For small models (about 100 objects), the load of a CPU core remains under 100 percent and the graphics card is about 50 percent capacity.
The model can be rotated smoothly and you can navigate very well.
For larger models (assembly with 142020 subordinate objects); a CPU is fully utilized and the use of the graphics card falls to a maximum of 10 percent.
The model can not be rotated properly (delay of several seconds, strong jerking).
It seems that modeling doesn’t provide the graphic card with the necessary data fast enough for large models.
That With the latest hardware, CED can not retrieve the available power / capability from the new system.
This is of course a pity or an absurdity.
Here, there is an urgent need for action that the graphics engine of modeling is adapted to the modern HW / operating systems.
