Use aiken for commit validator #1680

ch1bo · 2024-10-04T08:02:56Z

We want to update to a recent cardano-api version to get #1543 (which we contributed upstream). This update implied a bump on plutus-tx (via the ledger) and resulted in growing script sizes. At that point we explored various ways to decrease the script size again to keep the single transaction publishing and --hydra-scripts-tx-id config (or even expanding that).

This mikado / dependency graph shows our option and the chosen path of this PR:

Concretely this PR contains:

Update to cardano-api version 9.3
Rebased and updated to latest aiken the commit validator from this work: Aiken: commit validator #1072
Support PlutusV3 validators next to legacy PlutusV2 validators (the aiken one is V3)

Consequences of switching to aiken:

🚀 reduces cost to open (collect) a head, for example 5 parties from 1.00 -> 0.81 ADA =~ 20% cost saving
🚀 reduces cost to abort a head, for example 5 parties from 1.27 -> 0.85 ADA =~ 50% cost saving
🚀 supports collect transaction with 9 parties (abort 12) = new maximum number of parties is 8 (tests showed we need some margin)

CHANGELOG updated
Documentation update not needed
Haddocks updated
New TODOs introduced
- One as a reminder that we have duplication of types between aiken/haskell

We intend to create one of the hydra-plutus scripts with aiken.

This separates the aiken build, which yields the plutus.json blueprint and the haskell build which provides access to the compiled validators.

Instead of using the plutus-tx "compiledScript", we use the loaded-from-blueprint aiken script in the init validator and when constructing/observing transactions in the hydra-node.

This is not yet working and it does not correctly detect changes to plutus.json

Plutus-tx seems to be encoding tuples as a Constr on-chain.

Also add TODO comments for things that are missing to be removed.

… script info

Seems like the aiken compiler uses plutus core version 1.1.0 syntax and we need to declare it as a PlutusV3 script.

We do not need the plutus-style envelope and this would not match with the PlutusV2 used there.

This will keep the first trace argument which we expect to see in our mutation tests.

github-actions · 2024-10-04T16:35:42Z

Transaction costs

Sizes and execution budgets for Hydra protocol transactions. Note that unlisted parameters are currently using arbitrary values and results are not fully deterministic and comparable to previous runs.

Metadata
Generated at	2024-10-04 18:07:30.923828404 UTC
Max. memory units	14000000
Max. CPU units	10000000000
Max. tx size (kB)	16384

Script summary

Name	Hash	Size (Bytes)
νInitial	b512161ccb0652d7e9a0b540e4a3c808f73d6558a4bcabf374d85880	3969
νCommit	ea444d37d226e71eef73ac78d149750da977feb588900135bf9e8221	692
νHead	2253ddd95837c7aacc8635a971caaea743434152dd8dd2849bdf4162	10797
μHead	4d648ca239040b0e87901835aa11423e7aa3bd947ce6befe7db1bae8*	4508
νDeposit	1a011f23b139a6426767026bde10319546485d553219a5848cdac4e5	2993

The minting policy hash is only usable for comparison. As the script is parameterized, the actual script is unique per head.

`Init` transaction costs

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	5097	5.81	2.30	0.44
2	5298	7.09	2.80	0.46
3	5502	8.56	3.39	0.49
5	5901	11.32	4.48	0.53
10	6906	18.02	7.12	0.65
57	16353	82.89	32.79	1.78

`Commit` transaction costs

This uses ada-only outputs for better comparability.

UTxO	Tx size	% max Mem	% max CPU	Min fee ₳
1	569	10.84	4.26	0.29
2	756	14.31	5.80	0.34
3	942	17.92	7.39	0.39
5	1316	25.56	10.73	0.49
10	2254	47.11	19.97	0.77
19	3951	94.71	39.81	1.38

`CollectCom` transaction costs

Parties	UTxO (bytes)	Tx size	% max Mem	% max CPU	Min fee ₳
1	57	560	19.87	7.59	0.39
2	114	671	28.85	10.97	0.49
3	170	786	37.67	14.32	0.59
4	227	893	41.33	15.76	0.64
5	281	1004	48.49	18.48	0.72
6	339	1120	61.07	23.20	0.86
7	394	1227	75.79	28.74	1.02
8	448	1338	82.50	31.32	1.10
9	504	1449	83.79	31.89	1.12
10	560	1560	92.85	35.30	1.22

Cost of Decrement Transaction

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	625	18.64	8.15	0.39
2	723	18.46	8.82	0.39
3	981	22.44	11.17	0.46
5	1121	22.68	12.79	0.48
10	2135	34.80	21.37	0.68
49	7919	97.72	75.96	1.84

`Close` transaction costs

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	664	20.83	9.32	0.42
2	741	21.93	10.44	0.44
3	956	24.07	12.49	0.48
5	1349	27.79	16.08	0.55
10	2059	35.85	23.85	0.70
50	7850	97.75	84.99	1.90

`Contest` transaction costs

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	680	26.77	11.46	0.48
2	860	28.98	13.38	0.52
3	1021	30.60	14.88	0.55
5	1277	34.23	18.03	0.61
10	2094	44.47	26.95	0.80
39	6483	99.14	75.26	1.78

`Abort` transaction costs

There is some variation due to the random mixture of initial and already committed outputs.

Parties	Tx size	% max Mem	% max CPU	Min fee ₳
1	4984	15.34	6.57	0.54
2	5049	21.29	9.07	0.61
3	5214	26.49	11.36	0.68
4	5251	29.50	12.48	0.71
5	5548	41.79	18.29	0.86
6	5591	46.04	19.90	0.91
7	5828	54.16	23.59	1.01
8	5921	64.24	27.96	1.13
9	6098	67.40	29.46	1.18
10	6202	78.14	34.10	1.30
11	6262	81.87	35.58	1.34
12	6465	91.45	39.92	1.46
13	6557	93.05	40.39	1.48

`FanOut` transaction costs

Involves spending head output and burning head tokens. Uses ada-only UTxO for better comparability.

Parties	UTxO	UTxO (bytes)	Tx size	% max Mem	% max CPU	Min fee ₳
10	0	0	5090	9.79	4.09	0.48
10	1	57	5124	11.35	4.98	0.50
10	5	284	5258	15.63	7.71	0.56
10	10	569	5428	21.56	11.36	0.65
10	20	1137	5767	33.75	18.81	0.81
10	30	1707	6109	45.04	25.88	0.97
10	40	2275	6447	56.93	33.20	1.13
10	50	2849	6790	68.83	40.53	1.30
10	77	4384	7704	99.85	59.86	1.73

End-to-end benchmark results

This page is intended to collect the latest end-to-end benchmark results produced by Hydra's continuous integration (CI) system from the latest master code.

Please note that these results are approximate as they are currently produced from limited cloud VMs and not controlled hardware. Rather than focusing on the absolute results, the emphasis should be on relative results, such as how the timings for a scenario evolve as the code changes.

Generated at 2024-10-04 18:08:54.574052356 UTC

Baseline Scenario

Number of nodes	1
Number of txs	3000
Avg. Confirmation Time (ms)	4.907318410
P99	11.039467829999907ms
P95	6.361053499999997ms
P50	4.2919015ms
Number of Invalid txs	0

Three local nodes

Number of nodes	3
Number of txs	9000
Avg. Confirmation Time (ms)	23.495834438
P99	114.87703296000002ms
P95	30.58308695ms
P50	20.973222ms
Number of Invalid txs	0

This does not require us to be in a git repository and hence makes the test more flexible and isolated.

The hydra-plutus library seems to be consistently rebuilt when the plutus.json gets changed.

This was failing with PPViewHashesDontMatch because the transaction was referencing the commit script which is does not always need, and the commit script being of different version than the other scripts spent.

github-actions · 2024-10-04T18:17:54Z

Test Results

518 tests ±0 512 ✅ ±0 32m 38s ⏱️ + 9m 28s
163 suites ±0 6 💤 ±0
7 files ±0 0 ❌ ±0

Results for commit 665cc17. ± Comparison against base commit d3db623.

This pull request removes 1 and adds 1 tests. Note that renamed tests count towards both.

Hydra.Options/Hydra Node RunOptions ‑ validateRunOptions: using more than 5 parties should error out

Hydra.Options/Hydra Node RunOptions ‑ validateRunOptions: using more than 8 parties should error out

CHANGELOG.md

Addresses #1543 and tests whether IntersectMBO/cardano-api#632 results in `GET /snapshot/utxo` to contain `inlineDatumRaw`. Bonus: First step on consolidating `TxOut` and `UTxO` generators (deduplicating and moving them to a common module) TODO: Somehow get IntersectMBO/cardano-api@17eb46f into this branch. Most likely requiring #1680 (which updates to most recent `cardano-api` release `9.30`) and a `source-repository-package` onto some unreleased cardano-api. --- * [x] CHANGELOG updated or not needed * [x] Documentation updated or not needed * [x] Haddocks updated or not needed * [x] No new TODOs introduced or explained herafter

ch1bo assigned ch1bo and locallycompact Oct 4, 2024

ch1bo mentioned this pull request Oct 4, 2024

cardano-api: 9.1 -> 9.3 #1673

Closed

4 tasks

ch1bo linked an issue Oct 4, 2024 that may be closed by this pull request

Raw CBOR datum in transaction outputs #1543

Closed

ch1bo removed a link to an issue Oct 4, 2024

Raw CBOR datum in transaction outputs #1543

Closed

This was referenced Oct 4, 2024

Raw CBOR datum in transaction outputs #1543

Closed

Optimise script size for PlutusV3 #1665

Closed

plutus: 1.30 -> 1.34.1 #1544

Closed

locallycompact force-pushed the aiken-commit-validator branch from bae8866 to 38e1743 Compare October 4, 2024 13:23

ch1bo force-pushed the aiken-commit-validator branch 2 times, most recently from 7bafe31 to 296395c Compare October 4, 2024 14:12

ch1bo requested a review from a team October 4, 2024 15:23

ch1bo force-pushed the aiken-commit-validator branch 3 times, most recently from 9d5ca8d to 27b506b Compare October 4, 2024 16:18

locallycompact and others added 15 commits October 4, 2024 18:26

Update cardano-api: 9.1 -> 9.3

3e288c3

Add aiken to our nix develop shell

bbee1ba

We intend to create one of the hydra-plutus scripts with aiken.

Draft an aiken project in hydra-plutus

c9e19c9

Draft an empty commit aiken validator and commit plutus.json

4343b66

Create a module to load blueprint json

3b98df5

This separates the aiken build, which yields the plutus.json blueprint and the haskell build which provides access to the compiled validators.

Replace plutus-tx commit with the aiken commit script

9f4ac0f

Instead of using the plutus-tx "compiledScript", we use the loaded-from-blueprint aiken script in the init validator and when constructing/observing transactions in the hydra-node.

implement 'ViaCollectCom' path in Aiken

ef8b7d5

Implement 'ViaAbort' road on commit validator in Aiken.

4c6bc03

Try to make plutus.json a extra source file

eb368a7

This is not yet working and it does not correctly detect changes to plutus.json

Decode commit datum as a record

f706f7a

Plutus-tx seems to be encoding tuples as a Constr on-chain.

Add golden spec to verify plutus blueprint

0e320b1

Add commit errors to the validator script

f0551ca

Add comments to the aiken commit validator script

7e3e680

Remove commit validator

e7fb038

Also add TODO comments for things that are missing to be removed.

Add test to check plutus blueprint hash against the code reference in…

15f587d

… script info

ffakenz and others added 11 commits October 4, 2024 18:26

Upgrade aiken project version to match next release

54f1a26

Add a CIP-57 todo

210f65c

Re-add golden tests for commit validator

a3f4865

Add aiken as a build dep to hydra-plutus tests

4ad41dc

Update changelog

7ff4ecc

Compiling aiken script

5c4c79f

Pretty print reference scripts in renderTxWithUTxO

e4851b2

Use aiken scripts as PlutusV3 script

2527fac

Seems like the aiken compiler uses plutus core version 1.1.0 syntax and we need to declare it as a PlutusV3 script.

Add/update plutus v3 cost models

8d5590e

Drop commit from inspect-script exe

d49a2a2

We do not need the plutus-style envelope and this would not match with the PlutusV2 used there.

Rebuild commit aiken validator with -t compact

c52c7af

This will keep the first trace argument which we expect to see in our mutation tests.

ch1bo force-pushed the aiken-commit-validator branch from 27b506b to 5dee79a Compare October 4, 2024 16:26

locallycompact and others added 6 commits October 4, 2024 19:00

Simply compare the plutus.json after regeneration

90b32b8

This does not require us to be in a git repository and hence makes the test more flexible and isolated.

Drop a redundant test and FIXME

55dd765

The hydra-plutus library seems to be consistently rebuilt when the plutus.json gets changed.

Fix publishHydraScripts to use V3 for the commit script

db10f25

Only reference commit script if spending commits in abort

75a0775

This was failing with PPViewHashesDontMatch because the transaction was referencing the commit script which is does not always need, and the commit script being of different version than the other scripts spent.

Update changelog and set maximumNumberOfParties to 9

f4d379a

Bump versions and regenerate scripts

f9e07ce

ch1bo force-pushed the aiken-commit-validator branch from 5dee79a to f9e07ce Compare October 4, 2024 17:02

ch1bo mentioned this pull request Oct 4, 2024

Add inlineDatumRaw to transaction outputs in API #1684

Merged

4 tasks

Only 8 participants are possible consistently

665cc17

v0d1ch approved these changes Oct 7, 2024

View reviewed changes

CHANGELOG.md Show resolved Hide resolved

ffakenz approved these changes Oct 7, 2024

View reviewed changes

locallycompact added this pull request to the merge queue Oct 7, 2024

Merged via the queue into master with commit 5e6eca0 Oct 7, 2024
28 checks passed

locallycompact deleted the aiken-commit-validator branch October 7, 2024 10:42

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Use aiken for commit validator #1680

Use aiken for commit validator #1680

ch1bo commented Oct 4, 2024 •

edited

Loading

github-actions bot commented Oct 4, 2024 •

edited

Loading

github-actions bot commented Oct 4, 2024

Use aiken for commit validator #1680

Use aiken for commit validator #1680

Conversation

ch1bo commented Oct 4, 2024 • edited Loading

github-actions bot commented Oct 4, 2024 • edited Loading

Transaction costs

Script summary

Init transaction costs

Commit transaction costs

CollectCom transaction costs

Cost of Decrement Transaction

Close transaction costs

Contest transaction costs

Abort transaction costs

FanOut transaction costs

End-to-end benchmark results

Baseline Scenario

Three local nodes

github-actions bot commented Oct 4, 2024

Test Results

ch1bo commented Oct 4, 2024 •

edited

Loading

github-actions bot commented Oct 4, 2024 •

edited

Loading

`Init` transaction costs

`Commit` transaction costs

`CollectCom` transaction costs

`Close` transaction costs

`Contest` transaction costs

`Abort` transaction costs

`FanOut` transaction costs