I am currently working on a network with multiple subnets. So far I have implemented one reverse DNS entry that is a monolithic 10.* record covering everything. Is there any benefit or performance boost for creating a granular structure allowing look-ups to occur quicker? Or is this more work than I need to perform?
Generally speaking this is more work than you need to perform -- Most DNS servers can handle a /8 dumped in one file without batting an eyelash in terms of performance.
Having said that, traditionally reverse zones (for IPv4) are broken out as Class C (/24) netblocks --