feat(avm-simulator): individual storage access

[fcarreiro]

WARNING: PR waiting for resolution of blowup. See below.

---

This PR updates the oracle storageRead/Write interface to take and return just 1 value, instead of an array of values. However, the function storage_write<N> actually used in Noir still takes/returns an array, and does a loop calling the oracle.

• The oracle storageWrite was returning a value which was not used. This is due to an ACVM limitation, see below.

The reason for the change is, quoting Sean: "that it's easier for the AVM to prove single-access to memory". IIUC, we want the traces to contain single access, and our AVM opcodes handle that.

This change is NOT trivial, since it implied changing the ACVM and oracle interface. In particular

• The ACVM oracle interface assumed that every oracle function returned a value (or array of values).
• I made some changes so that void oracles are possible.

NOTE FOR THE REVIEWER: Sean gave an informal approval on the AVM side, please review everything else. If you are not familiar with the code please let me know of a possible reviewer.

---

This causes a blowup in instructions for storage access.

Before

open avm_setAdmin() 
        function signature: avm_setAdmin()
        selector: 0xb2a1e760
        bytecode: 555 bytes (sha256 174bb4bfe990d33a453a93706abd25cbe921cbb67bdb3acaa04e4308cdc425ef)

After (name changed)

open avm_setStorageSingle(a: Field) 
        function signature: avm_setStorageSingle(Field)
        selector: 0x928e2ba7
        bytecode: 31619 bytes (sha256 40219438c1fd8f261482505c305bd0530dbc19f02c35ad8067b8a5263db2b55a)

[fcarreiro]

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

• feat(avm-simulator): individual storage access #5095 👈
• chore: increase permitted bytecode size #5136
• 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:

• tx_sequencer_processing_time_ms (0): 14.7 (+92%)
• note_history_successful_decrypting_time_in_ms (5): 2,195 (+72%)
• note_trial_decrypting_time_in_ms (8): 76.8 (+185%)
• note_trial_decrypting_time_in_ms (64): 125 (+256%)

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.

Values are compared against data from master at commit 7f216eb0 and shown if the difference exceeds 1%.

L2 block published to L1

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

Metric	8 txs	32 txs	64 txs
l1_rollup_calldata_size_in_bytes	5,668	18,820	36,356
l1_rollup_calldata_gas	65,908	238,780	469,436
l1_rollup_execution_gas	666,388	953,489	1,336,592
l2_block_processing_time_in_ms	1,392 (+5%)	5,018 (+5%)	9,500 (+1%)
note_successful_decrypting_time_in_ms	213 (-5%)	531 (+11%)	991 (-1%)
note_trial_decrypting_time_in_ms	⚠️ 76.8 (+185%)	48.7 (-50%)	⚠️ 125 (+256%)
l2_block_building_time_in_ms	12,402 (+1%)	49,916 (+3%)	97,921 (+1%)
l2_block_rollup_simulation_time_in_ms	7,382 (+1%)	30,164 (+4%)	58,425 (+1%)
l2_block_public_tx_process_time_in_ms	4,982 (+1%)	19,678 (+1%)	39,288 (+1%)

L2 chain processing

Each column represents the number of blocks on the L2 chain where each block has 16 txs.

Metric	5 blocks	10 blocks
node_history_sync_time_in_ms	14,510 (+2%)	28,888 (+4%)
note_history_successful_decrypting_time_in_ms	⚠️ 2,195 (+72%)	2,788 (+13%)
note_history_trial_decrypting_time_in_ms	166 (+46%)	156 (-18%)
node_database_size_in_bytes	19,091,536 (+2%)	35,684,432 (+1%)
pxe_database_size_in_bytes	29,859	59,414

Circuits stats

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

Circuit	circuit_simulation_time_in_ms	circuit_input_size_in_bytes	circuit_output_size_in_bytes
private-kernel-init	280 (+3%)	44,338	27,700
private-kernel-ordering	212 (+2%)	52,324	14,326
base-rollup	686 (+4%)	177,083	925
root-rollup	76.7 (+12%)	4,176	789
private-kernel-inner	376 (+5%)	73,197	27,700
public-kernel-app-logic	184 (+3%)	34,582	27,671
public-kernel-tail	93.7 (+2%)	34,206	27,671
merge-rollup	9.80 (+18%)	2,696	925

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	512 leaves	1024 leaves	2048 leaves	4096 leaves	32 leaves
batch_insert_into_append_only_tree_16_depth_ms	10.8 (+8%)	16.9 (+5%)	N/A	N/A	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_count	16.8	31.6	N/A	N/A	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_ms	0.630 (+8%)	0.524 (+5%)	N/A	N/A	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_32_depth_ms	N/A	N/A	50.3 (+10%)	72.5 (-2%)	230	443 (-1%)	864	1,723 (-1%)	N/A
batch_insert_into_append_only_tree_32_depth_hash_count	N/A	N/A	96.0	159	543	1,055	2,079	4,127	N/A
batch_insert_into_append_only_tree_32_depth_hash_ms	N/A	N/A	0.517 (+10%)	0.448 (-2%)	0.420	0.416 (-1%)	0.411	0.413 (-1%)	N/A
batch_insert_into_indexed_tree_20_depth_ms	N/A	N/A	58.6 (+9%)	106	334	656	1,312	2,616 (+1%)	N/A
batch_insert_into_indexed_tree_20_depth_hash_count	N/A	N/A	104	207	691	1,363	2,707	5,395	N/A
batch_insert_into_indexed_tree_20_depth_hash_ms	N/A	N/A	0.523 (+10%)	0.478	0.455	0.452	0.454	0.453	N/A
batch_insert_into_indexed_tree_40_depth_ms	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	65.3 (+7%)
batch_insert_into_indexed_tree_40_depth_hash_count	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	109
batch_insert_into_indexed_tree_40_depth_hash_ms	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	0.571 (+7%)

Miscellaneous

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

Metric	0 registered classes
tx_size_in_bytes	22,014 (+15%)

Transaction processing duration by data writes.

Metric	0 new note hashes	1 new note hashes
tx_pxe_processing_time_ms	3,019 (+7%)	1,480 (+3%)

Metric	0 public data writes	1 public data writes
tx_sequencer_processing_time_ms	⚠️ 14.7 (+92%)	629 (+3%)