Took me a while, but was finally able to find a GTX 1080 FE available for sale.  Thought I'd share results from my test PC comparing going from a 970 to a 1080.  This is specific to P3D.

 

970 Medium            FPS 43, 31, 28  - 2560 x 1440  

Fraps and Graphics Settings 

 

970 Extreme           FPS 30, 21, 19  - 2560 x 1440  

Fraps and Graphics Settings 

 

1080 Medium           FPS 45, 33, 30  - 2560 x 1440  

Fraps and Graphics Settings 

 

1080 Extreme          FPS 34, 24, 22  - 2560 x 1440

Fraps and Graphics Settings

  

1080 Medium WideView  FPS 37, 28, 22  - 7680 x 1440

Fraps and Graphics Settings 

 

1080 Extreme WideView FPS 28, 20, 19  - 7680 x 1440

Fraps and Graphics Settings 

 

Was not able to run my GTX 970 at 7680 x 1440 (Widview) so couldn't compare.

 

Flight Files

 

Summary

GTX 1080 FE at Medium graphics 4.6% increase over 970

GTX 1080 FE at Extreme graphics 13.3% increase over 970

GTX 1080 FE going from 2560 x 1440 to 7680 x 1440 17.7% decrease in FPS

 

My Test PC Specs

Windows 10 64bit

FlightSim: P3D V3.3 (no add-ons)

CPU: 3960X @4.2 GHz 2133Mhz 32GB RAM

SSD: Samsung 2X EVO 840 1TB

Monitors: 3X Acer K272HUL - DisplayPort 

GPU: Asus 970 Strix, nVidia 1080GTX FE

 

Note:

1920 x 1080 (baseline) =  2,073,600 pixels (HD)

2560 x 1440            =  3,686,400 pixels - 78% increase in pixels

3840 x 2160            =  8,294,400 pixels - 300% increase in pixels (4K)

7680 x 1440            = 11,059,200 pixels - 433% increase in pixels

 

Since I have no means to direct record 7680 x 1440, I used one of my video cameras for external:

 

 

Note 2: 

This is my test PC not my main FlightSim PC.  I'm not a fan of 3 monitor setups (bezels) but received several requests for me to run a 3 monitor test on these two GPUs (970/1080), so I finally got around to it.

 

Cheers, Rob.

So in a nut shell no way your going to be replacing you Titan X's - great post very informative Rob

Correct, there isn't enough in the GTX 1080 (at least in terms of P3D anyway and extrapolating from 970 to 1080) that would warrant me to replace the Titan X's ... I'll wait until 2017 for next gen nVidia cards.  

 

Cheers, Rob.

Thanks for this Rob

Been considering a change of Card and looking at your results think I might wait and see if the price on 980Ti drops a bit and go with that. Cheers Pete

Great idea takes some years to pass for the next wow video card to come out but a good choice for those who have not invested in a Titan X  - and test above was not even against a Titan so those numbers would probably the same or even a bit worse compared to the mighty Titan 

Isn't this mostly just showing how cpu bound the program is?

Isn't this mostly just showing how cpu bound the program is?

 

That's exactly what I was going to say. It's really showing that P3D depends more on the CPU once GPU power goes beyond a certain point, for typical resolutions anyway. In other benchmarks, the 1080 is wayyyy beyond a 970.

 

It's still a useful test though because it does show exactly what one can expect if P3D performance is their top priority.

Thanks for the post Rob. Exceptional information as always. I thought that the 1xxx series cards were going to be a leap in tech and performance rather than just a marginal increase. Do you know more about that?

Thanks for the post Rob. Exceptional information as always. I thought that the 1xxx series cards were going to be a leap in tech and performance rather than just a marginal increase. Do you know more about that?

 

See above. It's a sign of P3D's reliance on CPU power as well as GPU power. In typical gaming (games like Battlefield, Witcher 3, and so on) the performance delta is far greater between the 9xx series and 1xxx series.

See above. It's a sign of P3D's reliance on CPU power as well as GPU power. In typical gaming (games like Battlefield, Witcher 3, and so on) the performance delta is far greater between the 9xx series and 1xxx series.

Well that is good to hear. Hopefully either Dovetail FS or P3D v4 will truly depend wayyyy more on the GPU, which I would imagine would require a ton of different coding.

Hello Rob, I moved from a GTX 770 to 1080 and its a huge improvement of course.

But in some airports my cpu now limits like latin vfr kmia or imaginesim atlanta. In other big airports its running well my cpu is a i5 4670K at 4,3ghz. I am not planning to upgrade my cpu so:

I want ask you what setting is more cpu depended? I know you already mentioned this somewhere but I cant find that thread.

 

Thank you.

 

Edit: @MikeT707: yeah i can confirm that. I am able to play GTA 5 in 4k with max settings. I am getting 60fps evertime.

Too bad you couldn't compare the GTX970 to the GTX1080 in Wideview, because this is where I would have expected a more significant step-up in performance. I'm currently running a triple monitor setup on a GTX780 and thinking about upgrading to a GTX1070 which should give some improvement at least in dense clouds. I will report back if I do the upgrade.

Thanks for the test Rob. Informative. Sticking with my 980ti for now....

Nice, thanks. Such a shame more GPU is not used.

Wow, thanks Rob, that is interesting .

 

I am waiting since weeks for my 1080. I might change my order to a 1070 that is available now and spend the 200 bucks elsewhere.

 

I am not sure if you published the results on a single monitor already? What is the performance there, or is that compare able with above results?