HOW one Nvidia card can perform awesome FPS , while other is terrible,in same system?

In the beginning I had a mysterious "very low FPS performance" problem in FSX (it was always around 6-7/sec average) and I tried many things to solve it.

I played very much with "fsx.cfg" file and other settings, I worked on BIOS and hardware configuration, drivers, CPU settings..etc. Finally I could not find even any little clue about the reason, and I could not be able to increase FPS even a little.

But after I attached a second NVidia card to the same system (which has lower capacity than my current card), I see with this card I can reach to 40/s FPS without making any special settings.

So I understood nothing was about "fsx.cfg" file or other setting or other hardware configuration. It was only about my display card!

My first display card was NVIDIA GeForce GTX 950

The second card I attached was NVIDIA GeForce GTX 580

Now I am connecting them to 2 monitors at the same time (one card connected to one monitor), and shifting FSX windows from one monitor to another.

When I move the FSX window to "GTX 580" monitor FPS is around 40/sec, but when I move the window inside other "GTX 950" monitor FPS is immediately decreasing under 10/sec. (by the way in this monitor always scenery spikes, flashes happen too)

I recorded this situation in a video

How this can happen?

What can be wrong with the card GTX 950?

But I must say that my NVIDIA GeForce GTX 950 has no problem in other applications. It is running perfect... Problem is only related with FSX.

For example with the same card I can play Grand Thief Auto 5 very smooth at high performance levels.

So what can be the cause of this NVIDIA GeForce GTX 950 & FSX cannot match each other?

Very interesting ...

Probably due to primary versus secondary card choice.

You don't normally run different monitors off of different cards on the same computer.

You try 10 different video cards with the SAME monitor. All 10 of them work at 40 fps.

Then you try 1 video card with 2 different monitors. One monitor gives you 40 fps. The other monitor gives you 6 fps.

The monitor that gives you 40 fps has a native resolution of ........... and FSX is set to a resolution of ............ . (Both the same).

The monitor that gives you 6 fps has a native resolution of ........... and FSX is set to a resolution of ........... . (Not the same).

Any questions as to what is THE DIFFERENCE?

Chuck B

Napamule

Wow, could it be that much difference in fps (75% drop) because of native resolution settings!

Sent from my iPhone using Tapatalk

No! It's not the res of either. It's the NATIVE (ie: max res) of the monitor that is involved. Not all monitors (even if the look exactly alike) can handle higher res if their native res is 800x 400 and you are feeding it 2048x1024 video. The monitor has it's own electronics. It is not a 'boob tube' like old CRT used to be. Some monitors can do it dynamically when they are first hooked up to PC. Some you have to go into the monitor's MENU and CHOOSE the res you want/need. It is then that you find out IF it can/can't 'process' higher res DATA from PC video card. It's that simple.

Chuck B

Napamule

Maybe the op should list all his specs for once. Because now he's letting things out.

First if mentioned a pc and a gtx1060. Now there's another card. And also god knows how many monitors (with various different resolutions).

Without giving all the info trying to help makes little sense.

Napa is right that having fsx scale up, then having win/card/monitor scale down, makes on sense.

More hardware? List it all. For example, what's in the other PciE-x16 slots? I saw there are 8 (!) of those when I looked up your mainboard to see how your Ram should be installed. Again, don't use two Ram sets that aren't matched. It's worse then using just one. And use one videocard at a time if you can. You can attach 4 monitors to a card. Using 2 cards (or more) won't help fsx. The opposite actually.

Probably due to primary versus secondary card choice.

You don't normally run different monitors off of different cards on the same computer.

His problem is that the 950 wasn't performing well in FSX while other games ran fine. The second card was added later and seems to run the sim fine. Something certainly isn't right with the 950 and FSX combination.

Where does the OP mention a 1060? Someone else does, but the OP only lists the 950 and 580 GPUs. Also, the OP isn't trying to run two cards with the second one was added as part of the troubleshooting.

Again, don't use two Ram sets that aren't matched. It's worse then using just one.

Why not? Unless you are trying to overclock and the specs vary considerably between the two sets, there shouldn't be any problems with more than two sticks of RAM. The system will automatically pick the settings that both sets are capable of and use those. If there is a bad stick of RAM, it wouldn't matter if the system was running 2 or 4 sticks of RAM.

Napa is right that having fsx scale up, then having win/card/monitor scale down, makes on sense.

Yes, it is best to run the sim at the native resolution of the monitor, however, any scaling the monitor does will only affect image quality and have zero impact on performance. FSX will only care about the resolution in its settings. The higher the resolution the more demanding it will be on the CPU and GPU.

While the 950 and 580 may perform similarly in many cases, the 580 does have more memory bandwidth and will handle higher resolutions better.

Clearly listing all specs is always a good idea.

yes, I thought it was 1060 instead of 950.

If the pc was a prebuilt that came with 4 gb and a 2XXX cpu, adding 2x8 may not be compatible. (remember HP ram anyone?)

Even if compatible, mixing and matching a slow 4gb with a fast 16gb is not that benificial. It will run at low speed of the 4gb. Having just 16 running at its full speed works better.

And setting 2100x something in fsx, when your monitor is only 800x600, is not a great idea.

Anyway, I'll stop guessing until he lists some better details.

Just have one more. Connect just one gpu, one monitor, then install its driver, and get a baseline fps of a normal setup. Then add more stuff, to see what causes the issue. Would probably be the simplest approach.

Because in some reviews they are similar.

http://www.anandtech.com/bench/product/1444?vs=1670

Nothing to do with a higher number at all.

Because in some reviews they are similar.

 

http://www.anandtech.com/bench/product/1444?vs=1670

 

Nothing to do with a higher number at all.

My performance decreased in both FS9 and FSX when I installed on another machine with a faster (but less cores) CPU and a gpu with double the memory

But I think both those platforms, along with all the background tasks, actually prefer a CPU with more cores. Totally against popular opinion.

My 'lesser' GPU performed better. Quicker texture loading. And better fps. On paper the newer GPU looks as good or better, I base my perfomance on the texture and pixel fill rate for the cards which was better on the old card.

All my previous cards I have done the same thing and got better performance based solely on this factor.

After looking at your post I checked both cards out. And no surprise the 580 had basically double the pixel/texture fill rate of the 950.

But each their own...

Thank for sharing your comments everybody..

My first thread was

https://www.flightsim.com/vbfs/showthread.php?307968-I-have-Very-Interesting-and-Mysterious-quot-LOW-FPS-PERFORMANCE-problem-quot&p=2030289#post2030289

(it was before I found out the reason of the problem is only Display Card)

In that thread we discussed about system hardware details too...MY SYSTEM :

- Intel Core i7-2600K CPU 3.40GHz

- Mainboard Big-Bang Marshal (MS-7670)

- RAM 20 GB 1333 MHx

- NVIDIA Geforce GTX 950

- Windows 7 x64

In last days, I asked this situation to many people. Most of the people could not give a very clear answer. But one expert shared this below;

"Reason is obvious, FSX needs memory band width and memory interface width and for the 950 this is way to low for FSX ( 105 and 128 bit ) where as the 580 has 192 and 384 bit, a "NVidia GTX..50" card is to low end for FSX unless it is a TI"

I don't have deep knowledge about hardware working details, but as I understand, hardware structure of this card (about memory using) is not very suitable for FSX.

If I am not wrong, after all, I came to a final conclusion as "For a good FSX performance I must change my display card Nvidia GTX 950..This seems the only solution."

I don't have deep knowledge about hardware working details, but as I understand, hardware structure of this card (about memory using) is not very suitable for FSX.

When you compare the performance of the 950 and the 580, they are very, very close despite the significant disparity in memory bandwidth. There are plenty of people who have run FSX quite well with x50 and x60 graphics cards - your performance delta is more likely based on something on your specific system rather than the card architecture in general.

Cheers!

Luke

There are plenty of people who have run FSX quite well with x50 and x60 graphics cards - your performance delta is more likely based on something on your specific system

Luke

You may be right, but I cannot have any idea about what can be the different thing in my system which can cause this...

My mainboard has no low capacity... The only the possible reason I can think was; I have 4 memory cards (1 pair of 2GB, and 1 pair of 8GB) can cause a problem.

So I totally removed both 2 GB cards, and only two 8GB (total 16 GB) cards remained in slot 0 and slot1 (both RAMs are identically same as brand and model). But this did not make any difference also.

Finally, I had to come to a point to believe that "this card and FSX are not matching each other".

You may be right, but I cannot have any idea about what can be the different thing in my system which can cause this...

My mainboard has no low capacity... The only the possible reason I can think was; I have 4 memory cards (1 pair of 2GB, and 1 pair of 8GB) can cause a problem.

 

So I totally removed both 2 GB cards, and only two 8GB (total 16 GB) cards remained in slot 0 and slot1 (both RAMs are identically same as brand and model). But this did not make any difference also.

 

Finally, I had to come to a point to believe that "this card and FSX are not matching each other".

Your RAM configuration shouldn't be causing this, or your other games would also be impacted. BIOS settings also shouldn't be an issue.

Have you tried doing a clean reinstall of the Nvidia drivers and deleting the FSX.cfg file (lost track of what has been tried between this and the other thread)?

Here's something else to try.

https://www.google.ca/amp/s/fsxtimes.wordpress.com/2011/07/11/clean-up-shader-cache/amp/

Have you tried doing a clean reinstall of the Nvidia drivers and deleting the FSX.cfg file (lost track of what has been tried between this and the other thread)?

Yes ...As I tried many things until now, I tried that also...

I recorded a video with Nvidia Inspector live parameters at flight..Maybe it can give more clue about the real reason of this problem..

Here is the link:

I made two videos with two cards with Nvidia Inspector

I cannot explain the performance difference with CPU, or with other hardware configuration, or settings of FSX..

Because both display cards are on same system, using same CPU, same system configuration, same FSX settings...One of them has terrible performance, while other has very acceptable performance...

I could not find any explanation about this difference yet ??

And nobody could find an explanation yet ...

You tried one advice, but then undid it. Tried another, then undid that also.

When installing the new card you need to do them all. New driver, new fsx.cfg, no inspector settings, definitely not higher inspector settings. And, all using just one monitor.

What are you doing? Getting your new card installed properly? Or comparing situations that you don't describe well? For example, you still have not said what monitor(s) you have attached, what their resolutions are, what other pciE devices are in the system, what other hardware and controller you have attached, etc, etc.

I suggest you get focussed on getting it working, instead of on asking "why?".

Hi Damlays -

I'm not sure if I understood correctly, but I concluded that you have both video cards in the same machine at the same time? If so, is there a priority setting, either in NVidia, or on Windows screen setup that has to be resolved ? For example - is one video card preferred over the other in the system based on a config setting - for that matter is it in the Bios?

If this is not the case - how is the heat factor on each of these video cars? In other words does one run hotter that the other ?

You tried one advice, but then undid it. Tried another, then undid that also.

When installing the new card you need to do them all. New driver, new fsx.cfg, no inspector settings, definitely not higher inspector settings. And, all using just one monitor.

 

What are you doing? Getting your new card installed properly? Or comparing situations that you don't describe well? For example, you still have not said what monitor(s) you have attached, what their resolutions are, what other pciE devices are in the system, what other hardware and controller you have attached, etc, etc.

 

I suggest you get focussed on getting it working, instead of on asking "why?".

Thanks for your reply, but sorry I cannot agree your comments;

I dont understand which advice I undid..I tried all..and of course I am trying many things to find a solution....I tried more many things with fresh install without any settings, with some settings, with more settings..etc..etc..

My monitor resolution is 1920 x 1080, FSX resolution is 1024x768

I tried in many different monitors also... With my videos I am making it visible very clear about "what is the difference in both cards while all system parameters were same".. Plus I am showing "live graphic card" parameters for people who may interpret these data with their hardware expert knowledge...

In one try monitor was Asus mx279h, in another Acer H236HL, in another Samsung 55k5300 Television..all connected by HDMI

So if the both cards are tested in the same monitor one by one and always all results show the same difference between cards, why it is important "which monitor" ? Isn't it obvious this is not a monitor related problem?

I wrote my system configuration already before...There is no other PCI devices attached (if attached I would write them too).. My system is simply designed with basic elements for high performance as much as possible..it is not full of many hardwares attached...

so what is "etc..etc" I can say about my hardware? I could not understand this too (I am not an hardware expert ..If I am, then I would not post here for help, I would just do the tests by myself alone)..So if only somebody asks me exact questions about my system, I can try to answer them....isn't this understandable?

And, I think it is also clear to see "what am I doing"...

I am showing all these parameters and just asking "do you have an idea what can explain this big difference between both cards in the same system"?

(Asking "why" is for making it "working"..if we cannot diagnose, we cannot cure!)

If still "no idea", then thanks for your attention in anyway..

Hi Damlays -

 

I'm not sure if I understood correctly, but I concluded that you have both video cards in the same machine at the same time? If so, is there a priority setting, either in NVidia, or on Windows screen setup that has to be resolved ? For example - is one video card preferred over the other in the system based on a config setting - for that matter is it in the Bios?

 

If this is not the case - how is the heat factor on each of these video cars? In other words does one run hotter that the other ?

Hi okdaley,

Thanks, this was a different and interesting point I didn't hear from anybody before.. After reading your comment I checked BIOS but I could not find any settings similar as you mentioned..

I will check about heat factor.

I also get the impression you have 2 cards installed. The 580 and the 950.

Also, I don;t see why you want to "explain the difference". I think you should focus on getting the 950 working. Not on comparing two cards.

you had one 850

you now compare it to an 850 + a 950

you should instead just install only the 950.

And, after installing a new card (the 950) you should:

attach only one monitor

install the card driver,

rebuild the fsx.cfg to restet certain settings

make sure in fsx settings the resolution is the same as the native resolution of the monitor that is attached.

leave the NVinspector alone for a while. Whatever you do, do not use the pre-saved settingsd you were using with the other card.

NV inspector is a lot of trial and error to get set up right. So no shortcuts. And only start trial and error when you know the video basics are working correctly.

Those are always the required steps after installing a new card.

Then test the result for fsx.

Focus not just on fps, but also on smoothness, and graphics quality.

And be aware, newer card-->

higher card menu graphics settings possible-->

result of setting those that high is lower fps, so maybe don't set those to maximum.

And what I mean by focussing on the real issue is that, if the above approach does not work, you should not immediately undo all changes, but stick with it. All the way untill you do get it set up fully and working.

Btw, a large monitor--less fps

3 monitors--less fps then one monitor

2 videocards at once--good chance of driver issues

Just swapping monitors, cards, etc, in the hope you'll "get lucky" sometime, is not going to work.