[do-not-merge] - packing gas testing

[iansuvak]

Why this should be merged

It shouldn't just a demo of testing packing efficency. If there are no problems with the testing method I will cut a PR to replace the existing packing method with the old one and keep the testing helpers to cover cases with multiple validators.

How this works

• Creates helper methods to generate subnetconversiondata with parametrized number of initialvalidators in _defaultSubnetConversionData
• Defines a _oldSubnetConversionDataPack to mimic the previous method of using abi.encodePacked in a loop.
• Defines 3 tests, 2 to test gas usage for old (abi.encodePacked)and new (manual byte by byte packing) methods, and a third one to ensure that their outputs match
• Changed the input type from bytes calldata to bytes memory. This is a significant change since reading from calldata is a lot cheaper than memory but I don't think that it should change the outcome of the test since it's still an apples to apples comparison.

How this was tested

Ran manual forge tests with different num InitialValidators counts:
Below are the results showing that doing abi.encodePacked in a loop is a lot more efficient given this testing methodology.

- numValidators: 1
[PASS] testSubnetConversionDataPacking() (gas: 69619)
[PASS] testSubnetConversionDataPackingNew() (gas: 66295)
[PASS] testSubnetConversionDataPackingOld() (gas: 5204)

- numValidators: 5
[PASS] testSubnetConversionDataPacking() (gas: 222986)
[PASS] testSubnetConversionDataPackingNew() (gas: 211833)
[PASS] testSubnetConversionDataPackingOld() (gas: 14351)

- numValidators: 10
[PASS] testSubnetConversionDataPacking() (gas: 419291)
[PASS] testSubnetConversionDataPackingNew() (gas: 393842)
[PASS] testSubnetConversionDataPackingOld() (gas: 30222)

- numValidators: 50
[PASS] testSubnetConversionDataPacking() (gas: 2202742)
[PASS] testSubnetConversionDataPackingNew() (gas: 1851161)
[PASS] testSubnetConversionDataPackingOld() (gas: 363856)

How is this documented