Yes indeed! That's the implementation we're using for the extraction step at the moment. At some point we'll adapt a version of the recently developed NOISE algorithm (Near Optimal Isosurface Extraction) for out-of-core use.
Tom On Mon, 17 Jan 2000, ballard andrews wrote: > Tom, > > You may know this (but just in case), > the Marching Cubes algorithm in VTK is faster > than the version implemented in DX. > > Tom Bartol wrote: > > > > Volume rendering is not what we want to do because we're not just trying > > to make pretty pictures. We need the polygons as input to computer > > simulations of biochemistry within the 3D space defined by the isosurface. > > > > Thanks for the suggestion, though. :-) > > > > Tom > > > > On Mon, 17 Jan 2000, ballard andrews wrote: > > > > > Have you tried volume rendering? > > > Its not hard to write a DX module that uses > > > the SGI Volumizer API to render large data sets > > > (512^3) in real time on a single pipe. > > > > > > Tom Bartol wrote: > > > > > > > > On Mon, 17 Jan 2000, Chris Pelkie wrote: > > > > > > > > > >My colleagues and I are currently considering presenting our work > > > > > >with > > > > > >OpenDX at VDE2000. > > > > > > > > > > > >Thanks for your offer! > > > > > > > > > > > >Tom > > > > > > > > > > > > > > > > > > > > > > Good! Tom B. has some frighteningly large data sets (bits of brain and > > > > > stuff), as I remember. Plus, I assume you'd take your FreeBSD openDX, > > > > > right? > > > > > > > > > > Chris Pelkie > > > > > > > > You bet! Joel and I would come equipped with our Thinkpad 770's (mine > > > > with FreeBSD, Joel's with Win95) running OpenDX. Yes, our datasets are > > > > rather large 3D reconstructions of muscle and brain synapses. Our > > > > "small" > > > > datasets contain 10's of millions of triangles. The marching cubes > > > > isosurface extraction step takes a couple Gigs of RAM. We're not sure > > > > how > > > > we're going to do process our "large" datasets. We'll need algorithms > > > > that operate "out of core" (i.e. to and from disk cache). And you can > > > > just > > > > forget about rendering these in DX. RenderMan is very efficient and can > > > > render out of core (Level of Detail only gets you so far -- how do you > > > > think they rendered every leaf and blade of grass in "Bug's Life"?). > > > > Issues related to this would be an interesting topic for discussion at > > > > VDE2000. > > > > > > > > Tom > > > > > > -- > > > Dr. A. Ballard Andrews > > > Senior Research Scientist > > > Schlumberger Doll Research > > > Old Quarry Road Ridgefield, CT 06877 > > > tel: 203-431-5522 fax: 5521 > > > > > -- > Dr. A. Ballard Andrews > Senior Research Scientist > Schlumberger Doll Research > Old Quarry Road Ridgefield, CT 06877 > tel: 203-431-5522 fax: 5521 >
