3 hours ago, Cpt_Piett said:

Nope, it's correct. That's all I can squeeze out of my rig on ultra settings with iniBuild's EGLL and AI traffic. It's the same at iniBuild's KLAX, and at other major hubs. Of course I can turn down the settings and get higher fps. But the purpose of the comparison was to use the highest settings in the most demanding scenario. 

And perhaps it would've been better not to quote my whole post 🙂 

That does seem rather low (dipping below 30 fps) given your hardware.  But then again I don't have Inibuilds EGLL.  But I do run at 4K with all the bells and whistles with the PMDG 737-700.  It's also possible you have your AI traffic turned up pretty high.

I've heard Inibuilds EGLL is hard on performance.  

15 minutes ago, Mace said:

I've heard Inibuilds EGLL is hard on performance. 

It is. It's possible I could squeeze more fps by tuning things, I'm not sure. As @GSalden says these hubs are performance killers. 

As this has been brought up I'm getting curious. Since I've saved the flight I'll do another comparison video with just the PMDG in community. 

I'm hoping that someone with a 13900 or a 5800X3D would join in on repeating the flight to compare performance. 

Edited November 6, 20223 yr by Cpt_Piett

5 hours ago, Cpt_Piett said:

But then we should agree on some common settings

and not use any variables that are hard to reproduce 100% like:

• in the DLSS2 video I used ReShade with a LumaSharpen shader among others. Hence, the image looks sharper. 
• in the DLSS2 video I used FSLTL with the injector, whereas in the DLSS3 video I used real-time online traffic with FSLTL base models

7 minutes ago, turbomax said:

and not use any variables that are hard to reproduce 100% like:

• in the DLSS2 video I used ReShade with a LumaSharpen shader among others. Hence, the image looks sharper. 
• in the DLSS2 video I used FSLTL with the injector, whereas in the DLSS3 video I used real-time online traffic with FSLTL base models

I concur. Should be pretty easy agreeing on some common settings. 

Further, I realize you want to tax your system in this test, but using add-ons that others don't necessarily own makes it difficult to make valid comparisons. What about only default scenery and aircraft? Pick a challenging airport or location like KDFW or KJFK? Fenix or PMDG probably okay since everybody owns them 😆, or maybe FBW A320 since it's kind of a hungry beast too. 

1 hour ago, Cpt_Piett said:

It is. It's possible I could squeeze more fps by tuning things, I'm not sure. As @GSalden says these hubs are performance killers. 

 

30 minutes ago, odourboy said:

What about only default scenery and aircraft? Pick a challenging airport or location like KDFW or KJFK? Fenix or PMDG

when I say default I mean default! 😀

also: should post screenshot with all other relevant ingame settings that unfortunately are not included in a saved .FLT file, like your preferences (graphics, traffic etc) either ALL 0 or ALL 100% max. this is how it was done in the old Bob Ainscough/P3D days.

1 hour ago, odourboy said:

Further, I realize you want to tax your system in this test, but using add-ons that others don't necessarily own makes it difficult to make valid comparisons. What about only default scenery and aircraft? Pick a challenging airport or location like KDFW or KJFK? Fenix or PMDG probably okay since everybody owns them 😆, or maybe FBW A320 since it's kind of a hungry beast too. 

Good suggestions. Although I didn't make the videos/benchmarks with comparisons with other systems in mind. For that I have to re-do the flights with some agreed-upon settings. The addons are however the same in the videos I used. 

I did another SU11 flight with DLSS2 with only the PMDG in the community folder.

Well, almost as I forgot about AccuSeason and FSDreamteam KORD. 

These are the CapFrameX benchmark results. 

Average fps 43 vs 32 with the addons. 1% low 26 vs 22 with the addons. I'll upload the video once it's finished processing. Spoiler: only around 30fps on landing. 

The video:

 

Edited November 6, 20223 yr by Cpt_Piett

21 minutes ago, GSalden said:

 

This is what I'm talking about. Even the best CPUs struggle with everything maxed out + AI traffic at major hubs. 

Edited November 6, 20223 yr by Cpt_Piett

13 minutes ago, Cpt_Piett said:

Even the best CPUs struggle with everything maxed out + AI traffic at major hubs.

so it has always been, and so it shall be as long as flightsimulator lives.

difference though: 30 fps today looks so good you can even show to your gaming friends (Cyberpunk, Red Dead Redemption) without being too embarrassed anymore. 

37 minutes ago, Cpt_Piett said:

This is what I'm talking about. Even the best CPUs struggle with everything maxed out + AI traffic at major hubs. 

With the 12900 + 3090 and testing at EHAM with PSXT Live Traffic I used 25 parked ac  + Live ac and was able to stay above 30 fps ( used Vsync / 30 hertz )

Then with the 13900 + 3090 I could set the amount of parked ac to 40.

Now with the 13900 + 4090 I can set it to 46 to stay above 30. I use DLSS / 50% Vsync / 60 hertz.

The issue is not the hardware. The issue is me who always tries too put the maximums amount of detail and realism into the sim….

 

1 hour ago, GSalden said:

The issue is not the hardware. The issue is me who always tries too put the maximums amount of detail and realism into the sim….

Lol, I'm 100% the same. 

2 hours ago, GSalden said:

With the 12900 + 3090 and testing at EHAM with PSXT Live Traffic I used 25 parked ac  + Live ac and was able to stay above 30 fps ( used Vsync / 30 hertz )

Then with the 13900 + 3090 I could set the amount of parked ac to 40.

Now with the 13900 + 4090 I can set it to 46 to stay above 30. I use DLSS / 50% Vsync / 60 hertz.

The issue is not the hardware. The issue is me who always tries too put the maximums amount of detail and realism into the sim….

 

Perhaps if you had a 3800X3D you could run 60 parked AI 😈😄  (ducks, and runs...)

6 hours ago, Mace said:

Perhaps if you had a 3800X3D you could run 60 parked AI 😈😄  (ducks, and runs...)

Sure …. 😂 😂 😂

Fs98 is the solution : guaranteed 100+ fps in every situation…..

Edited November 7, 20223 yr by GSalden

On 10/30/2022 at 8:40 AM, TravelRunner404 said:

Tom’s hardware shows the 5800x3D to be about 25% faster than the 13900k. There is a giant thread at the msfs forums of people upgrading to it and sharing their impressive results. The 13900k is a fine processor for a high end computer but if you want MSFS performance it’s slower.   In the MSFS forums we were positing results that are 15-25 Frames better than 12900k at busy airports.

Nahhh, I just came from a 5800X3D and upgraded to 13900K and Z790 board. The 13900K beats the 5800X3D hands down. Now, this is on 4K. I have never used 1080P on MSFS or any of my other games. Maybe in 1080p the 5800X3D does better, but once you throw in 4K in there, 13900K is king. I’m talking about 10 more frames in average and defined a lot smoother fps than 5800X3D. 
 

All this benchmarks showing 1080P has it’s purpose, but let’s be real, unless you are a pro gamer, who in the world buys a high end CPU to game at 1080P. Especially on MSFS.