should default option for object optimization mode in object viewer is low or high instead none ?

hello

yesterday i tested my addon with object viewer that contain 100.000 vertex and some 1K texture without using alpha channel, when i test using none and low mode the fps is around 8-13 fps with around 55000 opaque faces detected but when i use high mode the fps jump to 100-120 fps, that a lot performance boost isnt it? so i wonder if should the default option is high or low not none.

see image below for comparison
*none and low mode is make same fps and faces detected result

 

but why on high mode opaque faces detected is so low ? although the total opaque faces not effecting the mesh visually.

I'd suspect (without seeing the model) that high object optimization has managed to combine most faces & junk a bunch of duplicate vertices.

If opaque & using the same material / properties, faces can be combined into a single contiguous draw call. This however takes time, and in Object Viewer may not be the best idea if the object is being repeatedly reloaded, hence why it's set by default to off.

Other thoughts-
GL3 is likely to be an order of magnitude faster with this many vertices.
I would also strongly suspect you don't require 1k textures. Look at the size your texture ends up when rendered to screen. If you're worried about blurriness when downsizing, just use a decent resample algorithm (after all that's all GL is doing internally)