Automatically specify batch size for parallel iteration

[alice-i-cecile]

What problem does this solve or what need does it fill?

The parallel iteration tools that currently exist in Bevy require users to manually set a

As seen in https://github.com/bevyengine/bevy/blob/main/examples/ecs/parallel_query.rs.

What solution would you like?

As in #3183:

1. Rename the existing methods to par_for_each_manual and so on.
2. Remove the ability to set the batch size for the user.
3. Set the batch size automatically using some simple and low-overhead heuristic.

What alternative(s) have you considered?

More sophisticated strategies could be specified, but that is beyond the scope of this issue, which is focused on ergonomics.

Additional context

Works hand-in-hand with #3183, to finish improving the default ergonomics of parallel iteration.

Smarter control over the thread count should help improve the serious performance issues noted in #2173.

[alice-i-cecile]

Some design work (or benchmarking) is needed to determine what good defaults might be, but otherwise this is a simple PR.