[FEATURE]: Sync optimization

[mvandeberg]

Is there an existing issue for this?

• I have searched the existing issues

New feature

The sync code currently tries applying all reversible peer blocks. This is usually 60 blocks and the difference between the current node is likely only a block or two (if any).

Instead, the sync code (once the node is synced) perform a rough binary search to find the first block that differs from the peer and the current node. The search should be fast and does not need to be incredibly accurate. It should err on the side of applying blocks again rather than missing blocks. Any reduction is duplication is significant at this point.

Anything else?

No response