I actually only intended to make my point once and not have to make any follow up points at all, but you obviously misunderstood so I clarified, I was not being argumentative, I was just saying something you didn't like or particularly want to read properly.

The first place to look for when errors occur is your own code, were your users, treat us professionally, so far you've assumed one of your users can't do basic control tests and then attacked another user for pointing it out, all unnecessary for you to do, and damaging

I would suggest since you admit you don't know what is happening, that you investigate using some simple metrics. Logic would suggest that since he is running on a 9600, ( IIRC you are running in a 9xxx class nvidia gpu no? ) that shaders are less likely to be at fault, but timers would improve this guess a lot.

A modified version of both codebases that printed out time overall would give a good indication of the differences in CPU usage, and if the results are promising you can fine-tune with more timers to figure out where to start looking.

OK, it's the shader- that's what I needed to know. Good news is that it should be easily fixable.

Yay!

I'm sorry I wasn't more descriptive from the start, but I don't have the foggiest idea how to debug something like this. it wasn't till your post before this last one that I understood what you were looking for (otherwise I would have taken those screens the other day). Sorry for being so thick O_o

No, that's my fault, because I wasn't clear what I needed to know. It's a lot of code to look at, so this tells me what to focus on. I will try to make some time soonish and give you something new to test.

I can test it for ATI, I have multiple machines with ATI crap for gpu.

edit: Argh I tried the new code, and I can even notice if it is working. Do i have to enable certain visual settings for spring to notice it?

Silly question, but you did define a normal map for your unit to use (in custom params) didn't you?

I got 75 fps with the two balls demo with ST, with settings: shadows 1, reflect+refract water, 2xAA, 1600x1200, PBO, reflective units, 3d trees, render detail: everything on highest.

hardware: 3800 mhz E6750, 985 mhz AMD 3870 w/10.3.

too low ?
I afraid that this was both just a demo and just 2 "units", I usually get down to 75 fps after at least 10 minutes, sometimes 15 into the game with those settings.

oh noes FUUUUUUUUUUUUUU I need more mhz again

*revives*
Is there a chance for this to become a engine feature?

no need for it to be an engine feature. The capability is there, games jsut need to sue it (protip, Evolution RTS does )