I have added a new experimental render mode to ModelConverterX. This mode gives the object a colour based on the amount of texture vertices at a certain location. So this mode should help you to see where your object can be optimized further to increase performance. I have made a quick video to demonstrate it, since this is tricky to explain in words:
When analyzing the result keep in mind that this is calculated per modelpart. So where two different parts, with different material, meet, you will not see a higher texture vertex count. That is by design. As I already mentioned in the video, this is an experimental mode. I hope it helps to understand performance better, but please let me know if you have suggestions on how to make it more useful.