chore: extract merge from UC and simplify

[ledwards2225]

Removes create_merge_prover/verifier from UltraComposer. They are now constructed directly from their respective class constructors. This PR also simplifies the merge classes and removes the templating on Flavor which was unnecessary.

closes AztecProtocol/barretenberg#804

[AztecBot]

Benchmark results

No metrics with a significant change found.

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 8569e7c0 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	45,636	180,516	720,036
l1_rollup_calldata_gas	222,564	871,044	3,465,900
l1_rollup_execution_gas	315,051	988,784	3,685,557
l2_block_processing_time_in_ms	1,172 (+2%)	4,405	17,198
note_successful_decrypting_time_in_ms	311 (+2%)	975 (+1%)	3,633 (+1%)
note_trial_decrypting_time_in_ms	21.9 (+34%)	104 (+4%)	142 (+1%)
l2_block_building_time_in_ms	17,856 (+1%)	70,377	282,423
l2_block_rollup_simulation_time_in_ms	13,131	51,762	208,405
l2_block_public_tx_process_time_in_ms	4,690 (+1%)	18,470 (+1%)	73,743

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	13,858	26,145 (-1%)
note_history_successful_decrypting_time_in_ms	2,308 (-4%)	4,715 (-2%)
note_history_trial_decrypting_time_in_ms	72.5 (-23%)	144 (-20%)
node_database_size_in_bytes	17,330,256	33,050,704
pxe_database_size_in_bytes	29,923	59,478

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	310	44,481	26,193
private-kernel-ordering	193	44,513	14,929
base-rollup	1,405	128,970	881
root-rollup	82.3	4,088	677
private-kernel-inner	415	71,640	26,193
public-kernel-private-input	239	32,615	26,193
public-kernel-non-first-iteration	239	32,657	26,193
merge-rollup	7.56 (-1%)	2,608	881

Tree insertion stats

The duration to insert a fixed batch of leaves into each tree type.

Metric	1 leaves	2 leaves	8 leaves	16 leaves	32 leaves	128 leaves	64 leaves	512 leaves	1024 leaves	2048 leaves	8192 leaves
batch_insert_into_append_only_tree_16_depth_ms	10.3 (+1%)	10.6 (-5%)	12.7 (-3%)	17.1 (+1%)	23.3 (+2%)	63.7	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_count	16.9	17.5	23.0	31.6	47.0	143	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_ms	0.596	0.590 (-5%)	0.540 (-4%)	0.530 (+1%)	0.486 (+1%)	0.440	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_32_depth_ms	N/A	N/A	N/A	N/A	N/A	74.3	47.7 (+1%)	241 (+1%)	463 (+1%)	909 (+1%)	3,543
batch_insert_into_append_only_tree_32_depth_hash_count	N/A	N/A	N/A	N/A	N/A	159	96.0	543	1,055	2,079	8,223
batch_insert_into_append_only_tree_32_depth_hash_ms	N/A	N/A	N/A	N/A	N/A	0.460	0.489 (+1%)	0.439 (+1%)	0.433	0.432 (+1%)	0.426
batch_insert_into_indexed_tree_20_depth_ms	N/A	N/A	N/A	N/A	N/A	104 (+1%)	57.0 (+1%)	355 (+2%)	687	1,375 (-1%)	5,395
batch_insert_into_indexed_tree_20_depth_hash_count	N/A	N/A	N/A	N/A	N/A	197	104	691	1,363	2,707	10,771
batch_insert_into_indexed_tree_20_depth_hash_ms	N/A	N/A	N/A	N/A	N/A	0.491	0.497 (+1%)	0.482 (+1%)	0.473	0.475	0.470
batch_insert_into_indexed_tree_40_depth_ms	N/A	N/A	N/A	56.4	N/A	N/A	N/A	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_count	N/A	N/A	N/A	94.1	N/A	N/A	N/A	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_ms	N/A	N/A	N/A	0.578	N/A	N/A	N/A	N/A	N/A	N/A	N/A

Miscellaneous

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

Metric	0 deployed contracts	1 deployed contracts
tx_size_in_bytes	15,575	37,974

Transaction processing duration by data writes.

Metric	0 new commitments	1 new commitments
tx_pxe_processing_time_ms	549	1,215

Metric	0 public data writes	1 public data writes
tx_sequencer_processing_time_ms	0.547 (-12%)	571

[maramihali]

Looks mostly good to me as changes, requested just some comments to be kept and left a question

[maramihali]

Is the plan to eventually do this for all Provers/Verifiers or is there sth specific to the Merge ones that motivated this change?

[ledwards2225]

For the Merge classes it didn't make any sense for the Composer to be involved, this was just leftover from when the merge protocol was baked into the main Ultra protocol. More generally, it makes sense for classes to be constructed directly via their constructors when possible. The only argument for having methods like create_prover/verifier on the Composer is that the composer is doing some behind the scenes work that's required for the construction of both of these classes. I suspect that we could do what I've done here for most of the other classes currently constructed by the composer and I think we probably should.

[maramihali]

shouldn't still be taken from the flavor, even though GoblinUltra flavor is going to be the fixed one

[ledwards2225]

No because essentially the only data being extracted from the Flavor here was curve data. We'll only ever run the merge protocol on bn254 so there's no need for templating (similar to how the Translator is not templated).

[maramihali]

I think the comments here are valuable and are not entirely reflected in the changes to MergeProver

[ledwards2225]

Can you be more specific? Looks like there are two comments here and I think they both made it into the new MergeProver constructor (albeit in a slightly modified form)

[maramihali]

yes, indeed, I rechecked, my bad