The only reason i stated what i did about comparing the games is that game engines can be so far different and what actually happens in game can also be different even if the game is of a similar style .. in example 2 rts games. It would be like comparing the old command and conquer form the 90's with the original TA from the 90's TA ran like crap compared to C&C on my old 486 but i understand why. the games use different rendering tech's. C&C was all sprite, where as TA used 3d rendered objects on a 2d map. Understanding the underlying tech behind a game gives insight into how it performs on some systems. If you want to know more about why PA runs like so on your system im sure SXX and i can help you find the reason if we do test's. That's if you want to assist us. also i dont think uber can really tone down the poly count any more than they have.. i have viewed a lot of their models and most models are quite low in poly count. The issue with performance is the way the ATI card is handling the openGL shader coding and probably a lot of other internal coding issues that are being ironed out. Biggest one is all the units are individually rendered, this puts mad strain on the CPU and GPU, when the coding is done they will be rendered using instancing and all drawn in batches instead of individual calls per unit.
First we are not developers. All devs have red nicknames. Still there is known issues with performance on AMD graphics cards that developers said about already. Some players just affected more than others. E.g for many people using of non-Steam version works just fine, but unfortunately your problem is bigger.
Any chance you have an nvidia card sitting around? Ever since they merged with AMD the competitiveness of an intel/nvidia based system vs an AMD/ATI system has been more one sided in I/Ns favor then a then a fox vs a hen. Completely non helpful post at this time, but when building your next computer, remember there's a reason some of those components are so damn cheap compared to there competitors. Gigabyte: you get what you pay for. (I also file asus under this descriptor except that is because if you have to deal with their support or have to RMA something. Complete pita)
Seriously, it has been stated many times that true optimization stuff started a few weeks ago. This including most basic optimization technics which are not in place like batching units rendering. Do not expect to see major improvement before a few weeks. They have lots of stuff to do especially because of issues with AMD/ATI opengl implementation. Unless you get something like 1 or 2 FPS, i would simply patiently wait. It will come, Devs are targetting 60FPS
As Doud says - there's still a lot of optimization to do. I just finished a major pass on rewriting some of the low-level code, that will enable some big improvements to performance. I'm also in the process of re-writing how we do instancing for features (trees, etc), because the current implementation really, really hurts on AMD cards. Since I have an AMD card at home, this makes me very sad, so I want to fix it. I know it sucks that the perf is not where it should be... but we're all working very hard on making the improvements we need to make to get there.
