Graviteam

I've actually done quite a bit to try and improve my FPS recently, cleaned the registry, defragged the discs, removed as many background apps and services as possible, tried CPU infinity and priority settings and disabling CPU parking and reduced all windows appearance settings. Cleaned the memory, overclocked CPU and/or GPU a little, reduced in-game graphics setting and nvidia settings and reduced resolution as much possible but the FPS increase, if there is one at all, is absolutely tiny.

So my 2.5GHz Quad core + 4 GB RAM + Nvidia GTX550TI just won't cut the mustard with this game anymore even in smallish engagements. I have to wonder if there is some sort of memory leak as performance in the game drops off even more the longer you play it and the same thing happens with Steel Fury after a while which makes life harder. It is a shame as, IMO, this is the best game of the genre that I've ever played. One thing that I have noticed is that say you have 11 FPS and then move around the map with the mouse the FPS can drop low and stay low at 6 FPS for example.

Cheers

Aces

It seems to me that this game has problems with multi-core processors, something like ToW series, processor has always been a bottleneck for that game.
Also I managed to gain back some performance by tweaking with the project lasso software.
http://bitsum.com/processlasso/

Process lasso has been saving my gaming bacon since ToW2.   
'Windows dynamic thread priority boosts' seems to always cause me FPS problems, turning it off for games (via PL) has made games suddenly playable on more than one PC of mine.

I do feel the need to do a HW upgrade in the not too distant future but I really can't say when/how.  That said I don't expect GT to make their new project super-specs PC only.

We do not know what GT is doing to optimize things, & GT must have seen the problems on the release of the last iL2 iteration which it seemed very few people could actually  play becos of the HW reqs. 
That kind of thing would possibly be a financial disaster for GT.  I'm guessing a lot of fans especially in Eastern Europe haven't got the $ to throw at new rigs.  (not that the West is a whole lot better these days...) so GT will try to make the game accessible HW-wise.

Dane - I feared it might be a virtual paper-avalon hill type-game thing  WTF no eye-candy !!!   Drop & give me 20!    :o ;D :P

Hi, thanks for the link, I looked at process lasso briefly the other day. I will have a more in-depth look this weekend. I think that you might be correct regarding multi-core CPUs and the game, I've reached a similar conclussion. I have tried other CPU affinity settings without much difference in FPS. I think that the pan around the map and FPS slowdown thing where Framerates do not "recover" after stopping panning and stay lower is odd.

Cheers

Aces

Edit: Hi chashka17 I hope that you are right regarding the specification thing. I don't buy (or play) many games so I'm fortunate that I can still just about get by with my current hardware setup as there are not many "next gen" type games that I would be buying. Is there any particular tips for using process lasso that you recommend?. I shall try Windows dynamic thread priority boosts thing that looks promising. I don't really want great FPS, anything over 25 FPS consistently would do me if I could ever get it.

Close Combat - The Bloody First is coming next year and it's gonna be in 3D.

So far the only realstic, strategy-wargames that I know of and that I actually enjoyed next to the Achtung Panzer series are Theatre of War, Combat Mission and Close Combat series, and that's really it... nothing else on the market worth playing. If anyone knows any other similar game, please sare that info :)

GTOS dont have any problems on multi-core processors.


Four donkey harnessed to a cart dragging it with the speed of the slowest donkey. Think about this :)

Not need to use strange programs, they don't gain speed.

OK, It doesn't have problems with multicore but it's more likely that it doesn't utilize power of the multicore systems. It's does look that way. And, I actually gained few FPS with project lasso :)

I'm still going to try process lasso, a few frames for me could DOUBLE my framerate "down low n' dirty" in Shilovo. I let the game decide the best settings for me the other night and it pretty much maxed out everything, in fact its assessment 99.99% matched my chosen settings so, if the game thinks that my rig is capable of delivering playable framerates having accessed the capabilities of my rig, assuming that 6FPS isn't considered a playable framerate then where's the difference?.

Andrey, I know it is a loaded question but with my PC spec (Quad core 2.5GZ CPU  and 900MHz GPU) would you expect to see 6FPS in Shilovo?

Regards

Aces

Its not "real FPS" its like placebo, which are obtained as a measurement error, sorry.

Since 1960, found that, if 95% of the program can be parallelized, the theoretical maximum speedup using parallel computing would be approx 20 time, no matter how many processors are used. Is called Amdahl's law.
But games can't be parallelized at least 50% (in excellent case) in real world we have 5-30% in mass, this give to us need in 2-3 core (in best case) not more.

Think about this, surely if some small independent program would provide a gain of FPS, would the game developers (who make many  more and more complex programs) can not be directly inserted it into the game?  :D
Well, purely logically think, can this program improve something is real or not :)

Hi Andrey placebo or not if gameplay is smoother and more fluid then I'll try it. I believe that there are measures one can take regarding closure of background processes and applications, disc-defrag, defragmenting/freeing memory etc.. which can improve framerates in a game.  I'm not hooked up on framerates only as a basic indicator of performance, 10FPS, 60FPS or whatever I don't care but I do know when gameplay is slow and un-responsive versus fluid gameplay.

There doesn't seem to be many, if any settings in the game that can be reduced to increase performance so it has to be CPU-bound IMHO.

My PC meets (CPU) and exceeds in other respects the recommended system requirements for Shilovo as printed on Graviteam's website, maybe the CPU spec. should be upped in this information to take into account developments in the game unless, of course, higher FPS are to be expected than I'm experiencing with this DLC. I can't believe that 5-6 FPS with the recommended CPU etc. could be considered as satisfying gameplay when drawing up these spec..

Yes I do not mind though. Just as it offers a real look at things :)
In the current state of multi-core, it is well suited to play in 2-4 games at the same time on the same PC. But it is bad to play in one game at 2-4 times faster. That's it.

And you can believe in anything, as some believe that DX11 allows to make a beautiful picture in the game than DX9 or OGL  ;D



Its good, but Windows not real-time OS, all measures very dependend and vary from time to time. If FPS vary twice or more - you can _objective_ measure it, if on 5% its lower than measurment error.
But of course you can _subjective_ think of that was better  ;D

"Its good, but Windows not real-time OS, all measures very dependend and vary from time to time."

I agree :)

My FPS don't vary, I've re-tested the same "test" mission many times over many days, after a clean boot and at the end of the day and the FPS range is very much the same. the mission has lot's of dust so I tried reducing this setting to no effect, not even 1 FPS gained!. I decreased lighting, and textures quality in game, disabled the high texture pack as well for good measure but I have a 2GB video card FWIW. nothing has made a difference so it's not a graphics thing for me. But some of these settings are at least in part computational and I would expect that disabling or reducing them drastically would have at least a minor effect on reducing CPU loading and that is not the case for me. So it must be CPU IMO, 2.5GHz not enough to achieve sustain decent FPS/fluid gameplay. My only question regarding multi-core as opposed to CPU speed is is there any penalty using a multi-core CPU compared to a single core CPU of the same speed?. I appreciate that there's not necessarily any advantage in a multi-core CPU.

All that you list, it is calculated mainly on the GPU. If the game delays in the calculations on the CPU (and when the number of units committed to 1000+ in them like in Shilovo with unlimited radius - it delays), disable graphical options does not make sense. You get bad picture and the same FPS. Programs like listed above dont help too.


Now there was some strange situation. When the frequency is not growing, but improved architecture, and increases the number of cores (thank god only to 8  ;D ). So to say that are fasters is difficult. I think newest CPU (not in low-end) is better than oldest.
But more cache, more GHz and advanced architecture is the best choice.

P.S. For multi-core fans: Mius (64-bit version) is better parallelized and utilize more of core power :D than GTOS in many many infantry units case especially.

Hi An drey, this isn't with 1000+ units it is with a handfull of units in a quick battle!.

You might try this.I use this method occasionally when using the Ultra High grass setting.

Go to the options Time settings(7.0) and change them all to no action.
This enables you to up your game game speed and not have to keep upping the speed everytime something happens.
I usually only have to up the game speed to 1.5 to get everything to run good.
On the Summer maps about x2.

I also trim all the dust and smoke even at the lowest setting it still produces quite a bit.
Also,try enabling simple shaders and disable shading in the options.(you have to do both in tandem).
It's important that you enable simple shaders when disabling shading and the reverse applies also.Otherwise the game won't play right.
These two options menu tweeks make a big difference in FPS and the game still looks and plays decent.

My input on this subject.
Ive been trying for a while to get to the bottom of this, but without success.
Ive had Fraps running and a CPU/Memory usage indicator running in the background.
At game speed 1.0-0.9 between 40-60FPS. Down to 0.5 @ 11FPS.
CPU usage is at 12%, memory at 18%. CPU increases to max 16% with as many units as i can cram into a QB all moving at the same time.
Dust and smoke seem to have an impact on FPS. As does the "brown grass", high polygons, i dont know? But knocks it down to about 8FPS.
Im guessing its not a CPU but more of a graphics issue.

What spec PC does Andrey have to run this game i wonder?

Something I already confused. If you have a FPS drops when a lot of smoke on screen, it is clear and it should be (actually there is such in any game where a lot of smoke). To reduce the negative effect is used decimation setting (as an option - lower screen resolution).
CPU or multi-core dont help in this case in any way only GPU with higher fillrate (or reduce fillrate through decrease dust/smoke density or resolution).

But as I understand on above posts - the problem is not in the smoke or dust? What to do with a multi-core?  ???

Thanks guys,

I did enable simple shaders, trimmed back dust and smoke and diabled shading and got about 3 FPS gain on average which more or less nudges me into double figures FPS-wise in Shilovo quick battle.

Guess my poor 'ol PC isn't really up to the taskm anymore.

Cheers

Aces

And how many units engaged in combat?
Its only in Shilovo or in other areas too?

I cant see it being a CPU, multi core issue,
I have a water cooled 3930k @ 4.0ghz, thats 6 cores, 12 threads, and it never gets past 16% of usage. (although the game doesnt use all cores anyway)
So it has to be graphics related.
If i position my camera at ground level between a green field and brown grass field and rotate the camera (this is with no smoke and 1 unit per side in QB)
The FPS goes from 50-60 in the green field down to 11 in the brown field, and back up to 50-60 as the camera rotates.
Is it a Card issue? Lack of Ram maybe, mine only has 2GB. Or manufacturer, i have an AMD 6970? (with latest drivers).
Is the game opptimised for NVIDIA?
Anybody using a Crossfire/SLI set up get any better?

Andrey, its worse in Shilovo and Volokonovka. Summer maps.
The number of units doesnt seem to matter (well not to me anyway)

Yes of course Amdahl's law say for GTOS is 2-3 core in peak. 12 threads (cores) is 8% full load for one. 16 is approx 2 core/thread load.
Threads is really not usefull at all, and in most cases is bad.

Yes fields and ultra high grass settings is experimental, not use it in real gaming, please :)

Is too more grass blades (and more overdraw) in UH settings especially if you on ground level. If we have some of free time we remake grass algo for UH settings.

No, is too much for hardware (GPU) to draw it all. Too more overdraw at ground level. The same problem in case of dust/smoke - slow down if many pcs draw one over another.

Andrey,
Thanks for the reply, that explains alot.
Will Mius be any different?
Utilise more cores, or tone down the grass abit?
I generally play the winter maps at the moment to avoid the issues, which is a shame, as i like the Shilovo map.
And the different Arty options (sWG 41!)
I'll just have to stay away from the grass! ;D

In somthing yes, may be in some somthing.
Until we have not touched tacticals only game core and operational.

No, as I say before is GPU problem called overdraw. Dont have any known good solution, only some hacks and fakes.
Even in last CryEngine wide grass areas in some aspects tuned by hacks.

I think in RTS/wargame most people fly over ground not on ground itself :)

Hey everyone - option 6.16 "change game speed automatically" - set this to "disable" and you will see the difference!! My quad core machine also had the slow downs until I did this. Simple.

I play this game mostly with the camera at ground level.That is where I enjoy watching and controlling my units in this game the best.
Other RTS games don't allow this that is why I don't play them.

I don't like the large battles because I have to lift the camera up higher to control all the extra units and that is why I don't play the campaigns much and prefer Quick Battles with smaller units.

Quick Battles account for 90% of the replay value of this game for me.
I would very much dislike this game if it turned into a European Escalation type RTS.
I also think the balance in this game is perfect as long as the campaign battles are not overly large and kept to regimental size at the most.

IMHO bigger isn't always better for this game unless we are talking about the map size.
The graphics are excellent in this game and having to put the camera up at a higher altitude destroys that perception along with the immersive value of this game.

This is my favorite game and I would hate to see it become radically different from what we currently have.

Hi folks,

I too have disabled auto gamespeed to off as I prefer it that way but it has never increased my framerates. The mission I've been using lately to test on the Shilovo map consists of two platoons of tanks on either side, so about 20 vehicles in total. The attacking tanks have to cross one of the brown flattened grass "fields" on their way to engage the enemy and it is this that mostly causes the framerates to be dragged down from an absolute  max of about 15/16 FPS to a maximum of about 6/7 FPS. I think that these brown grass areas are at least in part responsible for the big drop in FPS This is  happening without any fighting  taking place as the two forces are still closing on each other.

I too mostly play at low level as my interest is watching the action and doing movie capture, not that this is really possible when FRAP drags frames down by a minimum of another 2-3 FPS and therefore my captures can be as slow as 4 FPS which is totally useless for this purpose. I also don't wish to turn off too many graphics settings as this also defeats my purpose although there aren't many settings that really make any difference in terms of FPS except perhaps the dust/smoke which does win me a couple of frames without which movie capture would be totally impossible.

BTW. For any movie capture fans I've changed the hotkey that toggles the Key point flags to the same key that disables the HUD so that one keypress does both, I'd like to do the same with the function that displays the unit icons but that toggles through three states and thus one key press isn't enough. But it does help to quickly disable all on-screen indicators. I did a similar thing with Steel Fury and bound the disable HUD hotkey to the same key that turned of on-screen messages so hiding all onscreen indicators in that game too with a single key press.

I'd like to get one of these if and when funds permit "Avermedia Live Gamer Hd C985 - Video Input Adapter - Pci Express " as it gets very good reviews.

Kind regards

Aces

Well that's what I'm saying, it doesn't utilize the cpu power of multicore systems, and it has nothing to do with the grass, drops are happening when there is a lot going on at the field, everyone is shooting, shells are falling, tanks getting hit, AI maneuvering across the battlefield, and changing graphical options doesn't do anything and yet cpu usage doesn't go above 25%.

"50-60 FPS"

Oh how I wish! :D

I'm kinda surprised to see the highest grass settings called 'experimental'. 
Assuming this isn't a joke,  :o I can see why such a setting with all the drawing req'd would kill performance.  Even if it is experimental I still like to max settings & see what I can get to run, for me that's half the fun with eye-candy.  Also tells me when I need an upgrade... ;)

& OT
On the subject of Process Lasso - yes there are many fake optimization tools about.
The logic that if there was a magic fix then everyone would do it isn't really fair.  e.g. How come Windows still 'needs' a third party defragger to defrag anything properly, surely if you are in the business of writing an OS you want to keep it running smoothly ...or maybe not?  Given the resources at MS I would have expected a lot better of tools like defrag & back-up a looooong time ago, yet 3rd party stuff is about & works.

As far as I can tell, games try to optimize the environment for themselves.  PL tries to wrangle the OS / PC in real-time for everything that is active.
The user can specify certain priorities etc for processes of his choosing.  So I right click an exe & assign some values & thats that.  ToW2 after the last patch was lousy with stuttering & grinding, used PL to turn off 'Win dynamic thread boost' & the game runs perfectly again. I also pushed it to use cores other than 0 & 1 which is what it is programmed to seek (but so is a whole lot of other stuff - hello bottle-neck...) I can now play the game.  So I know it helps in this case.

If this test isn't enough methodology/evidence:-
http://processpriority.com/ (http://processpriority.com/)
then my own findings were that it does work beyond any notion of a placebo.  In order to reach that conclusion I used it during it's trial & then uninstalled it.  (I could have just used the free version of course) then spent the next couple of weeks being annoyed at how badly my machine ran.
Put it back on & smooth performance was back. 
Installed it on an old laptop & suddenly that was also usable. So it's not all empty promises as far as I'm concerned.

Aces - easiest way is to start the game in question, tab out & open the main PL window, right click the exe  (& .bin for ToW) & select 'classify as game'.  For ToW I also click 'Priority class' & uncheck the thread boost option at the bottom of the dialog.  It's easy enough to play around with despite how it may appear to a new user.  Free version will also do this job.  You can also use options/configure games & MM procs & select/add any exes you like.

Hmm I tried it and with high grass settings I have on the green field 60 fps and on the brown 53 fps.
But when I turn on HDR I have on the green 60 and on the brown field 34 fps.

I have Q9550 , AMD 6970 and 4gb ram

Well,
It would appear to me  that alot of us have the same issues with the "brown fields" to various different degrees,
And as Andrey says, its experimental. Im not really sure what that means?
Im also not sure why you would include something in a game that no ones PC can run at descent FPS?
But for the playability of the game at ground level i hope its something that can be looked at.
I really dont want to play Muis from a "Gods eye" perspective, as this would kill the game for me, and by the looks of it for many other people to.
The game has such great eye candy, you just cant see it if youre up in the clouds!

Agreed :)

The brownfields don't seem to be a problem for my PC as far as FPS are concerned.

The major problems for me are too many units and some optional graphics settings.
Such as-Ultra high grass,smoke and dust,shading  instead of simple shaders and
very heavily tree populated woods.

My sentiments also.
Being able to play real time at ground level is what sets this game apart from other RTS games and why it is my favorite game to date.

Losing this feature would not only be a game breaker for me,it would be a very sad day for my personal gaming enjoyment as well.

+1000!!!

+2000, I'll see your 1000 and raise you 1000! :D

The only issue performance wise is with ultra high grass which lowers the frame rate down to about 30 fps and lower. Usually with everything cranked right up (shadows, HDR and so on), I get over 40 fps.

I'm also have option 6.10 (Force vidmem size to MB) to 1024.

Just to compare, I have the following PC: i7-920 at 4GHz with a 660 Ti 2GB card.

I don't have Shivolo yet so I can't say how that will run on my PC.