Here is a simple test you can do to check your cards potential power:
Make a circle.
Duplicate in a 100x100x1 array where they do not overlap.
Zoom extent in perspective view
Run TestMaxSpeed
(my score: 1.61 seconds)
Note: When TestMaxSpeed runs it downgrades the circle curves to polylines for faster redraw
Undo so you only have one circle
Make a planar surface from the circle
Delete curve (but keep surface)
Mesh with default mesh setting in simple mode
delete surface (but keep mesh)
array 100x100x1
zoom extent
Run Test Maxspeed
Note: As you can see the meshing is finer than the polyline curve conversion, and there are many more lines showing the triangulation.
(My score: 1.81 seconds)
Now select all meshes and JOIN
Run TestMaxSpeed
(My scores: 0.5 seconds)
The speed up from 10000 meshes to one is big (which is not strange), so could V6 potentially handle a block (kind of) like it handles a joined mesh, and achieve similar speed gains?
Possibly, but there are a lot of cases where this falls apart. Specifically when there are any display conduits turned on that pay attention to per object attribute setup or drawing operations.
The routine used for drawing the circle or the brep edge is very different than the mesh edge curve drawer.
How about making that an option like “draw blocks simplified” and “draw blocks as bounding boxed”?
So if you edit a block, it’s drawn with all bells and whistles. Once you are finisehd editing, you can reduce drawing overhead but still see some representation of th block geometry.
That makes sense Steve. But maybe we could have a “speed mode” that could benefit from this, and that ignored some of the per object attributes… and that ONLY added these when the screen stood still. On large projects i would easily use that option.
And just for comparison:
Laptop with i5 and GeForce 330M speeds:
Curves: 2.11 seconds
10000 meshes: 6.79 seconds
1 joined mesh: 4.43 seconds