chore: Compute function tree root in ts.

[charlielye]

One step closer to not needing circuits.wasm.

• This creates a new class MerkleTreeRootCalculator. It wouldn't have had to exist except for the fact that the zero leaf of the function tree is the hash of a zeroed function leaf (5 fields) rather than just a 0 buffer. Kinda annoying. Needed?
• I tried putting the class in the merkle-tree package, but worryingly that created a circular dependency because everything depends on this generic "types" catch all package, and it depends on circuits... Seems sus. For now the class sits next to where it's used in abis.
• When performing a naive tree hash, this actually beats the wasm performance by around 5%. Perhaps due to the wasm always computing the zero leaf. Assuming that's the reason, the performance is equal despite overhead of calling into wasm (so we can assume such overheads are negligible).
• The algorithm performs much better than the wasm version however, when the tree is not full, as it leverages zero layer caches to not have to hash the entire tree. The test adds 4 leaves to a 16 leaf tree, and as expected is 4 times faster.

Just for clarity, this still uses wasm, but it uses bb.js wasm and at a lower level. Just calling pedersenHash.

[AztecBot]

Benchmark results

Metrics with a significant change:

• circuit_simulation_time_in_ms (public-kernel-private-input): 40.9 (-93%)
• circuit_simulation_time_in_ms (public-kernel-non-first-iteration): 25.8 (-94%)
• l2_block_building_time_in_ms (8): 13,239 (-36%)
• l2_block_building_time_in_ms (32): 52,686 (-35%)
• l2_block_building_time_in_ms (128): 209,515 (-36%)
• l2_block_public_tx_process_time_in_ms (8): 1,304 (-85%)
• l2_block_public_tx_process_time_in_ms (32): 5,160 (-85%)
• l2_block_public_tx_process_time_in_ms (128): 20,536 (-85%)

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 15a522ba 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	223,008	868,160	3,449,432
l1_rollup_execution_gas	842,095	3,595,268	22,204,801
l2_block_processing_time_in_ms	1,981 (-1%)	7,744 (+2%)	30,001 (-1%)
note_successful_decrypting_time_in_ms	296 (+2%)	870	3,158 (-3%)
note_trial_decrypting_time_in_ms	98.8 (+6%)	78.5 (+4%)	133 (-1%)
l2_block_building_time_in_ms	⚠️ 13,239 (-36%)	⚠️ 52,686 (-35%)	⚠️ 209,515 (-36%)
l2_block_rollup_simulation_time_in_ms	11,895 (-2%)	47,393	188,487
l2_block_public_tx_process_time_in_ms	⚠️ 1,304 (-85%)	⚠️ 5,160 (-85%)	⚠️ 20,536 (-85%)

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	21,719 (-2%)	42,454 (-1%)
note_history_successful_decrypting_time_in_ms	2,035 (-2%)	3,954 (-3%)
note_history_trial_decrypting_time_in_ms	122 (+1%)	144
node_database_size_in_bytes	1,631,322	1,099,541
pxe_database_size_in_bytes	29,748	59,307

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	763 (-1%)	61,697	18,905
private-kernel-ordering	125 (-1%)	24,297	8,153
base-rollup	1,765 (-1%)	656,311	814 (-7%)
root-rollup	172 (+2%)	4,072	1,097
private-kernel-inner	787 (-1%)	81,568	18,905
public-kernel-private-input	⚠️ 40.9 (-93%)	41,519	18,841
public-kernel-non-first-iteration	⚠️ 25.8 (-94%)	41,497	18,841
merge-rollup	15.1 (-1%)	2,592	873

Miscellaneous

Transaction sizes based on how many contracts are deployed in the tx.

Metric	0 deployed contracts	1 deployed contracts
tx_size_in_bytes	8,787	27,547

[ludamad]

LGTM