Depends what you mean by optimization.
On a PC you do everything you can to reduce draw calls, including copying buffers and coalescing things on the CPU, because the cost of that copying/coalescing is less than the cost of the draw call overhead.
On a console in general you try to minimize the copying involved because the bulk of the overhead in a draw call is the copy.
On a console you might want to submit world models broken into smaller pieces because the GPU has to do less work on offscreen or hidden polygons, on a PC you'd want larger pieces to reduce draw call overhead.
This is what people don't seem to understand about optimizations on a console, ignoring algorithmic improvements, rarely do you discover new tricks that allow things to run faster later in a console life, what you understand is how data flows through the system, and where time is spent, that let's you be smarter about how you organize the data. You simply can't do optimizations like this on a PC, because different hardware behaves differently, so you "optimize" for the worst cases across the hardware.