ocnxlRun() without the need to restart simulator

Hello again!

There is a chance that what I'm trying to do isn't possible, but I'll try my luck anyways:

Essentially, I am creating some algorithms that should help me optimize circuit parametrization (The scripts are outside of Cadence). For this, I have been editing and running OCEAN scripts to automatically run simulations. My issue is though, is that when I want to run a simulation using OCEAN, I am forced to set the target cellview (a maestro session) using ocnxlTargetCellview() before running the simulation using ocnxlRun(). Why exactly is this an issue, is that I cannot have the maestro cellview opened prior to the simulation, which means that for every single iteration (If I'm doing singlepoint simulations) the OCEAN script must open the simulator from scratch, causing the simulations to take a very long time per iteration.The ocnxlTargetCellview function will simply not run if the targeted cellview is already open.

I've measured time on some of my simulations and come to the conclusion that the time it takes for a single iteration using OCEAN versus manually doing it in maestro to be around 45 seconds versus 5 seconds. The manual method of course has the simulator already open, skipping the setup time for the OCEAN script. So my question is simple, is it possible for me to somehow keep the maestro session in memory, so that even after the previous singlepoint simulation ends, the next iteration can start with new design variables without having to load the entire session again? So far it seems to me that no matter what I do, the OCEAN script must slowly open everything up again at the beginning of every run.

I also acknowledge that I can run simulations in OCEAN in batch mode using lists of input variables. This cuts down the time per simulation by a considerable amount, as the setup is only done before the first simulation, but for my purposes being able to loop singlepoint simulations would be much more useful.

I've found some old threads talking about slow simulations, and even tested their methods, but came to the conclusion that the setup time before the simulation is the bottleneck in my simulations, and not the actual simulation time itself.

Any ideas or suggestions appreciated! Thanks in advance! I'm on IC23.1 if that is of any relevance or help.