Add gas benchmarks for BatchLockup

[smol-ninja]

Changelog

• Closes Add benchmark dir in tests #334
• Adds gas benchmarks for batchsize: [5, 10, 20, 30, 50] and segments/tranches: [24, 24, 24, 24, 12] respectively.

Question: Do we also want benchmark tables for MerkleLockupFactory and MerkleLL/LT contracts? IMO that should not be necessary since Airstream campaign creators are usually the companies which have very few incentive to care about transaction cost.

[PaulRBerg]

Good job!

Reminder to update Ward Analytics (see Telegram) about this once we merge this PR.

Do we also want benchmark tables for MerkleLockupFactory and MerkleLL/LT contracts?

No, as per your rationale.

[smol-ninja]

@andreivladbrg can you please review and approve it?

[andreivladbrg]

Before giving a full review, I have a question.

Adds gas benchmarks for batchsize: [5, 10, 20, 30, 50] and segments/tranches: [24, 24, 24, 24, 12] respectively

I guess the reason for adding 12 count for a batch of 50 size (instead of 24) is because it's close to the block gas limit. But what's the reason for not using the 12 everywhere instead?

Also, would it be useful if we add more counts for each batch size? e.g. 2,12,24 etc

[smol-ninja]

I guess the reason for adding 12 count for a batch of 50 size (instead of 24) is because it's close to the block gas limit. But what's the reason for not using the 12 everywhere instead?
Also, would it be useful if we add more counts for each batch size? e.g. 2,12,24 etc

Please see the discussion here.

[andreivladbrg]

LGTM 👍

Should we also add a benchmark test to see the size for the value close to 20 million gas for the linear batch function? The current gas usage for a size of 50 is far from exceeding the limit.

[smol-ninja]

Should we also add a benchmark test to see the size for the value close to 20 million gas for the linear batch function? The current gas usage for a size of 50 is far from exceeding the limit.

The batch size depends on number of segments/tranches as well. So for what values of segments/tranches, would we do it for?

[andreivladbrg]

The batch size depends on number of segments/tranches as well. So for what values of segments/tranches, would we do it for?

I am talking about LockupLinear, not Dynamic or Tranched

[smol-ninja]

For LockupLinear, the gas consumption should increase linearly with the batch size. So based on (50, 6.87M), it should be $50*(30M/6.87M) = 218$. I ran the test, and its indeed ~218.

| `createWithDurationsLL`  | Lockup Linear   | N/A               | 218        | 30,037,705  |
| `createWithTimestampsLL` | Lockup Linear   | N/A               | 218        | 30,050,659  |

I can add another test contract to generate this but do we need it? If you want to add to this table, then thats very easy. But if you want to write a script similar to the calculation of MAX_SEGMENT_COUNT in v2-core, then it is a bit of work.

PS: we can also add this ~218 number in the Sablier docs of BatchLockup.

PS: my opinion is there is no need to know the value for max batch size in LL as it can easily be calculated from the data in the gas table we are providing.