CM Performance

[Childress]

After a lengthy hiatus with CM I fired up BN and loaded a scenario (reinstalling with the multiple patches was hellish). In the meantime I had replaced the medium power laptop with a rather powerful one: I7, 16g ram, a high end, if not state of the art, graphics card (NV GTX960M). I notice that load times and FPS have improved but not dramatically as expected.

Still cannot enable maximum parameters without a noticeable drop in performance. Not much of a gamer so no comparisons with other games. Eventually an SSD drive will get installed.

What's the general experience?

[AttorneyAtWar]

CM only uses one core so your experience is pretty much par for the course, if there are a lot of trees and other clutter on the map that also slows the game down. I found my FPS stays around 20-30 with a 2550K and GTX 770, the game is really only using your CPU, the graphics card isn't working very hard.

 

Edit: The SSD should help with the loading times, that's about it.

Edited March 3, 2016 by Raptorx7

[Bulletpoint]

After a lengthy hiatus with CM I fired up BN and loaded a scenario (reinstalling with the multiple patches was hellish). In the meantime I had replaced the medium power laptop with a rather powerful one: I7, 16g ram, a high end, if not state of the art, graphics card (NV GTX960M). I notice that load times and FPS have improved but not dramatically as expected.

Still cannot enable maximum parameters without a noticeable drop in performance. Not much of a gamer so no comparisons with other games. Eventually an SSD drive will get installed.

What's the general experience?

From what I've read on this forum, nobody gets good performance in CM on high detail levels, even on massively powerful computers. I have a laptop with a GPU much stronger than yours - at least in theory. But I get unsatisfactory performance as well, and not even on maxed out detail levels. So it's likely to be a CPU problem.

I know they keep saying performance is limited because of "massive amounts of calculations" being done, but really.. those calculations are made in a couple of seconds when you click end turn. There's no reason for the game to do the fire, movement and spotting calculations while the game is in the orders mode, or when paused. Yet the game still lags a lot.

Though I very much enjoy the game for what it is, I'm convinced that it suffers from highly unoptimal coding, like a fast car stuck in the first gear.

Edited March 3, 2016 by Bulletpoint

[General Jack Ripper]

What's the general experience?

If you're okay with it, disable shadows. You'll pick up a few FPS. 

[A Canadian Cat]

There was a recent discussion thread that was really great.  I could never figure out why some people were so unhappy with performance but I was very happy with it.  Turns out there are a lot of things to consider.  It was an interesting discussion but I cannot find it any more. Damn it I would have sworn I book marked it but it seems I have not and I did a couple of searches and came up empty.  Sigh.  If any one can find it please link it here - then I will be sure to book mark it.

[Bulletpoint]

I could never figure out why some people were so unhappy with performance but I was very happy with it.  

Depends on personality and expectations as well. CM runs sluggish but playable on low end hardware, and slightly less sluggish and a bit more playable on high end hardware. When I played on a weak laptop, I was happy with the performance compared to my hardware. Now, not so much, even though my game performance did increase somewhat after upgrading.

Edited March 3, 2016 by Bulletpoint

[MikeyD]

CMBN will always labor under the burden of those darned close-packed Normandy hedgerows. I suspect its not so much about the graphics as the insane LOS calculations. When I'm running CMRT or CMBS (no bocage mazes) big maps cause me no problems at all. CMFB Beta is often hilly and densely wooded yet runs like a charm. And I'm playing on a rig that would never be called optimized for gameplay.

Complaining about CM performance while maxing out all your graphics parameters is like complaining about your car's performance when redlining the engine RPMs. The simplest advice is don't do that! Can you live without watching the grass and trees gently waving in the breeze? Turn it off and watch your performance spike.

[Blazing 88's]

There was a recent discussion thread that was really great.  I could never figure out why some people were so unhappy with performance but I was very happy with it.  Turns out there are a lot of things to consider.  It was an interesting discussion but I cannot find it any more. Damn it I would have sworn I book marked it but it seems I have not and I did a couple of searches and came up empty.  Sigh.  If any one can find it please link it here - then I will be sure to book mark it.

Might it be this link?  http://community.battlefront.com/topic/96330-some-performance-and-quality-tips-for-nvidia-users/?page=5

[AttorneyAtWar]

If someone has spent a lot of money on a gaming rig they expect a game like Combat Mission which looks very dated compared to modern games to run well. Of course there is a lot going on under the hood, but it is surprising how far FPS falls in certain cases. I agree though CMBN really does seem to be the primary culprit with performance problems, or any CM map with DENSE forests.

Edited March 3, 2016 by Raptorx7

[Childress]

Depends on personality and expectations as well. CM runs sluggish but playable on low end hardware, and slightly less sluggish and a bit more playable on high end hardware. When I played on a weak laptop, I was happy with the performance compared to my hardware. Now, not so much, even though my game performance did increase somewhat after upgrading.

I estimate that my new PC is at least 8x more powerful than the old one which now sits around as a backup. Performance has improved by 10% at best. Surprising to say the least. It is what it is. But if it's just a matter of optimization BF would have done that long ago. They may be locked into an aging architecture.

[Childress]

There was a recent discussion thread that was really great.  I could never figure out why some people were so unhappy with performance but I was very happy with it.  

I was very happy with the  performance on the old, very modestly powered machine. Even impressed. But, it's true, I expected more after trading in the Clydesdale for Secretariat.

Still a great, engrossing game, however.

[Bulletpoint]

I estimate that my new PC is at least 8x more powerful than the old one which now sits around as a backup. Performance has improved by 10% at best. Surprising to say the least. It is what it is. But if it's just a matter of optimization BF would have done that long ago. They may be locked into an aging architecture.

They are a small company, with very few (maybe just one?) programmer. I'm sure they're really clever people, but it could be they just don't have the employees in-house to optimise the engine properly, and lack the money needed to hire someone to sit down and spend the time needed to track down the performance sinks and eliminate them.

[A Canadian Cat]

Might it be this link?  http://community.battlefront.com/topic/96330-some-performance-and-quality-tips-for-nvidia-users/?page=5

Not that is not the one I was thinking of.  That is a good thread.  The one I was thinking of I asked people what was it they were not happy with any why their desire to get X frame rate was an issue.  It was a very interesting discussion about what there were seeing that I was not noticing.  And the Steve posted a bit of information.  It was not really a practical here is how to optimize things for our machine but it was enlightening.

[c3k]

I am sympathetic to the desire for better performance in-game with higher-end computers. But...I also see the NEED for adequate performance for low-end machines. To me, the oddity is comparing the performance delta with the computer delta. Double the computer does NOT get double the game performance.

But, does it matter?

I say, "No".

I've found my gameplay to be far smoother when I cap framerate to 1/2 my screen refresh. This is NOT a first person shooter or a "twitch" game. Steady fps is more important to me than hitting the highest numbers. (And, while I don't consider myself a hard-core gamer, my wife and my gaming library may disagree. I've built over a dozen rigs for myself, used to use all sorts of 3rd party tweaking/overclocking/hardware mods/etc/etc. I keep 3 running game rigs at home (2 sons: what kind of father would I be otherwise?). About to upgrade my 1090T to an i7-6700k. That'll be nice. I guess admitting it is the first step? Crap. Maybe she's on to something?)

A lot of gaming "feel" is subjective. The only way to compare games from one user to another is frequently done by using a qualitative measurement. That has been framerate, historically. That may not be appropriate to some games...like this one.

I've found that using a high-end computer allows BIG maps. It allows FAST load times. It allows FAST turn times (blue bar). 

For example, the computer I mentioned above, running an AMD Phenom X6 1090T cpu (I've got it at 3.5GHz...I think. Shrug.) with 8GB of ram and an R9 390 (8GB) gpu (note the gpu is way "oversized" for the cpu/ram/mobo...because the cpu/ram/mobo are about to be replaced...) with an SSD for the OS and game: It is excruciating how long it takes to load a specific pbem series I'm in the midst of. I open the turn, the orange loading bar begins, and I walk away for another cup of whatever the time of day dictates I should imbibe. I can usually come back about the time the turn is ready. Or I beat it. (Computer is upstairs from kitchen. I frequently get my own drink so my intern can continue researching whatever it is that I've assigned.)

Another machine, i7-4790K (stock, 4.5GHz?), 32GB ram, GTX970, can load that same turn almost faster than I can say, "Intern! Fetch me my drink!" The framerates on both are VERY similar. (One may hold steady at 28-30, the other may run 24-28. Or so.)

 

The loading times (orange bar) and turn computation times (blue bar) are totally different. If one takes 5 seconds and the other takes 1 1/2 minutes, does that mean it's only 1:25 faster or 18 times faster? Improving "performance" by 1,700% is significant, isn't it? If I just compared min/max framerates, the two machines would seem to have performance numbers which are far closer than the loading times.

 

I would not try to measure CM by framerate.

Having said all that...I would be ecstatic if BFC found a way to run CM in such a manner that it could use all the cpu cores available. Yes, performance seems like it could be improved, but it depends on what you think "performance" means.

[Seedorf81]

I am very, very, very glad that BF allows us "poor people" to play CM with relative obsolete computers.

Recently I changed my Geforce 220 (talk about obsolete!) into a GeForce 610 on a 2,4 Ghz processor. So still not an up-to-date PC, but the Black Sea demo works fine (except for the fact that I can't get a win over those mechanized Russki's ), which I think is beautiful. No immediate need to buy a 1500 euro new computer in order to keep having fun with a game. With a lot of other games my financial capabilities lag those of the desired new versions PC specs.

This is why, in spite of schwimmwagen, sdkfz gunners, and those horrible waiting times for new releases, I still LOVE BF.

 

 

 

[Buzz]

I am sympathetic to the desire for better performance in-game with higher-end computers. But...I also see the NEED for adequate performance for low-end machines. To me, the oddity is comparing the performance delta with the computer delta. Double the computer does NOT get double the game performance.

But, does it matter?

I say, "No".

I've found my gameplay to be far smoother when I cap framerate to 1/2 my screen refresh. This is NOT a first person shooter or a "twitch" game. Steady fps is more important to me than hitting the highest numbers. (And, while I don't consider myself a hard-core gamer, my wife and my gaming library may disagree. I've built over a dozen rigs for myself, used to use all sorts of 3rd party tweaking/overclocking/hardware mods/etc/etc. I keep 3 running game rigs at home (2 sons: what kind of father would I be otherwise?). About to upgrade my 1090T to an i7-6700k. That'll be nice. I guess admitting it is the first step? Crap. Maybe she's on to something?)

A lot of gaming "feel" is subjective. The only way to compare games from one user to another is frequently done by using a qualitative measurement. That has been framerate, historically. That may not be appropriate to some games...like this one.

I've found that using a high-end computer allows BIG maps. It allows FAST load times. It allows FAST turn times (blue bar). 

For example, the computer I mentioned above, running an AMD Phenom X6 1090T cpu (I've got it at 3.5GHz...I think. Shrug.) with 8GB of ram and an R9 390 (8GB) gpu (note the gpu is way "oversized" for the cpu/ram/mobo...because the cpu/ram/mobo are about to be replaced...) with an SSD for the OS and game: It is excruciating how long it takes to load a specific pbem series I'm in the midst of. I open the turn, the orange loading bar begins, and I walk away for another cup of whatever the time of day dictates I should imbibe. I can usually come back about the time the turn is ready. Or I beat it. (Computer is upstairs from kitchen. I frequently get my own drink so my intern can continue researching whatever it is that I've assigned.)

Another machine, i7-4790K (stock, 4.5GHz?), 32GB ram, GTX970, can load that same turn almost faster than I can say, "Intern! Fetch me my drink!" The framerates on both are VERY similar. (One may hold steady at 28-30, the other may run 24-28. Or so.)

 

The loading times (orange bar) and turn computation times (blue bar) are totally different. If one takes 5 seconds and the other takes 1 1/2 minutes, does that mean it's only 1:25 faster or 18 times faster? Improving "performance" by 1,700% is significant, isn't it? If I just compared min/max framerates, the two machines would seem to have performance numbers which are far closer than the loading times.

 

I would not try to measure CM by framerate.

Having said all that...I would be ecstatic if BFC found a way to run CM in such a manner that it could use all the cpu cores available. Yes, performance seems like it could be improved, but it depends on what you think "performance" means.

Expressed very well c3k. Only question I have.... how fast does "Intern! Fetch me my drink!" take? 

[MOS:96B2P]

Another machine, i7-4790K (stock, 4.5GHz?), 32GB ram, GTX970, can load that same turn almost faster than I can say, "Intern! Fetch me my drink!"

I had a rig built about six months ago with this hardware and an SSD.  I am very happy with it.  I think mine is stock 4 GHz but it can be over clocked to 4.5GHz.  It runs very well at 4GHz so I have not bothered to over clock it but the option is there if I want to in the future.  I came from a laptop with a built in graphics card so to me this new rig is just amazing.    

[Childress]

1. I am sympathetic to the desire for better performance in-game with higher-end computers. But...I also see the NEED for adequate performance for low-end machines. To me, the oddity is comparing the performance delta with the computer delta. Double the computer does NOT get double the game performance.
2. But, does it matter?
3. I say, "No".

Right, frame rates don't matter much in CM. It's not an FPS or a racing sim where that's a critical element. And no laptop, however thoroughbred, can rival in power a beastly, homemade desktop if one accepts the space invasion and the loss of portability (I don't). Question: SSD drives tend not to be very capacious. Is there enough room for the Windows install and the CM games with their significant footprints?

[DesertFox]

Question: SSD drives tend not to be very capacious. Is there enough room for the Windows install and the CM games with their significant footprints?

I have a 500gig SSD and there is plenty of place for Windows 10 and all CMs and additional stuff. CM installs take about 10 gig average each...so for your question...yes go for a SSD. You can get good 500gig SSDs for about 150 EUR, 250gig SSDs for about 80 EUR, not that expensive as they used to be 2 years ago.

[MOS:96B2P]

1. <Snip> And no laptop, however thoroughbred, can rival in power a beastly, homemade desktop if one accepts the space invasion and the loss of portability (I don't). <Snip> 

The portability part is a negative if you go with a desktop.  However what I did was keep my laptop and use it mostly for the internet.  In fact I'm typing on the laptop now.  (I'm busy losing a PBEM on the desktop.)   I still have the CM games on the laptop but of course I just play them on the gaming rig now.       

<Snip> Question: SSD drives tend not to be very capacious. Is there enough room for the Windows install and the CM games with their significant footprints?

Hmmmm, I am not a computer guru.  I did some research to decide what I wanted in the gaming rig but I don't recall any concerns over SSD being small.  @c3k would probably know more about it and give a better answer.  Or delegate the question to his intern.....      

My SSD (Solid State Drive) is a Samsung 850 EVO (One Terabyte)  I have windows 10, all the CM titles except Afghanistan, One Total War game, screenshots, files, a bunch of folders with mods etc .........  No problem so far.  I brought up the storage setting thing on the gaming rig.  It advises 214GB used out of 931GB.  And one of the reasons I got a desktop for gaming is that hardware can more easily be added and switched out.  So if I ever did run out of space I would have the option to add more.         

[Bulletpoint]

It's not even really about the FPS (frames per second) number. It's more about CM feeling choppy and laggish, stuttering even when my FPS are reasonable. I have about 30-35 FPS usually, which should be enough, but it just feels really unsmooth, and I'm not sure why. I guess because the framerate fluctuates wildly? Sometimes I will select a unit and then try to scroll to a place to target, and the game will have a "seizure" for half a second where it locks up and then the camera will end up in some other place than i intended. It doesn't lose me any battles, but it's a nuisance, which is a nice way to say it's a pain in the family friendly neck.

In other news, I did some tests today, and it seems setting adaptive v-sync (half refresh rate) actually boosts FPS, which in theory it shouldn't! 

If I disable that, I go from 35 FPS to around 28. Which is really bizarre, as that setting is only meant to cap FPS to half the refresh rate of the monitor. It is not supposed to boost the FPS. In theory, if your game runs at 28 FPS without adaptive V-sync, it should also run at 28 FPS with v-sync on, because it doesn't hit the ceiling of 30FPS (if you're using a 60 Hz monitor). I'm completely confused by this.

The game definitely plays smoother with adaptive V-sync - half refresh rate on. But it's still not what I would call decent performance when I compare graphics to hardware.

Edited March 4, 2016 by Bulletpoint

[c3k]

It's not even really about the FPS (frames per second) number. It's more about CM feeling choppy and laggish, stuttering even when my FPS are reasonable. I have about 30-35 FPS usually, which should be enough, but it just feels really unsmooth, and I'm not sure why. I guess because the framerate fluctuates wildly? Sometimes I will select a unit and then try to scroll to a place to target, and the game will have a "seizure" for half a second where it locks up and then the camera will end up in some other place than i intended. It doesn't lose me any battles, but it's a nuisance, which is a nice way to say it's a pain in the family friendly neck.

In other news, I did some tests today, and it seems setting adaptive v-sync (half refresh rate) actually boosts FPS, which in theory it shouldn't! 

If I disable that, I go from 35 FPS to around 28. Which is really bizarre, as that setting is only meant to cap FPS to half the refresh rate of the monitor. It is not supposed to boost the FPS. In theory, if your game runs at 28 FPS without adaptive V-sync, it should also run at 28 FPS with v-sync on, because it doesn't hit the ceiling of 30FPS (if you're using a 60 Hz monitor). I'm completely confused by this.

The game definitely plays smoother with adaptive V-sync - half refresh rate on. But it's still not what I would call decent performance when I compare graphics to hardware.

Two points from your post:

1. Adaptive v-sync/half-refresh rate: yeah, kind of interesting isn't it? There's a debate on a hardware site I visit frequently (Several times a day.). The reviewers are asking if fps should still be included in their gpu reviews. A lot of the responses are trending towards "Yes, but only to lend an objective credence towards your subjective evaluation." Which is my fancy way of translating, "If you say card 'x' runs game 'y' better than card 'z', back it up with numbers". As has been mentioned elsewhere, traditional film uses 24 frames per second. (There are differences between film/projectors and computer monitors. Or so I'm told. ) If film looks good at 24fps, why do we need 144fps monitors? Or, why would someone say that running game 'x' at 131fps makes it look so much better/smoother then running it at 97fps? (And, is the monitor refresh synced to those fps numbers? Etc.) We're all starting to realize that fps is a measure of gpu horsepower, not actual gameplay experience.  So, you've got performance which subjectively seems better, but the objective numbers don't back it up. I would disregard the numbers.

2. The part I bolded: But it's still not what I would call decent performance when I compare graphics to hardware. Really? What are you comparing it to? Meaning, what other game tracks the ballistics of THOUSANDS of bullets; has individual AI for HUNDREDS of soldiers; computes LOS between MILLIONS of locations; and does it all in real-time while pumping out hi-res models? That's the issue: CM is NOT a traditional game. Using traditional measures of hardware performance will just frustrate you. My example, upstream, discusses this. Perhaps your hardware is screamingly good...and is reflected in your ability to load 8km x 2km maps of the Alps, filled with 6 battalions of panzergrenadier. Or, it loads a turn in 3 seconds instead of 2 minutes. THAT is where you should look for performance.

Take those two points and look at what you wrote about occasional freezes and camera jumps. They seem to be two facets of the same issue. That indicates, to me, that the memory/cpu is overly stressed, not the gpu. Is it an older chipset? Slow RAM? Limited amount of RAM? Slow/single core cpu with a lot of background apps open? Etc. (The two machines I wrote about, with specs, was posted because I wanted to show real builds, not try to enhance my epeen. The one has a more powerful gpu (similar class, but widely recognized has having more "horsepower"), but a much older cpu/chipset which, in addition to being slower (3.5GHz vs. 4.?xGHz) is also far less efficient on a clock-for-clock basis. THAT is the key driver of my CM experience: cpu, not gpu.)

Capping fps gives CM a better feel. Anything much below 20fps gives me judders...and also my on-screen display. A steady fps above ~20fps makes for butter smooth gaming in CM.

[AttorneyAtWar]

I hate to do this because its a game that should not be named, but Mius Front was just released and its a good comparison to CM.

Mius Front tracks individual soldier spotting in and out of vehicles, models maps up to at least 8x8 km at a time with up to entire regiments on either side. Enhanced smoke effects, ballistics tracking, vehicles interact with terrain (bogeys go up and down, shell holes created by artillery during game are crossed in a realistic fashion) and can become immobilized. Armor penmetration data and shell ballistics are tracked and can be viewed after the game to see exactly what hit it and where it penetrated and says what system it damaged. While all of this is going on I get a constant 50-60FPS, if Combat Mission has a few dense forests it crashes to 5fps frequently unless I disable tree trunks while maneuvering the camera. Does this make Combat Mission a bad game? nope, will I still play Combat Mission even though this game performs better? absolutely, but we need to acknowledge that there is a problem instead of saying "CM isn't a normal game FPS and stutters are too be expected because its doing things no other games do" that excuse just doesn't work anymore with a game like this out now, infact that excuse shouldn't be said in today's industry. Lets acknowledge that the games performance is lacking due poor optimization brought about by an aging engine.

Let me head off the inevitable here too-

* No I don't think BFC is evil and isn't working on it because they don't care, they have one programmer I understand completely.

 

 

Edited March 5, 2016 by Raptorx7

[Bulletpoint]

Two points from your post:

 So, you've got performance which subjectively seems better, but the objective numbers don't back it up. I would disregard the numbers.

No, actually just the opposite. I just gave an example of how the actual FPS numbers are better with 'V-sync half refresh rate' on than when it's off. And these better numbers are also reflected in my subjective feeling of gameplay smoothness.

2. The part I bolded: But it's still not what I would call decent performance when I compare graphics to hardware. Really? What are you comparing it to? Meaning, what other game tracks the ballistics of THOUSANDS of bullets; has individual AI for HUNDREDS of soldiers; computes LOS between MILLIONS of locations; and does it all in real-time while pumping out hi-res models?

Why is the game doing all that in the orders phase, while the game is paused?

It should do that when I hit end turn.

Whatever holds back CM performance is probably not the actual simulation.

[c3k]

 

No, actually just the opposite. I just gave an example of how the actual FPS numbers are better with 'V-sync half refresh rate' on than when it's off. And these better numbers are also reflected in my subjective feeling of gameplay smoothness.

Why is the game doing all that in the orders phase, while the game is paused?

It should do that when I hit end turn.

Whatever holds back CM performance is probably not the actual simulation.

End turn: yes, the game is still doing a LOT. It's just not doing the ballistics. It still has to create the models, etc., on the fly, as you move the camera. I don't know how much data-swapping that entails, but perhaps the cpu is doing more in CM then in other games for that.