I'll throw a guess too.
It seems that whatever is on screen at the time, regardless of the count
of the assembly, directly affects model manipulation. I have noticed
that if an assembly is "zoomed all", manipulation stays rather
responsive and stays that way even as the model is zoomed in.
However, once the model is zoomed in very close, the zooming out usually
takes considerable time (the pregnant pause) and will even worsen if a
highly (feature) patterned part becomes visible in the process. At this
point, a rotate becomes painful as IV's graphic seemingly tries to
"read" the information.
It would seem that at a "zoomed all" model somehow (as long as the
card's memory can hold it) can extrapolate any rotate manouvers. It also
seems that card memory is dumped as parts disapear off screen, but
re-built much slower/painfully as they come back into view.
As I understand it, IV doesn't use a "list" like MDT does, but at the
same time, it would seem that if a "zoomed out" assembly can be
manipulated in real time, why can't a "zoomed in" portion of the model
be manipulated as fast and without the pregnant pause.... isn't the
(viewing) information for the parts already "in" the card's memory?
I wonder if there is a "task manager" application for video cards. I
would just love to see how much memory is actually being used and how
dynamically it is being utilized.
QBZ
"deckart" wrote in message
news:f18ac0d.6@WebX.maYIadrTaRb...
> That is the most informed guess I've got and it is consistent with the
research I've done. That and Autodesk doesn't seem to hint at the
advantage of having a large amount of video memory. The Ti4200 128Meg is
looking like the best "bang-for-the-buck" upgrade choice.