"White Artifact" in scenery P2

Games can be written specifically to benefit from SLI, especially newer gamer and even more so the benchmark suites. 3Dmark is a prime example. Over the years it has been accused of favouring ATI or NVidia over a range of generations of cards. At the moment the accusation is that it favours NVidia.

CFS3 is one of the games that doesn't benefit.

Sli uses the second gfx CPU to load balance the graphics demands. You are still running a primary card and it does most of the work. The secondary card is there to chip in CPU cycles when needed. The ram it has on board doesn't get used at all. The manufacturers never say that, because it makes people happy to think they are getting twice the graphics. But the reality is very few titles perform twice as efficiently. Most games can achieve a 30% increase in frames, some excel, some fail miserably. OFF/CFS3 falls into the later category.
 
I appreciate the further detail/clarification on the functionality of SLI - I think it's generally understood that OFF doesn't benefit from SLI.

Actually, the results I posted were more particularly aimed at the suggestion Pol had made concerning SLI as a possible culprit in the 'stutter' I experience with AC detail slider at 3 or above. As even he said, there's something not right there.

So, to say it doesn't benefit is one thing - but it doesn't appear to be causing the issue I have, at least not as far as I can tell.

As an interesting sidenote, you'll notice I use 3DMark03 for one of my benchmarks; there are a few reasons for this. Among them: 3DM03 was released in February 2003. SLI, on the other hand, was introduced in 2004. The details I have on this version don't indicate any support for SLI configurations. It seems that 3DM03 would be completely 'unaware' of how to react differently to SLI - but it most certainly benefits, as my results prove.

In other words, I don't know that my benchmark would be 'fooled' into an unrealistic result by using SLI. I am familiar with the history of some of the benchmarks and accusations of being skewed favorably, and I can't deny it...but I'm not measuring ATi v. Nvidia, I'm measuring same system, same exact video card...just in SLI as compared to not in SLI.

Further, CFS3 was released in 2002. I find it very interesting that a benchmark written before SLI existed can obviously benefit from it, but that a game not much older (less than a year?) doesn't seem to at all.

I would submit, then, that (logically):

A game (an app) can be written explicitly to benefit from SLI arrangements, but an app can also benefit without having been explicitly written to do so. I'd bet money 3DM03 isn't the only example of this.

It seems, logically, then, to matter little whether CFS3 was written to support SLI or not.

It almost seems as if the issue(s) with CFS3 were/are of such a nature as to be incapable of correction even *with* the benefit of SLI. And, as I said earlier - I do believe one big reason for the lack of acceptance of CFS3 was graphics issues.

Don't know if this makes sense or not.
 
Cliff I think you are over investigating it, P2 is MUCH more resource hungry than CFS3. Slider 3 is slider 5 in CFS3 for a start. We have much more going on, more objects, more textures, more aircraft, more effects etc.

It's a fact, my 9800 GX2 ran like a DOG compared to my 8800 GTX with CFS3. But flew with modern games. Why I have no idea. I tried everything under the sun. Popped in a single GPU 1GB and wham much better.

Welcome to PC land. So many variables between game configs let alone the software installed, hardware settings, bugs issues, BIOSs etc. One reason consoles are popular. May not have the latest sharpest sim, but works out of the box until they become older.
More modern games and 3d benchmark tools (often look ahead to use features "coming soon" in the next Direct 3D) know how to use more features and more modern Direct X. There are too many variables you are leaving out of your assumptions.

I have never used Aircraft slider with anything less than 5 for the last 4 years so that is not the issue something else is.

You will always see triangles if you pan around. It's the engine. It can be lessened with sliders and more hardware yes but it's because CFS3 is not drawing a huge tile that only has say 2% in view where you see a large area from almost edge on mainly out of view, and only see the very very corner. The higher res the terrain like OFF the worse it can be - TIR can also make it worse if you have older drivers.

Try reducing the flight density to normal, try turning down AA as each notch up eats more memory (as you have have to render at a higher res than you see on screen). Yes others have a much better experience and fps than you report with similar setup, but it does sometimes need to go back to defaults first - clear out.

Try cfsconfig reset to default and slowly add any changes back in (backup your \app..data\....\Microsoft\CFSWW1 Over Flanders Fields\configoverrides.xml first).

OFF needs GPU memory of course, but better CPU Mhz is more useful (dual core etc does not help).

Other than that no idea.


P3 has a new high res tile set but optimisations in other areas so more eye candy with little difference in fps.
 
Ok, I think I'm one of the Lastof the Mohicans here, so I'll throw in my two cents, or p if you will :d.
I run Intel PIV 3.2 GHZ, 2G ram, XP Pro SP2, ATI Radeon HD3800 AGP w/ 512MB and TrackIr.
I run at 5,3,3,5,5 (seems to run better with higher clouds). I get frame rates in the 30's, which, on my rig, is greatly smooth and acceptable to me. If I turn my head very quickly at these settings esp. around airfields I will get the triangle for a split second in the edge of my vision. I understand the limitations of my rig and the short-comings of cfs3 so I deal with it. If I change my terrain settings to 4, I get the white artifacts in the distance, trees, bldgs, etc. I can see that the graphics card is being pushed too hard as I can see a noticeable increase in trees in the area going from 3 to 4 and for longer distances, so it has to draw millions more each frame.

I would like to upgrade to Nvidia and a faster GPU but that requires PCIe and a motherboard and probably CPU upgrade which would put a severe strain on the cash flow esp. with Christmas coming.

Now, the good news, and here's where my two cents get more valuable I think. Due t what's been posted by Pol and OVS, I am going to wait and see what my computer will do with P3 for the following reasons.

1: I understand that the team has been able to delve deeper into the cfs3 code to effect changes with the AI and other things that were limiting factors in P2 performance.

2: WM has been able to design the scenery to look better with less demand on resources.

3: The screenies OVS posted were on lower settings, so it appears that P3 scenery on lower settings like mine will look lightyears better than P2 scenery on 4 or even 5.

In conclusion, my advice would be to wait as I am going to do because I get the general impression that P3 will be much better performing than P2 and will possibly meet or exceed my expectaions on what I already have, or, at least keep me satisfied til I can afford the new components I desire.

Hope this helps a little.
 
Certainly your benchmark tests will show much better with SLI enabled, as you have twice the GPU processing power. Your results are exactly as would be expected, but I am not sure this has much to do with your OFF problems.

The problems you have been experiencing with OFF have been attributed to a lack of texture memory (per Winder's comments above), and since SLI apparently only uses the texture memory of one of the cards, this is going to be unaffected by enabling SLI. Yes, the SLI cards can process the graphics images very quickly, but if there is insufficient room for all the textures need by the game to fit into memory, then I assume there will be problems on-screen.
 
You must log in or register to reply here.
Back
Top