chore(bb): prereq work for polynomial mem optimization

[ludamad]

This PR: Splitting the implementation of plonk and honk into different polynomial classes.

Greater work: We will be moving to 'polynomials with more structure' -> aka tracking a size and a virtual size* where 'size' is actually allocated memory, and virtual size is conceptually all 0's. This should save a good deal of memory.

• I would call virtual size one more than the degree of the polynomial, but it's actually a bit weird to call it that when [size, virtual size) indices are all 0 coefficients. It's more like one more than the maximum degree of the polynomial, and only really used as an upper bound for loops or debug bounds checks.

[fcarreiro]

Nice to see! A few questions (without having looked at the code yet)

(1) Are these sparse in any place, or just a tail of zeroes "after size"? (the former is far better for the AVM)
(2) Did you see any noticeable performance overhead using these?

[ludamad]

Peeking a little early :) they are not called sparse anymore, they just have a trail of implied zeroes, and are not used yet (I would have done it sooner but I was out sick). We will support blocks of non-zero values, do you mean to say true sparseness is useful for AVM (aka not even using arrays)?

[fcarreiro]

Ah sorry, we were added automatically by CODEOWNERS.

We will support blocks of non-zero values, do you mean to say true sparseness is useful for AVM (aka not even using arrays)?

Interleaved blocks of non-zero values would be good enough for the AVM I think. That's why at some point I proposed the mmap solution (for native) which then you'd automatically get page-sized blocks (4k) and zero-filled blocks (that are not written to) I think would be "free" (virtual).

I'll remove myself from the PR and if at some point you'd like a review (or discuss things!) LMK :)

[AztecBot]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'C++ Benchmark'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.05.

Benchmark suite	Current: 4e59754	Previous: f6105d0	Ratio
commit(t)	3774980967 ns/iter	3580620778 ns/iter	1.05

This comment was automatically generated by workflow using github-action-benchmark.

CC: @ludamad @codygunton

[ledwards2225]

Nice. LGTM. Excited to get this into play

[ledwards2225]

This still needs to be addressed I guess? Could be one of the few situations where a duplicate method is needed for Plonk but maybe in this first pass.

[ludamad]

Removed this, it was a stray comment, there was no duplication

[ledwards2225]

Flagging this WORKTODO. Fine with me if you want to keep them here for now assuming this is going to be a continued thread of work

[ludamad]

Removed this, we'll go over all constructors

[ledwards2225]

It would be interesting to see if we can just delete all of these FFT methods. None of them should be used by Honk so it might just be that easy

[ludamad]

Yeah we can, I meant to, done

[AztecBot]

Benchmark results

Metrics with a significant change:

• avm_simulation_time_ms (Token:assert_minter_and_mint): 68.3 (+76%)

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.

Proof generation

Each column represents the number of threads used in proof generation.

Metric	1 threads	4 threads	16 threads	32 threads	64 threads
proof_construction_time_sha256_ms	5,741 (-1%)	1,555	700 (+1%)	755 (-1%)	765
proof_construction_time_sha256_30_ms	11,820	3,186	1,417	1,444 (-2%)	1,474
proof_construction_time_sha256_100_ms	44,117 (-1%)	12,354	5,744 (+1%)	5,612 (+1%)	5,525
proof_construction_time_poseidon_hash_ms	78.0	34.0 (-29%)	34.0	60.0 (+3%)	88.0
proof_construction_time_poseidon_hash_30_ms	1,527 (-1%)	419	200	231	271 (+1%)
proof_construction_time_poseidon_hash_100_ms	5,650 (-1%)	1,525 (-1%)	677	730 (+1%)	745 (-2%)

L2 block published to L1

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

Metric	4 txs	8 txs	16 txs
l1_rollup_calldata_size_in_bytes	4,324	7,844	14,852
l1_rollup_calldata_gas	49,684	92,504	177,512
l1_rollup_execution_gas	1,294,347	2,042,135	3,869,182
l2_block_processing_time_in_ms	257 (-1%)	435 (-2%)	826 (-1%)
l2_block_building_time_in_ms	9,326	18,189	36,230
l2_block_rollup_simulation_time_in_ms	9,325	18,189	36,229
l2_block_public_tx_process_time_in_ms	7,868	16,634	34,656

L2 chain processing

Each column represents the number of blocks on the L2 chain where each block has 8 txs.

Metric	3 blocks	5 blocks
node_history_sync_time_in_ms	2,937 (-2%)	3,739 (-2%)
node_database_size_in_bytes	12,738,640	16,789,584
pxe_database_size_in_bytes	16,254	26,813

Circuits stats

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

Circuit	simulation_time_in_ms	witness_generation_time_in_ms	input_size_in_bytes	output_size_in_bytes	proving_time_in_ms
private-kernel-init	95.8	394 (-1%)	21,846	44,858	N/A
private-kernel-inner	173 (-1%)	690 (-2%)	72,545	45,005	N/A
private-kernel-reset-tiny	506	1,192 (-2%)	69,614	44,844	N/A
private-kernel-tail	274 (+1%)	194 (-1%)	50,760	52,256	N/A
base-parity	5.55	N/A	160	96.0	N/A
root-parity	33.3	N/A	69,084	96.0	N/A
base-rollup	2,866	N/A	187,817	664	N/A
root-rollup	38.8 (+1%)	N/A	54,525	716	N/A
public-kernel-setup	100	N/A	103,760	71,222	N/A
public-kernel-app-logic	106	N/A	103,599	71,222	N/A
public-kernel-tail	571 (-1%)	N/A	409,190	16,414	N/A
private-kernel-reset-small	487 (+1%)	N/A	66,533	45,629	N/A
private-kernel-tail-to-public	1,319 (-2%)	749 (+5%)	507,554	1,697	N/A
public-kernel-teardown	94.0 (-1%)	N/A	104,005	71,222	N/A
merge-rollup	19.4 (+1%)	N/A	35,742	664	N/A
undefined	N/A	N/A	N/A	N/A	67,950 (+1%)

Stats on running time collected for app circuits

Function	input_size_in_bytes	output_size_in_bytes	witness_generation_time_in_ms
ContractClassRegisterer:register	1,344	11,731	343 (-1%)
ContractInstanceDeployer:deploy	1,408	11,731	18.1 (-1%)
MultiCallEntrypoint:entrypoint	1,920	11,731	456 (-1%)
FeeJuice:deploy	1,376	11,731	387 (-2%)
SchnorrAccount:constructor	1,312	11,731	173
SchnorrAccount:entrypoint	2,304	11,731	507
Token:privately_mint_private_note	1,280	11,731	226 (+5%)
FPC:fee_entrypoint_public	1,344	11,731	25.5 (+2%)
Token:transfer	1,312	11,731	408
Benchmarking:create_note	1,344	11,731	165 (-1%)
SchnorrAccount:verify_private_authwit	1,280	11,731	27.6 (+1%)
Token:unshield	1,376	11,731	630 (+1%)
FPC:fee_entrypoint_private	1,376	11,731	852 (+1%)

AVM Simulation

Time to simulate various public functions in the AVM.

Function	time_ms	bytecode_size_in_bytes
FeeJuice:_increase_public_balance	55.4 (+3%)	8,139
FeeJuice:set_portal	11.3 (+15%)	2,362
Token:constructor	101 (+22%)	31,107
FPC:constructor	55.5 (+8%)	22,380
FeeJuice:mint_public	38.3 (+2%)	6,150
Token:mint_public	53.8 (-29%)	11,720
Token:assert_minter_and_mint	⚠️ 68.3 (+76%)	8,028
AuthRegistry:set_authorized	44.8 (+25%)	4,537
FPC:prepare_fee	241 (+5%)	8,812
Token:transfer_public	30.7 (-6%)	47,374
FPC:pay_refund	65.2 (+3%)	12,114
Benchmarking:increment_balance	971 (-1%)	7,450
Token:_increase_public_balance	38.8 (-2%)	8,960
FPC:pay_refund_with_shielded_rebate	66.1 (-1%)	12,663

Public DB Access

Time to access various public DBs.

Function	time_ms
get-nullifier-index	0.164

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	256 leaves	512 leaves	1024 leaves
batch_insert_into_append_only_tree_16_depth_ms	2.23 (+2%)	3.87	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_count	16.8	31.7	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_16_depth_hash_ms	0.116 (+2%)	0.109	N/A	N/A	N/A	N/A	N/A
batch_insert_into_append_only_tree_32_depth_ms	N/A	N/A	11.5	17.8 (-1%)	31.1 (-1%)	59.5 (-2%)	114 (-2%)
batch_insert_into_append_only_tree_32_depth_hash_count	N/A	N/A	95.9	159	287	543	1,055
batch_insert_into_append_only_tree_32_depth_hash_ms	N/A	N/A	0.110	0.103 (-1%)	0.102 (-1%)	0.102 (-2%)	0.102 (-2%)
batch_insert_into_indexed_tree_20_depth_ms	N/A	N/A	14.3	25.9 (-1%)	44.0	82.7 (-2%)	162 (-2%)
batch_insert_into_indexed_tree_20_depth_hash_count	N/A	N/A	109	207	355	691	1,363
batch_insert_into_indexed_tree_20_depth_hash_ms	N/A	N/A	0.108	0.104 (-2%)	0.107 (+3%)	0.102 (-2%)	0.102 (-3%)
batch_insert_into_indexed_tree_40_depth_ms	N/A	N/A	16.4 (-1%)	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_count	N/A	N/A	132	N/A	N/A	N/A	N/A
batch_insert_into_indexed_tree_40_depth_hash_ms	N/A	N/A	0.105 (-1%)	N/A	N/A	N/A	N/A

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	64,779	668,997

Transaction size based on fee payment method

| Metric | |
| - | |