TZeer

Found this one on the net:

+1 to this Would nice to have some variety on the troops you place on the map. That a spec op guy actually looked a little like spec op, and not just like the average soldier.

If you have more units within your visible range, it will affect your cpu usage. See this thread for more info: http://forums.eagle.ru/showthread.php?t=101491

Is the reduced performance all the time in multiplayer? Or just certain occasions? When flying "fast missions" do you have a wingman? During multiplayer, how many other planes is in your close proximity, within viewing distance?

I'm currently at work and unable to do any tests. Wont be home until 6th of March.

Depends. But over Tbilisi with Kuky's high.lua mod I'm getting around ~80FPS Around Novo it can dip down to 50 depending on where I look. I have had a few times in multiplayer missions where we have a few planes ready to take off where I have been down in 30-40. But most of the time I'm at 60FPS stable (vsync on). But those drops are not due to the GPU, it's the CPU and number of objects that is clogging up the system. For more info check out this thread: http://forums.eagle.ru/showthread.php?t=101491 If you are already limited by the VRAM, a second card would not help. Since the amount of VRAM would not increase. For SLI, 2GB+2GB is still only 2GB unfortunately.

Hard for me to say. I can only speak for myself. I'm running single monitor 2560x1440. And I get almost double the FPS with 2 cards. As long as the CPU don't bottleneck me. I'm not sure how much VRAM your setup needs. And even if the VRAM is enough I dont know if the bandwidth on the memory is high enough to prevent it being a bottle neck.

I'm currently running 2x580 in SLI :smilewink: I have yet to see anyone with a working x-fire profile. All I have ever seen is people with dual GPU amd cards where only one GPU has been used, or x-fire users where only one card has been in use. If you you can show me someone with a wokring x-fire setup I would really appreciate it. Would like to keep both options open for my next upgrade.

It looks like a lot of individual trees, but they are really large models which consists of 1-3 objects and textures on them. So they are not really that resource hungry.

Would also be nice if we could get a definitive answer from ED regarding the doubling of objects and triangles when it comes to modelviewer versus in-game. It's not a big deal for models consisting of 2-5 objects etc. But when you get the A-10C, F-15E, AH-64D etc, they all have ~50 objects in them when checking them out in the modelviewer. And when they come into the game 2x of that. 10 of those units together and you are getting close to 1000 objects. My overclocked CPU @4,6GHz is at it's max when I reach 1100-1200 objects.

Hmm, seems a bit strange. Cause the tank model is all modeled and textured in the modelviewer. I can even start it's animations. And the object count never goes above 3. I'm one of the lucky ones with a overclocked CPU. But I can just imagine the reduced performance some may get due to non optimized models. Take around the harbor of Novo as an example. Over 2000 objects. Over 700 of them from the containers alone.

Did another test now. SLI vs non SLI. No difference at all. But I did notice now that the CPU was working quite hard when the number of objects went up. So the more objects, the harder the CPU works. The sweet spot for me was around 1100-1200 objects. When I passed that limit the CPU was at max load, and the GPU started reducing it's GPU usage. The more objects I introduced to the scene the lower the GPU usage went as did the FPS. So it seems low numbers of objects but higher number of triangles will work much better for models.

Did we get a conclusion/answer to why the triangles and object count doubled when the model got put into the mission editor?

Original file gave 68 FPS with 100 models. Your file gave me 120 FPS. But I noticed the object count on those two files are different. The old one had 17 objects. Your new one had 12.

Rgr

I have about an hour for testing right now. What do you want me to test?

Forgot something. What resolution are you playing at? What kinda games? Are you looking at standard versions or specialized version of the card. Like factory overclocked etc. One card might be very good up to a certain point. Then suddenly when you reach 2560x1440 resolutions the roles might be changed.

AMD 7970 GHz edition is even beating the GTX 680 in some tests. And if the AMD is even cheaper then the 670, I would go for the 7970. If you plan to go multiple cards in the future, I would go with nvidia, due to the fact that crossfire don't work with DCS. But SLI does.

It's not just objects doubling, triangulation doubles as well.

Yes, the models that can be destroyed have their own .edm file for the destroyed model. The 6 object tank model, has a moving turret, main gun, machine-gun and belts. That's 4 animations. The 38 object radar have moving belts and radar. 2 animations.

If you had read the entire thread you would have known that I have tested multiple objects. And for the last part about a SAM targeting a plane. It has nothing to do with this test. And even if it was, the AI would just add on top of the already reduced performance. All my testing as far as its possible I'm doing far far far away from any other objects on the map, in a clean mission editor. This is to prevent other factors to mess up the data. Where that's not possible, like Novorossiysk, I do adjustments, and observe the results I get. And just to use that scenario as an example. By removing the cranes, containers and smallboats I improved my FPS by 100%, while reducing the number of objects from over 2000 to under 700. Forgot to give a proper answer to this one, was late last night :) No, the model had no animated parts moving while I tested. This is something I'm wondering about myself. What I suspect, is that weaker systems have the same performance drops, but on a much smaller level. Simply because they have other bottlenecks in their system. Or they are already experiencing other performance issues like the heat blur bug or FPS loss due to internal flood light in cockpit. With the heat blur bug, my FPS would go down to around 50-60 FPS stable. Same with the flood lights in the cockpit. But when I combined them both, I still had between 50-60 FPS?? Since I have been able to remove that factor and get a very good performance, my system is very sensitive to changes. 20% change for someone running 30 FPS, is only 6 FPS, If I'm running 140-200 in a test environment that 20% is 28-40 FPS difference, and quite easy to pick up. 6 FPS is very easy to to just brush off as "normal" or probably something like "oh, your looking in that direction" Now, it's not much we can do except highlighting the models with abnormal amount of objects in them. It's up to ED to take action on it. Hopefully they will :)

Reduced shadows to low. No difference. Still ~30-40 FPS with the high object model.

All tests I have done, shows that when I'm reducing the number of objects in different settings, my performance increase. There are models with moving parts that consists of 1 objects in the modelviewer. Also, the rail car I showed earlier with 93 (186) objects is a static model in the world. And AI code has nothing to do with the pure visual. Based on what? All test's I have done points toward that the system is getting clogged by the number of objects. Increase number of objects and GPU usage goes down. Decrease number of objects, and the system get's breathing room to utilize the GPU and CPU properly. Number of triangulations or whatever with the high object model, was just under 400000. With the low object model, it was almost 2 mil. And still pushing 110 FPS more.

Here you can see the difference between optimized models and models who has not gotten the same treatment. In all pictures I have 100 units. First one is the M1A1. Roughly 140 FPS. Second set of pictures is the 1c91. 7980 objects!!! and 33 FPS. What was that SkateZilla?

I don't know why, but when you put the model into the game, the object count on them doubles. 2 examples: M1A1 In modelviewer, 3 objects. In game, 6 objects. Still not that bad. Now, a badly optimized model. The 1c91. 38 objects. In game. 76 Objects!!! That means the single rail car earlier actually gives you 186 objects. While the little crane will give you 80.