I'm new to VEP. Actually, just doing my first VEP project and being challenged to rethink my workflow - how much I want to rely on the DAW vs how much i want to load into VEP. I'm thinking I may as well go the whole hog and load pretty much everything into VEP and just use my DAW as a 'sequencer' as suggested by the continuous usage of that nomenclature in VSL documentation.

So my question is, what is the typical reason/useage-case for needing more than a single instance in VEP?

What is the disadvantage of just loading everything into a single instance.

Specifically, I am looking for reasons that go beyond using multiple instances as purely an organizational function.

What are the technical reasons for multiple VEP instances in a single project?

(I use a range of DAWs)

The only times I instantiate an instrument in Cubendo is 1) to establish and/or continue automation of parameters via live interaction with its interface, or it's something that rather than take MIDI note-ons to make it sound generates its own sounds, so a performance of it has to be captured by a record operation; or 2) this is the first in well over a decade, an instrument whose performance in VE Pro is atrocious and it happens to perform quite well in Nuendo currently. This thing is not supported for my hardware or OS, and for whatever combo of reasons does not work well with the iLok VEP, it brings my system to its knees at once. But performs very well in Nuendo 11.

I mix using VE Pro's automation mapping. I do some further work with rendered audio in certain cases but the orchestration and that mix is all from host automation lanes. And at this time I'm only outputting one stereo out per instance. And two instances that are pretty different from each other, mainly to not have one unwieldy auto map in the single instantiation, another reason is to see what is using CPU more (this being some very different approaches, where the smaller instance is using more CPU (cymbals, tails...).