chore: Measure block building times, history processing times, and db sizes

[spalladino]

• Measures total block building time, as well as the time to process public txs and the time to run the rollup circuits.
• Changes the elapsed methods so they return a duration and not a live timer, otherwise the timer would keep running after the nested function passed onto elapsed has finished.
• Introduces a new benchmark test for building a chain with multiple blocks, and at predetermined checkpoints measures the total time for a node and a pxe to sync it.
• Measures db sizes for both aztec node and pxe in the history processing benchmark.

[spalladino]

Testing comments

[AztecBot]

Testing

[AztecBot]

(Manually edited comment for testing)

Benchmark 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.

Values are compared against data from master at commit 37d9f9cc 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	128 txs
l1_rollup_calldata_size_in_bytes	45,444	179,588	716,132
l1_rollup_calldata_gas	222,984	867,956	3,449,696
l1_rollup_execution_gas	842,071	3,595,064	22,205,065
l2_block_processing_time_in_ms	1,060	3,981	15,688
note_successful_decrypting_time_in_ms	332	1,019	3,780
note_trial_decrypting_time_in_ms	34.0	108	138
l2_block_building_time_in_ms	9,114	36,117	152,315
l2_block_rollup_simulation_time_in_ms	6,771	26,781	107,164
l2_block_public_tx_process_time_in_ms	2,300	9,209	44,431

L2 chain processing

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

Metric	10 blocks	20 blocks	30 blocks
node_history_sync_time_in_ms	30,823	75,516	136,231
note_history_successful_decrypting_time_in_ms	4,653	12,961	20,148
note_history_trial_decrypting_time_in_ms	147	208	254
node_database_size_in_bytes	1,194,179	1,900,681	2,754,125
pxe_database_size_in_bytes	54,187	108,338	162,578

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	54.9	56,577	14,745
private-kernel-ordering	30.2	20,137	8,089
base-rollup	⚠️ 871 (-53%)	631,604	810
root-rollup	37.9	⚠️ 4,072 (+5556%)	1,097
private-kernel-inner	51.9	72,288	14,745
public-kernel-private-input	51.7	37,359 (-3%)	14,745
public-kernel-non-first-iteration	31.6	37,401	14,745
merge-rollup	1.00	2,592	873

[PhilWindle]

Looks great.

Some of these benchmarks are rather concerning. Especially as we aren't doing any proving yet!

l1_rollup_execution_gas is mental. Firstly the fact that it's 100K+ gas per tx but that it gets worse as the rollups get bigger.

Really all of the benchmarks at 128 txs. 22 million gas! 2.5 minutes to build the block with no proofs etc.