Improve equals performance on Address

[ahamlat]

PR description

We've noticed when profiling the SELFBALANCE during load testing that the throughput issue is related to the equals operation on the address class.

This PR introduces a new implementation of the equals method in the Address class to replace the one inherited from AbstractBytes. The issue with equals in AbstractBytes is that it compares each byte from both addresses individually, and the get method used is megamorphic.

Fixed Issue(s)

Thanks for sending a pull request! Have you done the following?

• Checked out our contribution guidelines?
• Considered documentation and added the doc-change-required label to this PR if updates are required.
• Considered the changelog and included an update if required.
• For database changes (e.g. KeyValueSegmentIdentifier) considered compatibility and performed forwards and backwards compatibility tests

Locally, you can run these tests to catch failures early:

• unit tests: ./gradlew build
• acceptance tests: ./gradlew acceptanceTest
• integration tests: ./gradlew integrationTest
• reference tests: ./gradlew ethereum:referenceTests:referenceTests

[macfarla]

do you need to override hashcode() because you've overridden equals()?

[ahamlat]

The current implementation of the hashcode in AbstractByte is good IMO, it is also stored to avoid recalculation. That's why I didn't override it.

[shemnon]

Consider using toArrayUnsafe, which will use local copies of the bytes if they exist, and will have less GC churn. (if a local copy does not exist it will go back to toArray)

As long as we never manipulate the bytes that are returned it will be safe. The Unsafe part only denotes we are exposing an internal array.