based on the responses it's clear that my question sounded like a leading one. it was not and i employ both methods actively.
- It means you can render partials to the user before you receive data from the server. So things seem to happen instantaneously to the user.
- It also means you can have your partials rendering in a single place and rendered by a single template engine.
performance was admittedly awfully misdirected in a general context. i suppose my situation is unique; several templates on my app require complex logic or additional requests based on properties of the initial request. this is where you start to see it get hairy.
again though, admittedly not the normal scenario. i appreciate you taking my question literally, as it was meant.