I do not, nor do I think that Tom, or 'anyone' at this level of involvement in VRs for that matter, has access to the precise algorithms or the rule set of the PMR VR. In general, my understanding, is that we are talking multiple levels by which the calculation requirements are reduced; what I call computational parsimony. If physical reality is expressed as fractals at multiple levels or scales, then you only have to deal with those levels of which an IUOC can be aware at a given time. That means you deal with lights in the sky rather than cosmological calculations and dirty water rather than bacteria, etc. for most of the existence in time of PMR and beyond into the future for most of the billions of IUOCs 'living' here. You don't suddenly have to start doing quantum calculations just because someone invented quantum mechanics or celestial calculations just because someone invented a telescope. Only if a device capable of making such an observation is active and an observation being made.
Then there is a further resulting reduction in required calculations that occurs because you, for instance, never have to calculate and keep track of the identity of ever grain of sand on the beach or star in the sky. Only if you are an IUOC lying there sunbathing with your nose close enough to distinguish individual sand grains or perhaps are a mollusc of some kind squirming it's way through the sand at the tide line. Even then, you don't have to keep or calculate detailed records. A sunbather is incapable of keeping track of or observing at a level that these details are available, much less that they need to be remembered. All they need is the sun glittering off of the wet granules. And any kind of micro fauna operates at such a low level that the model of PMR does not have to be fully populated with all of their varieties and numbers. A dung beetle rolling a ball of manure is a dung beetle rolling a ball of manure. A gut bacteria is a gut bacteria. The gut bacteria of all of the PMR reality in all of the bellies of PMR entities do not have to be allocated and have their environments modeled for them. When they are being observed, their population exists where observed and their environment is fully detailed when it is observed, perhaps by a medical scientist or technician investigating for the cause of an infection. Otherwise, your gut bacteria can be simulated perfectly well by the VRRE in terms of your sensory perceptions. And the environment of the IUOC taking the role of the gut bacteria can have it's environment also simulated perfectly well by the VRRE. A relatively few IUOCs are actually thus required to function in the roles of gut bacteria, only where they are under observation.
Smoke and mirrors. It's all illusion to the fullest extent possible. Every saving in computational resources is a saving in computational resources. I suspect, rather than know, that eventually they will come up with explanations for dark matter and dark energy. But then again, perhaps not. If you think about the above described techniques for creating this VR and the resulting VR experience, for the vastly long time of it's existence, there has been and in fact, might never have been, any reason to calculate the properties of quantum level 'particles'. The same goes for the full cosmos as revealed to us by the Hubble telescope and other advanced viewing devices such as radio and x-ray telescopes. Would all of the rules required to make sense to PMR scientists of the present and future have been put into the system if only needed far into the future and possibly not even then? And the purpose was not making a sensible system for future PMR scientists, but rather for creating all of the interactions for IUOCs to improve their quality of being. The system was not intended, after all, to last forever. You can always start up a new one if needed. I would not be at all surprised if at some time in the future that it is discovered that the explanations for dark energy and dark matter amount to 'fudge factors' needed to make the present equations of PMR science work as we expect. Something like the epicycles of Ptolemaic astronomy, used basically as fudge factors to make the model work. Perhaps some cosmic engineer has had to explain why his committee never thought it would be necessary to set the equations up to account for the anomalous observations of galaxies far, far away as opposed to just making it look good. Will it be good enough to say, 'how were we to ever know that anyone would ever be able to tell the difference?'
To go with a point by point style model and account for everything existing at a give point is clearly very intensive in computational resources. Then you are not only keeping track of every grain of sand, but of every sub atomic particle. We are not talking 'holodeck' here. Perhaps I don't understand the method of simulation you are talking about. After all, when I did early modeling, the IBM mainframe used probably had less power than my present desktop. And since between my NAS and installed drives, I have nearly 3TB of storage, I know I have more disk storage than we, the whole university, had then and mine is all on line. I don't remember how much a disk pack held, but I remember that it wasn't much and getting one assigned to you alone was an achievement. Otherwise why would we load in a carton or two of punched cards for every run rather than just store the program on disk and recompile the whole program, rather than just a subroutine every time as needed? That was long ago and in a galaxy far away. I am way out of date now.