Jump to content
  • Announcements

    • Battlefront.com

      Special Upgrade 4 Tech Tips   12/27/2016

      Hi all! Now that Upgrade 4 is out and about in large quantities we have now discovered a few SNAFUs that happen out in the scary, real world that is home computing.  Fortunately the rate of problems is extremely small and so far most are easily worked around.  We've identified a few issues that have similar causes which we have clear instructions for work arounds here they are: 1.  CMRT Windows customers need to re-license their original key.  This is a result of improvements to the licensing system which CMBN, CMBS, and CMFB are already using.  To do this launch CMRT with the Upgrade and the first time enter your Engine 4 key.  Exit and then use the "Activate New Products" shortcut in your CMRT folder, then enter your Engine 3 license key.  That should do the trick. 2.  CMRT and CMBN MacOS customers have a similar situation as #2, however the "Activate New Products" is inside the Documents folder in their respective CM folders.  For CMBN you have to go through the process described above for each of your license keys.  There is no special order to follow. 3.  For CMBS and CMFB customers, you need to use the Activate New Products shortcut and enter your Upgrade 4 key.  If you launch the game and see a screen that says "LICENSE FAILURE: Base Game 4.0 is required." that is an indication you haven't yet gone through that procedure.  Provided you had a properly functioning copy before installing the Upgrade, that should be all you need to do.  If in the future you have to install from scratch on a new system you'll need to do the same procedure for both your original license key and your Upgrade 4.0 key. 4.  There's always a weird one and here it is.  A few Windows users are not getting "Activate New Products" shortcuts created during installation.  Apparently anti-virus software is preventing the installer from doing its job.  This might not be a problem right now, but it will prove to be an issue at some point in the future.  The solution is to create your own shortcut using the following steps: Disable your anti-virus software before you do anything. Go to your Desktop, right click on the Desktop itself, select NEW->SHORTCUT, use BROWSE to locate the CM EXE that you are trying to fix. The location is then written out. After it type in a single space and then paste this:

      -showui

      Click NEXT and give your new Shortcut a name (doesn't matter what). Confirm that and you're done. Double click on the new Shortcut and you should be prompted to license whatever it is you need to license. At this time we have not identified any issues that have not been worked around.  Let's hope it stays that way Steve
    • Battlefront.com

      Forum Reorganization   10/12/2017

      We've reorganized our Combat Mission Forums to reflect the fact that most of you are now running Engine 4 and that means you're all using the same basic code.  Because of that, there's no good reason to have the discussion about Combat Mission spread out over 5 separate sets of Forums.  There is now one General Discussion area with Tech Support and Scenario/Mod Tips sub forums.  The Family specific Tech Support Forums have been moved to a new CM2 Archives area and frozen in place. You might also notice we dropped the "x" from distinguishing between the first generation of CM games and the second.  The "x" was reluctantly adopted back in 2005 or so because at the time we had the original three CM games on European store shelves entitled CM1, CM2, and CM3 (CMBO, CMBB, and CMAK).  We didn't want to cause confusion so we added the "x".  Time has moved on and we have to, so the "x" is now gone from our public vocabulary as it has been from our private vocabulary for quite a while already.  Side note, Charles *NEVER* used the "x" so now we're all speaking the same language as him.  Which is important since he is the one programming them
iMolestCats

CM performance thread

Recommended Posts

Layout of what you should have listed (Made by Steiner):

FRAPS RESULTS:

minimum - maximum - average framerate:

HARDWARE

Manufacturer + model:

OS:

CPU:

RAM:

Graphics card (overclocked?):

Graphics memory:

SETTINGS

Screen resolution of the game:

In game model quality:

In game texture quality:

Shadows on/off:

Shaders on/off:

GRAPHICS CARD SETTINGS

Anti-aliasing:

Texture filtering:

VOLUNTARY ADDITIONAL DATA

Share this post


Link to post
Share on other sites

My Specs...

FRAPS RESULTS:

minimum -14 maximum -35 average framerate:

HARDWARE

Manufacturer + model: self built

OS:Windows 7 64 Bit

CPU:Intel 3570k 3.8GHz, Overclocked 4.0

RAM:16GB 4x4

Graphics card (overclocked?):2x7850 Radeon HD

Graphics memory: 2GB

SETTINGS

Screen resolution of the game: 1920x1080

In game model quality: Best

In game texture quality: Best

Shadows on/off: On

Shaders on/off: Off

GRAPHICS CARD SETTINGS

Anti-aliasing: x8

Texture filtering: x8

Share this post


Link to post
Share on other sites

Fraps for what? They depend on what's visible.

You will need to pick a scenario and then we measure fraps on the opening screen before anything moves (including the camera).

Share this post


Link to post
Share on other sites

iMolestCats sadly only has posted the list but no explanation how the test procedure is about to work. :mad:

The idea of the test is to choose two representative 1 minute action replays of one medium and one large scenario with the view locked to a certain unit and make these files available for download.

Don't be confused by the test results he posted. They mean nothing and i have no clue what he measured. :D

First we need representative save-files showing 1 minute of action (CMBN v2.01).

Share this post


Link to post
Share on other sites
Guest

Things might look better (and possibly even perform better, depending on your card) with shaders on.

Share this post


Link to post
Share on other sites

I'm in the testing fiesta to - don't know what scenario saves games to upload besides having medium and large maps. How many units should be present, terrain diversity, weather type, etc.?

Share this post


Link to post
Share on other sites

The map/action should offer a good diversity of many terrain elements in the view. Maybe a urban only map gives very different results to maps without houses but instead full of trees?

I think the medium map should have at least a dozen houses and quite some trees.

Units: For the medium map i'd suggest a company of moving/fighting infantry per side and one or two platoons of vehicles in action.

The large map/action IMO should not be huge, because of the scarcity of such battles played. What is a result of x frames per second on a huge battle worth, that most players will probably never play anyway?

I think for the large map maybe a full batallion inf per side and 3 or 4 platoons of vehicles should be enough to get an impression of the scaling of the system if the results from the smaller testscenario are known, too (factor 3 more units + much bigger map).

The locked unit view should show big parts of the map, trees and houses and as many units as possible (wide-angle view activated) or during the action the different elements come into view.

It should also be tried from the scenario uploader at which viewing level the biggest impact on the framerate occurs and be chosen accordingly. I guess locked at level 3 or 4 to the unit will be better suited than the gound level 1 view, because of the further horizon the game has to draw and because we use the higher levels most of the time for navigating the map, where a fast and fluid graphic is more important.

I also think that the weather and lightning should offer quite some shadows to draw. Who wants the surprise by choosing a hardware that shows good framerates with a testscenario that runs at noon or overcast, but with long morning shadows the system maybe sucks? :D

Share this post


Link to post
Share on other sites

I'm in.

We should probably also add video card driver version.

Also, no apps other than fraps running.

I'll start hunting around for a scenario. What level should the lock view be?

Share this post


Link to post
Share on other sites

Is it possible to make this the first post or a new sticky thread?

Performance test for CMBN

Reproducable measurements are achieved by downloading a certain saved game file with a certain 1 minute action replay, locked to a certain unit at a certain viewing level.

The file is a SAVED GAME (not a PBEM-file!). Therefore you must put the file in the CMBN/GAME FILES/SAVED GAMES folder NOT in the INCOMING PBEM folder. ;)

The performance is measured with the freeware tool FRAPS: www.fraps.com

IMPORTANT!

The unit used for benchmarking is already selected: "Barkmann", PSW 234/1, 1.Zug HQ.

BEFORE BENCHMARKING MAKE SURE TO PRESS TAB UNITL THE VIEW IS LOCKED TO THE UNIT (otherwise the camera would only move with the unit, but not turn with the turret and therefore the benchmark would deliver wrong results).

DO NOT CHANGE THE VIEWING LEVEL (=3).

Download of the performance testfile (used scenario: "Closing The Pocket"):

I have submitted the file to the repository. Until it will be available, I uploaded it to my Dropbox folder:

https://dl.dropbox.com/u/96409796/CMBN%20v2.01%20Performance%20Test%20Medium%20Size.zip

FRAPS RESULTS

minimum - maximum - average framerate:

SETTINGS

Screen resolution of the game:

In game model quality:

In game texture quality:

Shadows on/off:

Shaders on/off:

Trees (full/distant):

GRAPHICS CARD SETTINGS

Anti-aliasing:

Texture filtering:

HARDWARE

Manufacturer + model:

OS:

CPU:

Mainboard:

RAM:

Graphics card (overclocked?):

Graphics memory:

VOLUNTARY ADDITIONAL DATA

Share this post


Link to post
Share on other sites

FRAPS RESULTS

13/25/19,2

SETTINGS

Screen resolution of the game:1920x1200

In game model quality: BEST

In game texture quality: BEST

Shadows on/off: ON

Shaders on/off: ON

Trees (full/distant): FULL

GRAPHICS CARD SETTINGS

Anti-aliasing: 32x

Texture filtering: 8x

HARDWARE

Manufacturer + model: DIY

OS: Win XP

CPU: Intel Q6600 (o/c @2,6GHz

Mainboard: Gigabyte P35DS3

RAM: 4 GB (3 GB used in WinXP)

Graphics card: GTX560

Graphics memory: 1 GB

VOLUNTARY ADDITIONAL DATA

Share this post


Link to post
Share on other sites

FRAPS RESULTS

19/37/26,8

SETTINGS (usually used)

Screen resolution of the game:1920x1200

In game model quality: BEST

In game texture quality: BEST

Shadows on/off: OFF

Shaders on/off: ON

Trees (full/distant): distant

Share this post


Link to post
Share on other sites

FRAPS RESULTS

16/21/19.417

SETTINGS

Screen resolution of the game: 1280x1024

Vertical Synchronisation on/off: off

In game model quality: improved

In game texture quality: improved

Shadows on/off: on

Shaders on/off: on

Trees (full/distant): full

GRAPHICS CARD SETTINGS

Anti-aliasing: 4x

Texture filtering: 16x

HARDWARE

Manufacturer + model:

OS: Windows 7 64bit

CPU: AMD FX-6300 BOX 3,5 GHz - 14MB Cache - 95W

Mainboard: Asus AM3+ M5A97 (970 ATX)

RAM: DDR3 1600 8GB CL8 Corsair 2x4GB Vengeance

Graphics card: Geforce 550 Ti (MSI)

Graphics memory: 1Gb

Hard Disc (SSD): Samsung 840 Pro 128GB

All at vanilla settings, nothing over-clocked.

@Steiner: You forgot to add in the instructions that the Z-folder should be renamed (from Z to A is suffecient - no need to delete it) before testing, otherwise mods could impact the measurements. Also it would be wise to add type of Hard Disc: HDD/SSD. What is also missing is if the game settings have Vertical Synch on or off - makes quite a difference so it is an obligatory info!

Edit1:

FRAPS RESULTS

17/35/30.733

Shadows on/off: off

*all other settings as in the first test

Note: FPS increase exactly by a third as I already noted before.

Edit2:

FRAPS RESULTS

18/22/20.283

In game texture quality: balanced

*all other settings as in the first test

Note: looks like changing it to balanced texture quality gives me 1 more FPS and 2 more minimum FPS while I haven't noticed any visual difference so I'm sticking with it for now. ;)

Consistency issue: Barkmans AFV got hit in one test so he backed off with it changing the FPS result considerably. Make a note that if he gets hit test needs to be remade and is not to be entered here!

Edit3:

FRAPS RESULTS

16/19/17.667

In game texture quality: balanced

Anti-aliasing: 8x

*all other settings as in the first test

Note: I lose 3 FPS going from 4x to 8x antialiasing - guess I'll have to stick with 4x.

Edit4:

FRAPS RESULTS

21/26/23.267

In game texture quality: balanced

in game model quality: balanced

*all other settings as in the first test

Note: It gives me 3 more average FPS. Don't really know if the visual difference is big or not. Doesn't seem so so I'll probably stick with it from now on.

Edit5:

FRAPS RESULTS

21/25/23.333

In game texture quality: balanced

in game model quality: balanced

Texture filtering: 8x

*all other settings as in the first test

Note: A bit less crisp visual difference and no FPS gain as expected (modern GPU's have no problems processing 16x texture filtering) so I'm sticking to 16x.

Edit6:

FRAPS RESULTS

23/28/26.200

In game texture quality: balanced

in game model quality: balanced

Anti-aliasing: 32xCSAA!!!

*all other settings as in the first test

Note: This is very strange and inconsistent with logic! Why would having such a high antialiasing produce 3 more FPS versus 4x antialiasing!?

Edit7:

FRAPS RESULTS

16/27/19.567

In game texture quality: balanced

in game model quality: balanced

Anti-aliasing: 8xCSAA

Note: This result is OK and logical. FPS scales down when compared to Edit4 test where all settings were the same with the exception of antialiasing which was set to 4x.

Edit8:

FRAPS RESULTS

25/30/27.700

In game texture quality: balanced

in game model quality: balanced

Anti-aliasing: 16xCSAA!!!

Note: This is again very strange and inconsistent with logic! Does the game not support 16x antialiasing and up? Getting 4 more avg FPS then with 4x antialiasing is just wrong. I can't really say if there is any visual difference between 4x and 16x antialiasing though - doesn't seem so. Looks like I'll be sticking to 16xCSAA antialiasing for the time being - what a revelation!

Definitely something that needs to be looked into by the developers!!!

Share this post


Link to post
Share on other sites
Guest

Good test. Right now you're likely looking at differences in driver optimization. We don't directly affect or support any of this (nor do we need to), it's down to how your drivers interact with OpenGL. It looks like higher-level CSAA is probably the most heavily optimized path for your card.

You didn't mention whether you had AA turned on in-game, or just outside?

Edit: the reason you're not seeing dramatic visual differences between 4xMSAA and 16xCSAA is that 16xCSAA is equivalent to 8xMSAA with coverage samples added in, and the difference between 4xMSAA and 8xMSAA for a given frame isn't really all that huge, necessarily. Basically, once you're above 4xMSAA (or 8xCSAA) you're really not seeing a huge visual benefit. There's a point at which more samples may not matter all that much. If your drivers can handle it (as apparently they can) you may as well turn it on, though.

Share this post


Link to post
Share on other sites

FRAPS RESULTS:

minimum - maximum - average framerate:

Shadows/Shaders: ON: 18 / 34 / 24.268

Shadows OFF/Shaders ON: 24 / 45 / 32.916

Shadows ON/Shaders OFF: 19 / 34 / 25.387

Shadows/Shaders: OFF: 24 / 46 / 34.129

HARDWARE

OS: Win7

CPU: i7-2600K CPU @ 3.4GHz

Mainboard: ASUS Sabertooth P67

RAM: 8GB DDR3

Graphics card (overclocked?): GeForce GTX 670 - No

Graphics memory: 2GB

Hard Disk: Western Digital Caviar (7200 rpm)

SETTINGS

Screen resolution of the game: 1920x1080

In game model quality: Best

In game texture quality: Best

Trees (full/distant): Full

GRAPHICS CARD SETTINGS

Anti-aliasing: 32x CSAA (on in-game)

Texture filtering: High Quality

No more tests, cuz this game looks funny without all the mods :eek::D

Share this post


Link to post
Share on other sites

Consistency issue: Barkmans AFV got hit in one test so he backed off with it changing the FPS result considerably. Make a note that if he gets hit test needs to be remade and is not to be entered here!

If you just run the replay, how does Barkman get hit? ;)

Edit: Just for fun, I ran the test again with all the mods, just to see what kind of hit I'm taking. My game is heavily modded with all of Aris' stuff, including his hi-res terrain mods. Here's what I get with shadows and shaders on: 17 / 32 / 22.935. Not too much of a hit... roughly 1 frame, not nearly enough to stop using the mods

Share this post


Link to post
Share on other sites
Good test. Right now you're likely looking at differences in driver optimization. We don't directly affect or support any of this (nor do we need to), it's down to how your drivers interact with OpenGL. It looks like higher-level CSAA is probably the most heavily optimized path for your card.

You didn't mention whether you had AA turned on in-game, or just outside?

Edit: the reason you're not seeing dramatic visual differences between 4xMSAA and 16xCSAA is that 16xCSAA is equivalent to 8xMSAA with coverage samples added in, and the difference between 4xMSAA and 8xMSAA for a given frame isn't really all that huge, necessarily. Basically, once you're above 4xMSAA (or 8xCSAA) you're really not seeing a huge visual benefit. There's a point at which more samples may not matter all that much. If your drivers can handle it (as apparently they can) you may as well turn it on, though.

Very interesting! Thank you for taking a look into this.

I have AA turned on in the game.

If I turn AA in game and leave Nvidia control panel to use "application controlled" then I get no AA whatsoever. When I set how much AA I want in Nvidia control panel then AA works.

If you just run the replay, how does Barkman get hit?

Every time I play the reply something different happens. In my testing Barkmans vehicle got hit 3 times. Once they drove back, once they all died and once they all bailed. Am I doing something wrong? When the test mission loads I click on the Barkan's AFV, press the "end turn" button, hit TAB twice and then press "end turn" button again. Preloaded mission starts.

Edit: Great news about the mods!

Share this post


Link to post
Share on other sites
Every time I play the reply something different happens. In my testing Barkmans vehicle got hit 3 times. Once they drove back, once they all died and once they all bailed. Am I doing something wrong?

When they die, yes. :D;)

You select the correct car and hit TAB - that's good...

When the test mission loads I click on the Barkan's AFV, press the "end turn" button,

You are in the REPLAY phase when the file is loaded. Why not simply pressing the PLAY button?

Share this post


Link to post
Share on other sites

FRAPS RESULTS

13/26/18

SETTINGS

Screen resolution of the game:1920x1200

In game model quality: BEST

In game texture quality: BEST

Shadows on/off: ON

Shaders on/off: ON

Trees (full/distant): FULL

HARDWARE (difference)

CPU: Intel Q6600 (o/c @2,925GHz)

FRAPS RESULTS

20/40/28,6

SETTINGS

Screen resolution of the game:1920x1200

In game model quality: BEST

In game texture quality: BEST

Shadows on/off: OFF

Shaders on/off: ON

Trees (full/distant): DISTANT

Share this post


Link to post
Share on other sites

My test results combined into one post:

FRAPS RESULTS

13/25/19,2

SETTINGS

Screen resolution of the game:1920x1200

In game model quality: BEST

In game texture quality: BEST

Shadows on/off: ON

Shaders on/off: ON

Trees (full/distant): FULL

GRAPHICS CARD SETTINGS

Anti-aliasing: 32x

Texture filtering: 8x

HARDWARE

Manufacturer + model: DIY

OS: Win XP

CPU: Intel Q6600 (o/c @2,6GHz)

Mainboard: Gigabyte P35DS3

RAM: 4 GB (3 GB used in WinXP)

Graphics card: Nvidia GTX560

Graphics memory: 1 GB

FRAPS RESULTS

13/26/18

HARDWARE (difference)

CPU: Intel Q6600 (o/c @2,925GHz)

FRAPS RESULTS

19/37/26,8

SETTINGS (usually used settings)

Screen resolution of the game:1920x1200

In game model quality: BEST

In game texture quality: BEST

Shadows on/off: OFF

Shaders on/off: ON

Trees (full/distant): distant

HARDWARE

CPU: Intel Q6600 (o/c @2,6GHz)

FRAPS RESULTS

20/40/28,6

HARDWARE (difference)

CPU: Intel Q6600 (o/c @2,925GHz)

Share this post


Link to post
Share on other sites

×