Thus spake Thomas Russ: > > On Feb 5, 2007, at 12:37 PM, Joel Uckelman wrote: > > > Thus spake Thomas Russ: > >> > >> Random thought: Given the improvement in graphics card capability, I > >> wonder a bit about whether using > >> only a single map image internally and doing the zooming using > >> graphcis transformations might not be more responsive? > >> > > > > That had occurred to me as well. The GPUs on modern graphics cards are > > basically overgrown matrix multipliers, and all scaling is is a matrix > > operation. The problem is, I don't know how to get at any of the 3D > > card > > stuff through Java. If you have any suggestions, I'd be happy to look > > into them. > > I would have expected the acceleration to come about via the > implementation of the JVM and thus be relatively transparent > to the programmer. What I could find with a little browsing of > the Java 2D interface documentation supports this conclusion. > > (See for example > <http://java.sun.com/j2se/1.5.0/docs/guide/2d/new_features.html> > <http://java.sun.com/j2se/1.4.2/docs/guide/2d/spec/j2d-bookTOC.html> > ) > > I don't know if that would be directly useful here, but taking a look > at VolatileImage may be worthwhile, since it provides the ability to > use hardware acceleration during operations.
Follow-up: I've since learned that the JVM won't use any 3D capabilities of my particular graphics card because the driver is too old, and it's unlikely that there will ever be a newer one. So, that means I can't test any changes which would use the Java2D OpenGL pipeline until summer when I plan to get a new laptop. Sigh. Thanks to everyone who suggested things in this thread. I plan to revisit this particular issue in a few months when I have hardware which permits it. -- J.
