Is FB more optimized than the others?

[AstroCat]

All I can say is that over the course of CMSF release date till now I've tried many different computers and graphics cards and I've never been able to keep a constant frame rate above 30fps. As I explained before depending on what's "on the screen" I get as high as vsync'ed 60 but as soon as I pan back around to the map or raise up to a higher position it drops.

Sometimes its 40 other times it's 10, it's all over the place but two things are constant. 1) They are usually low and often in the 20s, 2) It hasn't mattered much what hardware I used. 

I've seen others report the same experience as well so I know I am not alone. Seriously not making it up, you could probably even see my posts through the years of this happening.

Oh and if there is some hardware configuration that can give solid good fps I'd love to know what it was, I'd try and match it on my next upgrade which will be this year.

Edited January 10, 2016 by AstroCat

[Thewood1]

I don't think he is making it up.  I have a similar experience.  I have run CMBN in its iterations on at least 10 PCs over the years as I upgrade work and personal computers.  I have run it on AMD A6, A8, and A10.  I have also run it on i3, i5, and now i7s.  The i5 is a Surface Pro3 and crappy Intel graphics and CMBN runs OK on lower settings.   But what puzzles me is my i7 MSI with nVidia 980 GPU.  It only runs a little better than my Surface.

I have always used Brinkman's corner as my benchmark.  As soon as I get a new PC, I load it up...after contacting the help desk for a new activation.  My Surface gets about 20-24 on average, spanning 40 fps with few units or trees in view to 10-15 with a lot of units in trees in the shot.  My MSI averages about 30 with the span being over 60 to high teens.  My Surface CPU runs at 2.3Ghz on high power and my i7 runs 3.4 Ghz on high power.  I use Process Lasso on both to monitor performance.

What is different is my perspective.  People running scenarios with hundreds of units and 5-12 sq mi maps at the highest settings are being unrealistic.  It is not what CM is built for.  If I was BFC, I would just put a hard cap on units and map sizes to get people complaining about performance off of the forum.  

With that said, there is something wrong somewhere, most likely in the older OpenGL libraries, when a solid performing rig sees performance drops frequently on mid-sized maps.  I am always puzzled by CMBN putting limited stress on my PC.  I monitor the RAM, CPU core 0, and GPU usage during play and even when the fps is dropping significantly, the load on all those systems is only between 25% and 50%.  I have no idea why CM refuses to use the power of the single core and GPU power that is available to it.

So while I think there are solid issues aligning with what some of the posters are complaining about, I don't think most people who originally started playing CMBN and its sister games have high demands for graphical fidelity and performance.

 

[76mm]

If I was BFC, I would just put a hard cap on units and map sizes to get people complaining about performance off of the forum.  

 I think this is a really bad idea...why reduce the game to the lowest common denominator?  If someone has an old/cheap computer, they shouldn't run the big scenarios, or should do so with reduced graphics settings...why ruin it for everyone else?

[Vanir Ausf B]

Yeah, that wouldn't do anything. Map sizes are already capped and the geometry of a map has at least as much effect on performance as size. A a 3km x 3km map of open fields can perform better than a 1km x 1km urban map.

Edited January 10, 2016 by Vanir Ausf B

[BarbaricCo]

I'm sorry but I don't believe you unless you consider 5-10fps good performance. And I hardly consider a 980GTX ti card from the stone age. I have done massive testing with all kinds of setups and configurations over the course of years so I'm not buying it.

Don't know about PC but I tested everything on Mac again (5 years old i5, 8gb RAM, ATI 9600m 512mb).

In short, bad and unplayable fps is something I will get playing huge map and the if camera is looking at lots of the vehicles (tanks seems to cause the problems).

For example playing CMRT Gog and Magog as a soviet with as I remember 6 SU-76 and lot of the infantry is smooth across the map. Lower, zoomed, raised camera just work without hiccups. All maxed except models and terrains to better.

Playing Gog and Magog as Axis is painful because just looking at the panzers (maybe 20 of them) drops fps to unplayable level. And it doesn't make a much difference if lowering details or anything to fast and low.

It just seems like Mac hit the wall.

Again this is something I saw in other games too, and my bet is on the number of 3D audio sources. Probably multithreading can help, but I just don't play those huge maps with lots of panzers causing problems

 

 

[Hister]

Knowing which hardware combos allow the game to run 30 FPS or more constantly with all the setting maxed out in any available map size and unit numbers would be of much interest to me.

 

[c3k]

Knowing which hardware combos allow the game to run 30 FPS or more constantly with all the setting maxed out in any available map size and unit numbers would be of much interest to me.

 

This is hard to quantify. One of my regular pbem partners created a map 2km x 8km. It had a LOT of elevation changes, woods, water, and some buildings. Each side fielded about 3-4 battalions of infantry and tanks. I could play it, but not at 30fps. (Of course, all graphics are always maxed out.  ) It would run in the teens, for fps. A lot depended on the camera angle. It eventually stopped working.  (I think the pbem file sizes exceeded an in-game limit. Something about 2 to the power of 32.)

I don't know if any hardware combo could run that. If the camera is at height 1, and facing the edge, sure 30fps is easy. If it's at height 4, and facing an 8km length of map, well, bog-city. There are soooo many variables...

Ken

[BarbaricCo]

Knowing which hardware combos allow the game to run 30 FPS or more constantly with all the setting maxed out in any available map size and unit numbers would be of much interest to me.

 

That can only happen if BF port CM to PS or Xbox .  One hardware spec with known limits.

It is in the essence of PC gaming to never be satisfied with performance and to spend days in preferences panel  . That drives the industry.

If he (PC gamer) can buy i9 8Ghz with 1Tb RAM, he will probably try to simulate Prohorovka and again would be disappointed with performance. At least I will try to do that  .

Anyway, It seems that CMBS has little bit better performance than CMRT, so probably we can except same gradual improvement from the CMFB. It also has yet to be seen what never game technologies like Metal or Vulkan if implemented, can do for game like CM.

Also, can anyone post experience with newer retina iMacs and CM.

[Hister]

Edit: Double post.

 

Edited January 10, 2016 by Hister

[Hister]

Ha, I suppose that I would have to changr my words to "reasonable map sizes and unit numbers" then wjatewer that might be.

[A Canadian Cat]

What is different is my perspective.  

So while I think there are solid issues aligning with what some of the posters are complaining about, I don't think most people who originally started playing CMBN and its sister games have high demands for graphical fidelity and performance.

I have a feeling that differing perspectives is  what is driving this frustrating debate. For me I do not regularly monitor the fps. U do not measure my happiness over performance by watching the fps. I do it by is the game animation smooth? Is the camera panning smooth? Is the keyboard control smooth? And the fact is it is for me the vast majority of the time. Sure if I push a scenario with a lot of units or a huge map there can be issues. 

 I can play the largest shipping scenario s consistently with good performance.

 

Ha, I suppose that I would have to changr my words to "reasonable map sizes and unit numbers" then wjatewer that might be.

My rig can do that  no problem. It is a less capable rig than @AstroCat's . it is an i5 with an NVidia 760 and 16Gb of memory now running Windows 10.

[AstroCat]

I have a feeling that differing perspectives is  what is driving this frustrating debate. For me I do not regularly monitor the fps. U do not measure my happiness over performance by watching the fps. I do it by is the game animation smooth? Is the camera panning smooth? Is the keyboard control smooth? And the fact is it is for me the vast majority of the time. Sure if I push a scenario with a lot of units or a huge map there can be issues. 

 I can play the largest shipping scenario s consistently with good performance.

 

My rig can do that  no problem. It is a less capable rig than @AstroCat's . it is an i5 with an NVidia 760 and 16Gb of memory now running Windows 10.

Something to remember that in my experience the "faster" the config doesn't equal usually any increase in performance or super minimal at best. You could from a 680gtx to a 980gtx and only get small performance increase, which in my opinion is very odd.

[CarlWAW]

I think a dedicated page displaying performance results from different systems would be very helpful for existing and future customers.
How could CM performance be measured objectively? Maybe with FRAPS and a dedicated scenario in replay mode for each game family?

[BarbaricCo]

Something to remember that in my experience the "faster" the config doesn't equal usually any increase in performance or super minimal at best. You could from a 680gtx to a 980gtx and only get small performance increase, which in my opinion is very odd.

Only part of the CM calculations are graphics GPU related. There's a lot more real physical things simulated that can be cheated if game doesn't want to be simulation or only pretend to be one.

Again, there are always room for improvements and optimizations, and there will be some. There are legacy code problems but CM is not FPS. If it is, actually it would be easier, as it is super smooth on ground level (5 years old Mac with prehistoric graphic card).

And if there are slowdowns on huge maps, they can be used to double check what on first thought seems like a genius tactical idea  

[Hister]

I think a dedicated page displaying performance results from different systems would be very helpful for existing and future customers.
How could CM performance be measured objectively? Maybe with FRAPS and a dedicated scenario in replay mode for each game family?

That was already done

 

 

. Thread is to be found in the CMBN tech forum.

 

[Baneman]

...

And if there are slowdowns on huge maps, they can be used to double check what on first thought seems like a genius tactical idea  

 Well put.

[Ron]

Does anyone have a link to the optimization guide for AMD/ATI video cards? I searched but could not find them.

Just bringing this forward again as I would like to see this guide also. Thanks.

 

[Battlefront.com]

I've seen others report the same experience as well so I know I am not alone. Seriously not making it up, you could probably even see my posts through the years of this happening.

 

I don't think he is making it up.

Please read more carefully.  The objection was the statement that people are playing only at "5-10 FPS" and that people consider that "good performance".  Nowhere has anybody said that is their framefrate and certainly only a crazy person would say they are happy with that.  So yeah, it's a bunch of hyperbole which only makes this sort of discussion less likely to be helpful to anybody.  Which is why I called AstroCat out on it.  I wanted to nip that train of thought in the bud because this is a good discussion in other respects.

What is different is my perspective.  People running scenarios with hundreds of units and 5-12 sq mi maps at the highest settings are being unrealistic.  It is not what CM is built for.  If I was BFC, I would just put a hard cap on units and map sizes to get people complaining about performance off of the forum.  

As others have said, this is a bad idea.  While we do cap map size, there's way too many variables to put a reasonable cap on what is included.  A small, dense terrain map with tons of terrain variety will run slower than one that is much larger but less dense or varied.  A map with tons of units that are spread out or come in in waves will run better than a smaller scenario with them all crushed together in one spot.  So on and so forth.

With that said, there is something wrong somewhere, most likely in the older OpenGL libraries, when a solid performing rig sees performance drops frequently on mid-sized maps.  I am always puzzled by CMBN putting limited stress on my PC.  I monitor the RAM, CPU core 0, and GPU usage during play and even when the fps is dropping significantly, the load on all those systems is only between 25% and 50%.  I have no idea why CM refuses to use the power of the single core and GPU power that is available to it.

Because the engine would have to be recoded to take full advantage of multiple processors. As I've said before, one of the problems we have is we need to be careful about how much of our resources we invest in rewriting existing code/content vs. new code/content.  Since we wrote the engine before multiple processors were standard the inherent engine is single processor only.  With the v3.0 Upgrade we did offload a bunch of stuff to use multiple processors, especially when loading maps/scenarios.  We will do more of these in the future.  But a whole rewrite?  Not going to happen.  We'd be better off to abandon CMx2 in favor of CMx3 rather than reengineer CMx2 to that extent.

So while I think there are solid issues aligning with what some of the posters are complaining about, I don't think most people who originally started playing CMBN and its sister games have high demands for graphical fidelity and performance.

 

Yes and no.  For sure the performance could be better than it is.  I've never disputed that and never will.  The question comes more down to where people think the line is between "playable" and "acceptable".  Absolutely, for sure, without any doubt at all we have people in this thread that are fine with 20fps and we have people who think anything less than 30fps is an abomination.  We have people who understand that we don't have vast amounts of resources to throw at performance improvements and therefore set their expectations for improvements accordingly.  We also have people who do not understand and simply say "Game X can do this, so no excuse for CM not being able to match it" regardless of the details.  Our ability to deliver is somewhere between this and it will always be very scenario/game specific.  Playing with a Regiment on a max map in an urban environment will always be on the pokey end of things compared to a company on company fight on a smaller map.

We have steadily improved the performance of CMx2 over the years.  Games that are based on the v3.0 Engine are significantly faster than v2.0 and even more so than v1.0.  CMSF and CMA are lagging behind all of them, obviously, since it is even older and has not been updated at all.

Steve

 

Edited January 11, 2016 by Battlefront.com

[AstroCat]

Not trying to make a fuss but I felt my claims were baited so my response about being ok with 5-10fps was based off of CarlWAW below post:

How it is possible that someone with a modern desktop computer can have performance problems with CM engine 3?
I have a medium range gaming notebook only (Acer VN7-791G; GTX 960M) and I can play at highest settings big maps with several batallions without any performance problems.

I am very pleased with CM's performance which I find very well suited for mobile computing. No need for an energy wasting desktop-monster from the stone age of computing.

So, yeah running that computer I see 5-10fps, maybe 10-20 "play at highest settings big maps with several batallions without any performance problems." I know that claim is not true unless he is ok with super low fps. Also, this part... "No need for an energy wasting desktop-monster from the stone age of computing." I think that set the tone... I should have probably just ignored because it was bait.. but still no way that computer is rocking 60fps "highest settings big maps with several batallions without any performance problems." 

Edited January 11, 2016 by AstroCat

[A Canadian Cat]

So, yeah running that computer I see 5-10fps, maybe 10-20 "play at highest settings big maps with several batallions without any performance problems." 

So, I really think that your machine might have something off about it. Because I have played a game in CMBN with multiple battalions (I had two) in a large map (I forgot exactly but I think it was 3x5 with lots of elevation changes and lots and lots of forest). On max settings it did drop below 20fps many times but I was amazed at how well it did. As I moved around the map to give orders there where times when the camera panning and keyboard controls were sluggish but it was not all the time and much of the time the fps were above 20.

For anything else I run with the best settings and do not experience sluggishness or hesitation. When i measuer the fps it goes between 25 and 35 or 40. The feel of the game controls is fine and the play back is smooth.

but still no way that computer is rocking 60fps "highest settings big maps with several batallions without any performance problems." 

Right, I do not normally see 60fps. But I don't care, what I care about is that the camera and mouse controls work smoothly anyway. 

I see too places where there are disconnects. One is those of us that are OK with the performance are OK with 20-30 fps, as long as the mouse and keyboard controls creat smooth camera panning. For me the feel of the controls is how I measure my satisfaction not by a fps measure. Second many of us do seem to be getting better fps numbers than you are reporting. It sounds like there is something in your configuration that is causing a slow down. Or something in mine and others that is speeding us up.

[AstroCat]

I think it's expectations... I'm pretty sure I'm seeing the "correct" fps for my settings/hardware it's just I might have higher expectations for performance. I know it's mostly apples to oranges... but when I can run AAA games with everything max with graphics that "wow" even this old jaded gamer I come back to CMx2 and it's a bit jarring. I understand the reasons but that doesn't make the game experience any less real. It just seems with graphics at the CMx2 level a higher end system shouldn't even break a sweat, even on larger maps or ones with lot's of buildings but it does, comes to a crawl often.

I'm still geting CMFB of course, and probably whatever else comes out I am just hoping for some legit performance improvements along the way. And... 2 more pet peeves, the shadows are still pretty rough/jagged and "flickery", and I still not a fan of the "floating world" game space... couldn't we have a fake level horizon that looks connected? I think that would up the "immersion" factor a bit, I remember CMx1 had this.

 

Edited January 11, 2016 by AstroCat

[A Canadian Cat]

So am I reading this correctly? It sounds like you are saying you want to see higher frame rates. That you just expect that the fps number should be higher.

[76mm]

After reading this thread, I went and checked my frame rates with FRAPS--usually I never bother.  Sure enough, they were around 30, and I'm OK with that.

But one thing I really dislike about this game is the very distinct "blur line" or whatever people call it.  Trees/shrubs aren't visible, and fields of yellow flowers turn into bland brown stretches.  How to reduce this effect, either by increasing the radius of detail or increasing the amount of detail shown beyond the radius?  I've tried playing around with the settings but without much rhyme or reason.  Should I focus on the video settings in the game or on my video card?  I know nothing about this stuff...

Also, do I need to restart the game after changing video settings to see the effects?

BTW I've got a 3.4 gig processor/8 gig of RAM, and a GeForce GTX 580.  64bit Win7.

[CarlWAW]

Not trying to make a fuss but I felt my claims were baited so my response about being ok with 5-10fps was based off of CarlWAW below post:

So, yeah running that computer I see 5-10fps, maybe 10-20 "play at highest settings big maps with several batallions without any performance problems." I know that claim is not true unless he is ok with super low fps. Also, this part... "No need for an energy wasting desktop-monster from the stone age of computing." I think that set the tone... I should have probably just ignored because it was bait.. but still no way that computer is rocking 60fps "highest settings big maps with several batallions without any performance problems." 

@AstroCat I think to make it fit to your rant you are making things up I did not say.

But a picture says more than thousand words.

I searched for the biggest savegames on my HD. I wont say from which the screenshots are to prevent spoilers, but they are from CMRT (game engine 3). That's the position, where I achieved the absolutely lowest framerate according to FRAPS on that map. I could not get into the single digits and usually the game runs around 30 fps.

Movie lighting with customized shaders enabled:

As I said: Big map. Several battalions. Best settings. Even very long shadows. Full HD resolution.

Sorry, but I am very satisfied that this great game/sim runs so well and can look so good on an already very affordable mid-class gaming notebook.

[Hister]

Carlwaw and AstroCat, you should swap this savegame file and then see how much frames does Astrocat get using the same ingame settings.

 

I personally wouldn't bother with fraps but because I was/am experiencing non smooth gameplay with sometimes delayed mouse inputs I resorted to fraps to measure it and see what frames are cpnsidered too low for my eyes.

 

When fps is 27 I already spot non smoothness, when 24 it's bordering still acceptable but already ruining my immersion, below 20 is a no go any more. Just my personal peet pevees. Not saying this is everyone elses experience.