Something is DEFINATELY fishy...

Maybe because you're using DCSMax. It enables cores after 30 seconds after entering cockpit. You don't need DCSMax since the core bug was fixed with 1.01 patch. Try.

No change:

2010-01-09 15:52:20 - not using DCSMax and not using alt-tab trick:

Frames: 2844 - Time: 93570ms - Avg: 30.394 - Min: 26 - Max 38

It seems that alt-tab trick on my (4/8 core) CPU additionally boosts all cores. :confused:

Disable Hyperthreading in BIOS and test. Many people see increase in perfomance in many games.

Anyway, no need to use DCSMax if it's not for the special launcher (not because cores)

Exactly. My tests show that disabling HT gives the same performance boost in DCS BS as my little alt-tab trick:

HT on: 30 FPS

HT off: 37 FPS

HT on and alt-tab trick: 37 FPS

I use DCSmax only because it auto-starts my TrackIR and I can easilly change skins.

Thanks.

It just amazes me that the graphics are decent and the GPUs aren't taking that big of a hit. Do the Russians understand DirectX that much better than American developers? heheh... or is the app just not using DirectX properly? My machine eats games like Crysis and Dragon Age for breakfast, nearing 100fps, and those apps are super graphics intensive. Things that make you go hmmmm.....

I just want to make a quick point about graphics differences in games. Stand back and think about the differences between Crysis, Dragons Age, and DCS.

When you play FPS, RPG, MMOs. Typically there is more lighting, fog/smoke right up close to the camera view and generally, much more going on in a typical scene: more models up close, higher poly count models, more lighting calculations, more light sources, etc. You are typically much closer up to the other characters, objects, etc, all of these require more vertices, higher resolution textures, and all-in-all, much more work for the GPU. The GPU is pretty good at this stuff, so these games can typically fly pretty fast, and, this is why games like Crysis are usually used for benchmarking GPUs for hardware sites.

In DCS, typically you are viewing the world from inside your cockpit, so right away, much of the world is blocked from view, and thus doesn't need to be rendered (I am assuming that their engine culls objects blocked by the cockpit). When you view objects, they are typically very far away and can be represented adequately with quite low-polygon count meshes and textures (the so called level-of-detail, or LOD). Additionally, the ground can be adequately modelled using low-polygon count meshes and relatively coarse textures, same with buildings. It all adds up that in DCS, the GPU has much less to do than in a typical FPS or MMO game.

Now, versus a typical FPS or MMO game, there is much more going on in terms of physics. Of course there are the basic things that must be done in all games: input, collision detection and AI. DCS also has ballistics, weapons (can be a lot of these going on at once) and most importantly, the flight model (again, can be several objects flying at once), which involves solving a time-varying differential equation accurately and in real-time. This is probably the most CPU intensive step, and is probably the FPS-limiting step. Even though, AI planes probably use a simplified and easier-to-compute FM.

Is there a way to see a breakdown for each frame how much of the time is taken in the flight model, AI, collision detection and graphics rendering to get an idea of just how much computation the flight model takes up? I wonder if you can make an empty mission with the helo turned off to see if there then isn't any flight modelling to get a sense of the max FPS that the GPU can turn out with their graphics engine?

Anyway, just wanted to make a few comments as a game programmer, or at least, an amateur-hack of a game programmer, as this comment seems to come up quite a bit. My observations and comments may be a bit wrong though, so welcome corrections.

Stingray

Exactly. My tests show that disabling HT gives the same performance boost in DCS BS as my little alt-tab trick:

HT on: 30 FPS

HT off: 37 FPS

HT on and alt-tab trick: 37 FPS

 

I use DCSmax only because it auto-starts my TrackIR and I can easilly change skins.

Thanks.

What is the alt-tab trick?

What is the alt-tab trick?

As posted before: Alt-Tab out of the game and back in.

Thanks for the positive comments everyone. I was afraid flightsims and the exceptional communities that follow them were dead. I'm really glad I found ED and DCS.

Goldfinger35:

I don't know why, but I get 23% increase in FPS if I do the following:

When mission loads (cockpit view and game is paused), I press alt-tab to exit the game and than alt-tab again to return to the game.

I found some interesting information about this, Goldfinger35:

http://en.wikipedia.org/wiki/Microsoft_Direct3D

Fullscreen mode:

The Direct3D application generates all of the graphical output for a display device. In this mode Direct3D automatically captures Alt-Tab and sets/restores screen resolution

and pixel format without the programmer intervention. This also provides plenty of problems for debugging due to the 'Exclusive Cooperative Mode'.

http://www.codeproject.com/KB/directx/mdx_tutorial1.aspx?msg=387173

The RestoreSurfaces method:

The Restore Surfaces method is called when the user defocus the application and then refocus the application (like an Alt-Tab switch)

So what the above says is that even if the dev doesn't use the RestoreSurfaces method to refresh the textures after alt+tab, d3d provides it inherently. I assume the textures that aren't outside the cockpit window are loading gradually or something like that, which causes a decrease in FPS for the first few minutes of flying, whereas alt+tabbing causes d3d to push all the off screen textures at once... in maybe a more efficient way?.. I won't pretend to understand.. but.. I performed the same tests as previously mentioned, first without alt+tab, then with alt+tab. I received the same increase in FPS that you mentioned. To test the information I found above, I used the same mission and circumstances, but first I used alt+tab method then restarted the game and zoomed out from the F11 airfield and spun around two or three times to take in all the textures. I received the same FPS increase after F11 zooming as I did with an alt+tab. I might use the zoom method myself, as alt+tab in full screen games has sometimes caused me problems in the past. Probably due to the mentioned 'Exclusive Cooperative Mode'.

Stingray:

Is there a way to see a breakdown for each frame how much of the time is taken in the flight model

I'd like to see a wireframe mode.. if not for testing, then to make DCS look like the old Battle Zone game! heheh. I found a useful function where you can hit CTRL+PauseBreak to show your FPS, but if you hit it again it shows a breakdown of resources as well. Very cool. You can tell you're dealing with devs who care about their audience when....:thumbup:

Thanks again,

StrongHarm

I won't pretend to understand.. but.. I performed the same tests as previously mentioned, first without alt+tab, then with alt+tab. I received the same increase in FPS that you mentioned. To test the information I found above, I used the same mission and circumstances, but first I used alt+tab method then restarted the game and zoomed out from the F11 airfield and spun around two or three times to take in all the textures. I received the same FPS increase after F11 zooming as I did with an alt+tab. I might use the zoom method myself, as alt+tab in full screen games has sometimes caused me problems in the past. Probably due to the mentioned 'Exclusive Cooperative Mode'.

As part of my usual start up routine, from the cockpit, I will zoom to the full extent, and look around completely, and then do the same thing from the external view, with the hope that it loads all the textures/models at the finest LOD. This was a trick I heard about in the IL-2 1946 community, probably common in a lot of games. Also, before online combat, many IL2-1946 pilots will fire their weapons once, so that the sounds are loaded and in the sound cards hardware buffer to avoid any stutter when they first fire. Although, it is difficult to do this in the ka-50, as there is much less of each type of ammunition :)

So what the above says is that even if the dev doesn't use the RestoreSurfaces method to refresh the textures after alt+tab, d3d provides it inherently. I assume the textures that aren't outside the cockpit window are loading gradually or something like that, which causes a decrease in FPS for the first few minutes of flying, whereas alt+tabbing causes d3d to push all the off screen textures at once... in maybe a more efficient way?.. I won't pretend to understand.. but.. I performed the same tests as previously mentioned, first without alt+tab, then with alt+tab. I received the same increase in FPS that you mentioned. To test the information I found above, I used the same mission and circumstances, but first I used alt+tab method then restarted the game and zoomed out from the F11 airfield and spun around two or three times to take in all the textures. I received the same FPS increase after F11 zooming as I did with an alt+tab. I might use the zoom method myself, as alt+tab in full screen games has sometimes caused me problems in the past. Probably due to the mentioned 'Exclusive Cooperative Mode'.

 

Informative post.

Note that my testing method was different: I have creted a replay track where I fly 15 minutes above ground units and city. Note that I engage FRAPS counter after 10 minutes of flight and turn it off 5 minutes later, so I think I have avoided pre-loading of textures anyway.