Problem in the way the game uses pc resources.

[Battlefront.com]

Update...

We've given out a very special diagnostics build to a number of people experiencing the Slideshow Problem. It's still too early to say why this problem is happening, but we're starting to get an idea of what the problem itself is. As I said in the regular Forum in the thread about the v1.11 patch release timeframe, we're working on fixing this problem before we release v1.11 (which is basically done).

The really odd thing is that it's affecting mostly nVidia users (I've seen only 2 ATI people reporting problems, the dozen or so others all have nVidia) and only a very small number of them. We've got plenty of people using the same video cards, similar systems, etc. and are having no problems playing the same exact scenarios. Memory leaks aren't supposed to be so picky

Hopefully we'll get this tracked down for good this week, but it might take digging deeper to figure out the exact cause.

Steve

[pute01]

Steve

In no way did I mean to be impolite. I feel relieved after reading your last post. But the fact was that before this, when reading between the lines, I got the impression the two problems I described were'nt high on your priority list. Now I feel I might have wronged you and I apologize and say that no-one will be happier than me when the problem is fixed.

By the way I have had the slide-show problem with both n-Vidia and Radeon cards.

[Jak170]

add me to the list with these problems with an ATI Radeon graphics card

[MehMan]

Add me to the ATI list too.

But I don't think this is probably a GPU problem, but rather a CPU problem and probably mishandling of multiple cores and memory management.

[Battlefront.com]

Pute01,

No problem. Just rest assured that we always take performance issues very seriously, even when they do not appear to be "our fault". Sometimes we can hack around the problem (like the ATI OpenGL click bug).

Now, I think you guys will be happy to know that the problem has almost certainly been identified. It's still puzzling why the problem has been so apparent to some and not to others, however we're content to not have all our curiosity satisfied if the problem goes away Charles is finishing the first pass at a fix and we'll be giving it out to people we've been working with to get this problem nailed. When we're sure it's fixed I'll give you the details. But indeed, it doesn't appear to be VRAM related (once we had some ATI guys speak up that theory went down the drain).

Thanks!

Steve

[Battlefront.com]

Oh, and a big thanks to you guys for posing some great stuff here. Because of the quality of reports we quickly ruled out "user error", which sometimes is very difficult to do. Once we were sure something was wrong (either drivers or CM itself) it was just a matter of finishing up the pathing fixes so we could look into it further. This is a great example of how working together gets things fixed faster

Steve

[SlapHappy]

Hi Steve

I think my greatest concern here is that as we move closer to the Normandy game, it could possibly turnout that Western European maps might end up being more demanding than the ones currently found in CMSF. If that is the case, then the only alleviation for poor FPS with the current code is to further limit map sizes and unit counts.

It would be great to play maps like City Ruins with a high building count and larger number of units in the Normandy game and still get at least playable frame rates.

Hope you get it sorted out!

Thanks!

[MehMan]

Good news!

[Battlefront.com]

SlapHappy,

We are also curious to see what happens when we put temperate terrain into the game. However, we're not going to shrink the map size. Instead, if framerate is not acceptable we'll get cracking on rewriting portions of the code that are found to be bottlenecks. We already know of one or two sections of code that could use an overhaul, but such optimizations take quite a bit of time and a ton of testing. Possibly even new artwork/models. It's why we have put code optimization on hold for the moment.

The Slideshow Problem was something very specific that was most likely introduced in v1.10. We're still a bit unsure why so few people actually had the problem, since theoretically everybody playing certain scenarios should have. I mean, we're not talking about small stuff like one guy getting 15fps and another one 25fps. Instead it was one guy getting 0.5fps and another guy getting 25fps. When comparing hardware profiles we didn't see an obvious indicator as to who would experience the problem or not. But it doesn't matter as long as it's fixed

Steve

[JSB]

I know I have a low specs configuration, BUT, I have the slideshow problem from the beginning even with SOME little scenarios.

Let's open the scenario called "A patch of dirt" downloaded on cmmods from the blue side for example (which I don't consider as a huge one). If I just turn the camera to the right to have a panoramic view of the map, everything is ok until the sun appears and then begins the slideshow! (down to 2-3 fps when the town and the US units appear).

All this happening BEFORE hitting the GO button (no unit selected either)!!!

If I keep the camera still, facing the units, and I move the mouse cursor, the slowdown is still the same, still visible, appearing as a "mouse cursor slideshow" (a fun-killer while playing!)

I don't have to add that the trees, balconies, windows and other trenches are disappearing/redrawing all the time.

All the things described here occur with all settings to the MINIMUM (fastest textures, fastest 3D models, no vec Sync, no antialiasing, no shadows).

My specs

AMD Athlon64 3400+

1 Go RAM

N'Vidia Geforce 6800 (128 Mo) (driver: 6.14.11.6921)

Running Windows XP

[cmfan]

Thanks for looking into this Steve. The game campaigns (and larger scenarios) are basically unplayable for me because of the "out of memory" bug. My frame rate drops and my soldiers start sliding around. I can usually tolerate it for a bit, and restarting the game helps a little, but after a while the game just crashes.

Tonight I was on my way to finishing the second scenario in the Army campaign - I had hit all the objectives and was just about to blow up the special forces HQ when the game crashed. I'm bummed.

Pentium D CPU 3.00 GHz (dual core)

Nvidia GeForce 8800 GT

2 GB RAM

XP

Ps. Someone had mentioned that setting the game to run on only one core processor might help? How do you do that?

[Jak170]

lol

and here we are complaining that the game is only running on one core processor. Maybe your already running on one core processor?

[cmfan]

Never mind, I probably misunderstood the problem with the processor cores.

[lucero1148]

Hi, I'm using a MBP with the nVidia 9600GT chipset 256MB vram and 2MB RAM and I'm also experiencing the slideshow and at .5 frames/sec actually I think its even less than that. I'm playing theHasrabit campaign and it's quite difficult to enjoy the game.

All best

Patrick

[JSB]

I know I have a low specs configuration, BUT, I have the slideshow problem from the beginning even with SOME little scenarios.

Let's open the scenario called "A patch of dirt" downloaded on cmmods from the blue side for example (which I don't consider as a huge one). If I just turn the camera to the right to have a panoramic view of the map, everything is ok until the sun appears and then begins the slideshow! (down to 2-3 fps when the town and the US units appear).

All this happening BEFORE hitting the GO button (no unit selected either)!!!

If I keep the camera still, facing the units, and I move the mouse cursor, the slowdown is still the same, still visible, appearing as a "mouse cursor slideshow" (a fun-killer while playing!)

I don't have to add that the trees, balconies, windows and other trenches are disappearing/redrawing all the time.

All the things described here occur with all settings to the MINIMUM (fastest textures, fastest 3D models, no vec Sync, no antialiasing, no shadows).

My specs

AMD Athlon64 3400+

1 Go RAM

N'Vidia Geforce 6800 (128 Mo) (driver: 6.14.11.6921)

Running Windows XP

Since this post, I boosted my RAM up to 3 Go (not especially for CMSF :cool: ) and I see ABSOLUTELY NO IMPROVEMENT in the slideshow problem.

[Battlefront.com]

Version 1.11 will fix the Slideshow Problem. It came down to two different problems which, for some reason, were far more noticeable for some people than others. There is no glaringly obvious connection between the people experiencing the problem since other factors came into play (using lower quality video settings, for example). All that matters is that it's fixed and you guys will have that fix shortly

The two problems were:

1. Information sharing. When a large number of enemy units were under observation by a large number of friendly units in close proximity to each other (like within buildings, for example), the number of calculations necessary to pass on that information gobbled up too much CPU time. Charles fixed this by rewriting the code so that it was far more efficient.

2. Issuing Movement Commands in huge numbers. If the CPU was already overworked, by the above problem or just general game stuff, issuing Commands to groups of friendly units at one time could cause some significant slowdowns. This is a fairly rare thing, though, since people generally don't issue simultaneous Movement Commands to more than a couple of units at one time. Even a platoon or so shouldn't cause a problem. This problem was uncovered when a tester basically issued Commands to his entire force at one time! Of course the moves were terrible, but the system doesn't pass judgement, it just tries to carry them out The solution here was to cap the number of Commands that the pathing AI tries to execute simultaneously, queueing up the others for execution after. This causes a slight delay for some units, however you have to be issuing Movement Commands to a TON of units before this limit ever kicks in, therefore it's highly unlikely any of you will notice this.

Thanks for your patience and good reporting!

Steve

[Piecekeeper]

Thanks for the update...looking forward to Pooh with all settings high.

[Michael Withstand]

Oh it seems I'm one of the luckier one. CMSF is not that fast on my system and sometimes it bogged down to 10s fps but overall it's still playable as long as I don't load a crazy scenario maps made by people who play CMSF with the NSA mainframe.

I've heard rumors and personally suspected that the fault may lie in the GPU itself. Current releases of NVidia cards starting with the 8500GT, 8600GT, 8800GT and so on may not have been designed to run OpenGL based game at an efficient level. My suspicion arised from the fact that I could play IL-2 Forgotten battle game on Open GL with my old 7600GT card and had fantastic flying fps at max setting while with my 8600GT I had a very very poor performance when the game was set to run on OpenGL. What I mean by very poor is unplayable. 10-12s fps which sometimes bogged down to slideshow. That's how powerful a 8600Gt is on an old openGL based game which the older 7600GT tackles very gracefully

Another possible reason is that apparently some game developers are extreme leftist in terms of their political disposition so they decided to make their software to have equal performance independent of the hardware configuration. LOL.

I hope this poor performance issue will eventually gets resolved.

[Flanker15]

I'm not getting very bad framerates but below what you'd expect for a computer of my spec (a max out Crysis comp).

I tried a little tweak and even though it shouldn't really do much I got some more FPS. I set the core affinity in the task manager for CM:SF so it was allowed to use all 4 of my cores instead of the default 1, it shouldn't do anything since you need to code the program to take advantage of multi cores. However there was an immidiate performance increase for me.

You do it by alt-tabing to the task manager while playing the game and right clicking the CM:SF process, where there is an option for core affinitys. You have to redo it everytime you start the game since it doesn't remember it.

[stoex]

lol Flanker I just answered your post about the multi-core thing in the seperate thread with a nice walkthrough ALTHOUGH I thought I'd read everything in this thread. I think I even referenced to your last post here in that other thread.

Man I gotta go to bed

[MehMan]

My suspicion arised from the fact that I could play IL-2 Forgotten battle game on Open GL with my old 7600GT card and had fantastic flying fps at max setting while with my 8600GT I had a very very poor performance when the game was set to run on OpenGL.

That's because the 8600GT is a very very poor card.