Hc00: My PC is a non-HT Pentium IV, 2.4ghz, with 1GB RAM . I recently upgraded to an NVIDIA 5200 card with 256MB and that made a massive difference in terms of 3D view redrawing speed. Before that I was using an old NVIDIA card (420) with 64MB RAM, which was way overtaxed by all the high-poly projects I’ve worked on these past few years (including my “Great Sea Race” for the F1 competition). By using layers and deactivating subsurf for the 3D view (but leaving it active for render) I am able to minimize lag as a project progresses. Sometimes I push too hard though and my computer freezes up. My laptop, which has 2GB, suffers less from VM-swapping, but even where that has a Go-6800/256MB card, my busier projects often bring that to a near halt. Maybe in the next year or two I will move to a more capable system, preferably 4 or 8 cores just for the potential multithreaded rendertime gains there, and a better GPU for improved realtime response. There are many other ways of handling potential gigapoly projects though, including rendering out individual items or layers and then compositing in Blender or in post. In my experience, I’ve found that, while there are easily reached technical limits for the computer, there are usually workarounds to be found and different modeling/texturing strategies (such as normal mapping) to be tried to enable us to make far more complex works while maintaining a workable workspace and optimal geometry. About the skulls (there’s more than one in the lower half of the picture): they are actually out a bit in front of the larger skeleton, caught in the teeth of the machine, which has consumed countless ones before. There were more yet, but for compositional purpose I went with this final arrangement.