feat(avm): Track gas usage based on memory accesses

[spalladino]

Defines gas cost for each instruction based on its memory accesses: reads, indirect reads (priced as 2 reads), writes. Each instruction defines its expected number of memory operations, and gas cost is computed based on that and a base gas cost defined in the gas cost table. There is an additional check after the instruction is run that verifies that the number of memory operations matches the ones declared by the instruction. Note that, since we need to charge gas before execution (as specified in the yellow paper), we cannot just track accesses during execution and then charge that.

The new workflow for an instruction, which no instruction needs to fully override, is:

• Invokes loadInputs to load from memory the inpiuts needed for execution (not needed if ops can compute their gas cost from their wires only)
• Requests the expected number of memoryOperations to compute the memoryGasCost (defaults to zero)
• Computes gasCost based on the loaded inputs, using the sum of baseGasCost and memoryGasCost (no need to override unless there's an additional gas component)
• Consumes gas based on the computed gas cost, throwing an exceptional halt if not enough
• Executes actual logic from internalExecute (all instructions override)
• Asserts the expected memory operations against the actual memory operations (can be turned off for performance if we wanted to)

This needs a bunch more tests, but I can add them in a subsequent PR.

[spalladino]

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): Track gas usage based on memory accesses #5518 👈
• refactor(avm): Introduce fixed gas cost instruction abstraction #5514
• feat(avm): Gas usage for nested calls #5495 : 1 other dependent PR (#5563 )
• master

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

Join @spalladino and the rest of your teammates on Graphite

[AztecBot]

Benchmark results

Metrics with a significant change:

• note_successful_decrypting_time_in_ms (8): 243 (-18%)

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 9ffe4571 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	676	676	676
l1_rollup_calldata_gas	6,424	6,424	6,424
l1_rollup_execution_gas	585,757	585,757	585,757
l2_block_processing_time_in_ms	1,307 (-2%)	4,673 (-2%)	9,114 (-3%)
note_successful_decrypting_time_in_ms	⚠️ 243 (-18%)	577 (-14%)	1,013 (-1%)
note_trial_decrypting_time_in_ms	46.1 (-54%)	97.7 (+12%)	38.0 (-21%)
l2_block_building_time_in_ms	12,487 (+2%)	45,243 (+1%)	91,196 (+3%)
l2_block_rollup_simulation_time_in_ms	7,570 (+1%)	26,390 (+1%)	52,318 (+2%)
l2_block_public_tx_process_time_in_ms	4,888 (+3%)	18,801 (+1%)	38,776 (+4%)

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	14,436 (-1%)	27,499
note_history_successful_decrypting_time_in_ms	1,310 (+1%)	2,457 (+3%)
note_history_trial_decrypting_time_in_ms	110 (+18%)	145 (+18%)
node_database_size_in_bytes	18,677,840	35,000,400
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	184 (+1%)	44,379	28,214
private-kernel-ordering	170 (-1%)	52,880	14,296
base-parity	3,477 (-1%)	128	311
root-parity	1,315 (+4%)	1,244	311
base-rollup	15,244 (+1%)	165,760	861
root-rollup	49.4 (+1%)	4,359	725
private-kernel-inner	228	73,795	28,214
public-kernel-app-logic	97.4 (+2%)	35,252	28,217
public-kernel-tail	140 (+1%)	40,928	28,217
merge-rollup	8.90 (+6%)	2,568	861
public-kernel-teardown	89.8 (-1%)	35,252	28,217
public-kernel-setup	88.9 (-1%)	35,252	28,217

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	10.0	16.0 (-1%)	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	0.584	0.494 (-1%)	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	48.3 (+5%)	72.8	236 (+1%)	448 (-1%)	881 (+1%)	1,728	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	0.495 (+5%)	0.449 (+1%)	0.428 (+1%)	0.419 (-1%)	0.418 (+1%)	0.413	N/A
batch_insert_into_indexed_tree_20_depth_ms	N/A	N/A	54.9 (+1%)	107 (-1%)	342 (+1%)	666	1,325 (+1%)	2,613	N/A
batch_insert_into_indexed_tree_20_depth_hash_count	N/A	N/A	105	207	691	1,363	2,707	5,395	N/A
batch_insert_into_indexed_tree_20_depth_hash_ms	N/A	N/A	0.483 (+1%)	0.482 (-1%)	0.462 (+1%)	0.460	0.460 (+1%)	0.454	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	61.7
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	0.539

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	14,968	495,320

Transaction size based on fee payment method

Metric	none fee payment method	native fee payment method	fpc_public fee payment method	fpc_private fee payment method
tx_with_fee_size_in_bytes	773	905	1,161	1,377

Transaction processing duration by data writes.

Metric	0 new note hashes	1 new note hashes	2 new note hashes
tx_pxe_processing_time_ms	1,845 (+3%)	1,210	5,568 (+1%)

Metric	0 public data writes	1 public data writes	2 public data writes	3 public data writes	4 public data writes	5 public data writes	8 public data writes
tx_sequencer_processing_time_ms	10.1 (+3%)	588 (+1%)	482 (+12%)	1,136	567	1,971	557 (+1%)