Performance

Deprecated Sequoia Beta Discussion
1

I compared performance of sequoia with 3dsmax2016 on my laptop. My laptop has Nvidia Quadro K1100M GPU.
I used 26M points data. And almost similar setting for meshing.

sequoia / 3dsmax
point cloud viewport performance(both use 100% points display): 12fps / 14fps
Meshing speed(frost): 2min20sec / 1min23sec
mesh viewport performance: 24fps(20198714poly) / 8fps(21057362poly)

So, I don’t know why sequoia meshing is too much slow?? (I disable cull setting in sequoia mesher)
And want to more improve point cloud viewport performance.

-Makoto

2

I was able to reproduce the performance difference in my own benchmark and have logged it for investigation.

I don’t think we can do much about the drawing speed of points on your graphics card. Note that Sequoia is using OpenGL via a game engine serving as the viewport, while 3ds Max is running its own Nitrous graphics via DirectX, but the general bottleneck would be the graphics card’s ability to draw point primitives. So while we are 2 fps slower in your case, we generally approach the problem by reducing the point count to be drawn via LOD and adaptive degradation as we cannot make your card run faster. On desktop machines, we have seen performance increase 10x by switching to a newer GeForce gaming card, so it is very hardware-dependent.

3

So, Geforce mekes better performance than Quadoro on Sequoia?

4

I guess it depends on the generation of the chip, and even on the drivers. When I replaced my older Quadro with a GeForce GTX 660, I got 10x speed up (it draws about 3 billion points per second). GeForce cards advance very fast since nVidia’s main target audience is gamers. And each year they get faster…

Also a FireGL we have in one of the machines is a few times faster than my GeForce.