	3,648	47.0	4,194,304
root-rollup	111 (+1%)	⚠️ 173 (+95%)	23,259 (+3%)	25,309	620	3,456	41.0	1,048,576
public-kernel-setup	708 (-1%)	3,697 (-1%)	44,918	116,905	93,334	125,344	3,850	2,097,152
public-kernel-app-logic	616 (+1%)	4,737	46,255 (-4%)	116,905	93,334	125,344	3,850	2,097,152
public-kernel-tail	1,412 (+1%)	34,727 (-8%)	192,838 (-2%)	511,910	10,014	14,912	399	8,388,608
private-kernel-reset-small	594 (+1%)	2,232 (-10%)	45,136 (-2%)	123,313	67,190	92,352	2,819	2,097,152
public-kernel-teardown	611 (-1%)	4,680 (-1%)	47,230 (-2%)	116,905	93,334	125,344	3,850	2,097,152
merge-rollup	28.7	N/A	N/A	16,542	756	N/A	N/A	N/A
private-kernel-tail-to-public	N/A	7,552 (+8%)	97,165 (-6%)	N/A	N/A	125,344	3,850	4,194,304

Stats on running time collected for app circuits

Function	input_size_in_bytes	output_size_in_bytes	witness_generation_time_in_ms	proof_size_in_bytes	proving_time_in_ms	size_in_gates	num_public_inputs
ContractClassRegisterer:register	1,344	9,944	422 (-1%)	N/A	N/A	N/A	N/A
ContractInstanceDeployer:deploy	1,408	9,944	39.5 (-1%)	N/A	N/A	N/A	N/A
MultiCallEntrypoint:entrypoint	1,920	9,944	1,781	N/A	N/A	N/A	N/A
GasToken:deploy	1,376	9,944	972 (-1%)	N/A	N/A	N/A	N/A
SchnorrAccount:constructor	1,312	9,944	1,406	N/A	N/A	N/A	N/A
SchnorrAccount:entrypoint	2,304	9,944	2,819	16,768	55,974 (+1%)	2,097,152	457
Token:privately_mint_private_note	1,280	9,944	1,603 (-2%)	N/A	N/A	N/A	N/A
FPC:fee_entrypoint_public	1,344	9,944	349 (-2%)	16,768	11,607 (+3%)	524,288	457
Token:transfer	1,312	9,944	4,449 (-1%)	16,768	46,018	2,097,152	457
AuthRegistry:set_authorized (avm)	21,043	N/A	N/A	87,200	1,368 (-2%)	N/A	N/A
FPC:prepare_fee (avm)	28,521 (-1%)	N/A	N/A	88,032	5,656 (-3%)	N/A	N/A
AuthRegistry:consume (avm)	34,973	N/A	N/A	87,616	3,014 (+1%)	N/A	N/A
Token:transfer_public (avm)	44,849	N/A	N/A	87,616	⚠️ 3,038 (-37%)	N/A	N/A
FPC:pay_refund (avm)	33,381 (-1%)	N/A	N/A	88,448	11,060 (-1%)	N/A	N/A
Benchmarking:create_note	1,344	9,944	1,403	N/A	N/A	N/A	N/A
SchnorrAccount:verify_private_authwit	1,280	9,944	73.5	N/A	N/A	N/A	N/A
Token:unshield	1,376	9,944	3,683	N/A	N/A	N/A	N/A
FPC:fee_entrypoint_private	1,376	9,944	4,619 (-1%)	N/A	N/A	N/A	N/A

AVM Simulation

Time to simulate various public functions in the AVM.

Function	time_ms	bytecode_size_in_bytes
GasToken:_increase_public_balance	69.4 (+1%)	13,873
GasToken:set_portal	16.5 (-4%)	3,495
Token:constructor	92.9 (-9%)	24,207
FPC:constructor	61.9 (+1%)	13,893
GasToken:mint_public	52.0 (+1%)	10,241
Token:mint_public	592 (-3%)	19,216
Token:assert_minter_and_mint	59.2 (+4%)	13,034
AuthRegistry:set_authorized	33.7 (+10%)	7,869
FPC:prepare_fee	158 (+14%)	15,187
Token:transfer_public	51.0 (+9%)	31,425
FPC:pay_refund	152 (+5%)	20,080
Benchmarking:increment_balance	2,706 (+1%)	15,465
Token:_increase_public_balance	56.0 (-3%)	15,089
FPC:pay_refund_with_shielded_rebate	146 (-1%)	21,167

Public DB Access

Time to access various public DBs.

Function	time_ms
get-nullifier-index	0.160 (+5%)

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	10.4 (+1%)	16.8	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	0.603 (+1%)	0.516	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_32_depth_ms	N/A	N/A	48.0 (-1%)	76.1	132 (+1%)	245	472
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	0.491 (-1%)	0.468	0.452 (+1%)	0.443	0.441
batch_insert_into_indexed_tree_20_depth_ms	N/A	N/A	59.3	112 (+1%)	183	352	695
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	0.502	0.501	0.484	0.477	0.476
batch_insert_into_indexed_tree_40_depth_ms	N/A	N/A	72.6	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_count	N/A	N/A	133	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_ms	N/A	N/A	0.517	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	85,672	670,983

Transaction size based on fee payment method

| Metric | |
| - | |