Sigh... alright, let's run the figures one more time. One LAST time, because I refuse to utterly destroy your pathetic attempt at an argument again, it's just a waste of valuable time I might spend working on my mods.
With shadows off, an average computer such as my own gets a bare minimum of 150 frames a second, and as high as 300 if there's nothing at all going on and it's a simple map. All tweaks turned on, all bars full.
With shadows on, I get a framerate between 50 and 30, depending on what, if anything, I'm looking at. This is with all graphical tweaks turned on, all bars to full again.
With TWO HUNDRED of my fighters onscreen, that's 1000 triangles per fighter and 512x512 textures, I am reduced to 15-25 FPS, although generally in the higher range of that figure.
That is a total of at least 200,000 triangles in action. Now, for comic relief, check what XTA gets with 200 units onscreen. *GASP* it's the same thing! QUITE LITERALLY!
Now remember, these aren't even evolva high-detail models. These are stock XTA bulldogs which weigh in at a massive 56 polygons. I haven't modified the results in any way, in fact I puposefully put GEM at a disadvantage - XTA is rendering the same textures for every unit, GEM is rendering 4 seperate 512x512 textures for 4 different types of fighters. Oh, plus the GEM ones have things like real reflectivity.
So where's the bottleneck, my models or the engine? Put on your thinking caps here. The game performs identically with 11200 polygons worth of bulldogs onscreen as well as with 200,000+ triangles onscreen. Hmm, hmm...
Since clearly there is no difference whatsoever between XTA and GEM with shadows turned on, let's try it with shadows turned off.
Aha! Here XTA gets ahead. There is a difference of 5 entire frames per second between 11200 polygons and 200,000 triangles But, let's compare it to a similar screenshot, this one showing no units at all:
So, between nothing onscreen and 100 XTA tanks onscreen, there is a difference of 62 frames per second. With 100 GEM fighters onscreen, there is a difference of 67 frames per second. So going from 0 polygons onscreen to 11200 is a difference of 62 FPS. Between 0 and 200,000+ there is a difference of 67. Again, the bottleneck appears to be the engine, since if only 11,000 polygons can reduce it 62 frames per second, there's obviously a problem there. 11,000 polygons is nothing in modern terms.
The bottleneck remains: the engine. And as that bottleneck is improved, the difference between XTA and GEM performance will likely widen a bit, yes, but GEM will remain more than playable and will likely gain nearly as many FPS as XTA would, especially as by the time the engine is improved to that extent, computer hardware will have similarly improved massively.
