jdchrist@watcgl.waterloo.edu (Dan Christensen) (11/09/88)
I am working on a program that draws many spheres. These spheres are approximated by a set of polygons. Often I just want to draw a wireframe image of this sphere, so I construct a set of lines representing the edges of the polygons. I am interested in drawing these lines as quickly as possible. My questions are: 1) Will it make a significant difference if I draw the sphere as a series of connected paths rather than as separate line segments? In other words, is move( pos1 ); draw( pos2 ); draw( pos3 ); ... ; draw( posn ) going to be much quicker than move( pos1 ); draw( pos2 ); move( pos2 ); draw( pos3 ); ... move( posn-1 ); draw( posn ) ? 2) How much will the length of lines change this? 3) How much will precompiling the lines into an object affect this and how will removing the extra moves affect the precompiling speed? 4) What is a good way to experimentally check this? I am finding it difficult to do accurate timings, as timing how long it takes to perform the moves and draws, doesn't take into account how long they spend going through the pipeline. The timer stops before the drawing is actually drawn. Thanks in advance for any help. I'll post a summary. Dan