MSFS 2020 Auto FPS App v0.4.2

[Harrier77]

Hi, I would just like to say that I have been using MSFS2020 AutoFPS since the beginning, and I am now on 0.4.3.1 test, and it is working flawlessly. I am not using expert mode and I am not using Auto Target FPS.  I have 30 FPS set as the target (I have a modest system running in 4k). I fly both airliners and GA, plus some plane spotting with PSXT, and I use for VFR and IFR modes as needed. Your program has been a 'life saver' for me which is allowing me to continue using my current system for a little longer before the 'big upgrade' when I have saved up the money. Thank you.

EDIT: I also use DLSS Ultra Quality mode set at .8 which is working very well.

Edited March 19, 2024 by Harrier77
Further info

[Cpt_Piett]

I've completed yet another rigorous test session, and to be honest I was surprised by what I found. Video first, then my observations.

Observations:

• TLOD has a huge impact on how PG-buildings are being rendered.
• TLOD affects terrain draw distance, PG level of detail as well as PG draw distance
• OLOD does not affect PG whatsoever

Rendering PG-buildings is a very CPU-intensive task. This huge increase in PG level of detail increased the load significantly on almost all of the 12900KF's 16 P-core threads, increasing temperatures (occasionally hitting TjMax at 100 degrees celcius) and fan noise. After the rendering was complete though, the CPU load settled on a much lower level.

The GPU load was not significantly affected by this. However, the impact on VRAM usage was massive, from 12.5GB to 20.5GB.

Next one up is OLOD 🤓

Btw I'm finally on 0.4.3.1-test. 

Edited March 19, 2024 by Cpt_Piett

[Harrier77]

7 minutes ago, Cpt_Piett said:

Rendering PG-buildings is a very CPU-intensive task. This huge increase in PG level of detail increased the load on almost all of the 12900KF's 16 P-core threads work at maximum speed, increasing temperatures (occasionally hitting TjMax at 100 degrees celcius) and fan noise. After the rendering was complete though, the CPU load settled on a much lower level. The GPU load was not significantly affected by this. However, the impact on VRAM usage was massive, from 12.5GB to 20.5GB.

I can only dream of a system that is capable of such things 🙄 Maybe one day!

[Cpt_Piett]

24 minutes ago, Harrier77 said:

I can only dream of a system that is capable of such things 🙄 Maybe one day!

I can understand it could frustrating to be constantly subjected to my ridiculous high-end tests. Not thinking about you specifically, but I've had some reactions - especially after being one of the first on Avsim MSFS to share how things changed after I got the 4090. 

I will say this though: this insane test put a load on my rig akin to some 3rd party torture tests, and the case fans sounded like an A300 taking off! Not only that - twice did MSFS CTD with the "your graphics card encountered a problem" error message. And twice - as I got too ambitious with TLODs and OLODs, Windows became irrevocably unresponsive, requiring a forced reboot 🤪 I really hope I didn't fry my CPU doing this!

Edited March 19, 2024 by Cpt_Piett

[Adamski_NZ]

Apologies if this has already been covered in either of these two busy threads ... but ...

What's the procedure for taking AGL into account? For instance - I often do NZAA (Auckland) - NZQN (Queenstown) ... or from NZCH or NZWN. All those departure airports are pretty well at sea level. When arriving at NZQN (1171 MSL) - or at any airport significantly higher than departure - what should I be doing?

Can AutoFPS be set to automatically use AGL?

[Cpt_Piett]

Last test for now: OLOD and objects + TLOD and airport textures

Compared to my videos regarding the influence TLOD has on terrain and photogrammetry, the difference here is very subtle (at least to my eyes), even though I was switching between OLOD 10 and 1000. The only difference I could see is that the number of objects increased somewhat at 1000. I didn't see any significant impact on CPU or GPU load.

Previously I suspected that OLOD heavily influenced the number of buildings, but as seen in the video, this is not the case. Both number of buildings as well as building detail level didn’t seem to change at all. That's the reason I went to Rio by the way, as the area has a huge number of (non-PG) buildings. 

I did take the opportunity to show the effect TLOD has on airport textures though, as I'm sure this will be relevant to most simmers. Notice the change in detail level especially on the runways.

Also, I feel like I should apologize for hijacking this thread with these types of tests. I do think it's relevant to AutoFPS though, and I sincerely hope that someone finds it useful - as I might have shortened my CPU lifespan significantly in the process 😅 

Edited March 19, 2024 by Cpt_Piett

[Cpt_Piett]

5 minutes ago, Adamski_NZ said:

Apologies if this has already been covered in either of these two busy threads ... but ...

What's the procedure for taking AGL into account? For instance - I often do NZAA (Auckland) - NZQN (Queenstown) ... or from NZCH or NZWN. All those departure airports are pretty well at sea level. When arriving at NZQN (1171 MSL) - or at any airport significantly higher than departure - what should I be doing?

Can AutoFPS be set to automatically use AGL?

Since I'm (still) here, I might as well try to answer this. I often fly in NZ - the scenery is simply too good. The way I see it is that you don't have to take anything into account, as TLOD and OLOD changes are based on AGL. I recently departed from NZQN and TLOD and OLOD reacted exactly the same way as when I landed in NZNS. 

[Cpt_Piett]

3 hours ago, Cpt_Piett said:

Last test for now

...but - I never stop testing, do I? 😅

As the Lofoten archipelago is one of my favourite places in the world (been there several times IRL, I thought it would have been unfair not to subject it to my rigorous testing. I can't think of a better place to appreciate the effect TLOD has on how detailed the terrain is in the distance.

From the camera location just northeast of ENSH, the aerial distance to the southwestern tip of the islands (just visible in the far distance) is 54nm (100km).

I'll let the viewers judge what would be a reasonable TLOD setting here. Or, rather than using a fixed setting, we can let AutoFPS change it for us dynamically. I see no reason not to set TLOD to maximum as AutoFPS will only increase it as long as there's headroom for it. 

I realized something new: TLOD not only affects terrain textures, but also mesh resolution. Which does matter here if one wants to fully appreciate the very high resolution digital elevation model (DEM) introduced in SUXV. According to this post (not an official source btw), it’s 1m for mainland Norway. Even if that’s incorrect, it’s very hi-res.

By the way, can anyone spot the aircraft? That's my 787 at 5000ft 😉

Edited March 19, 2024 by Cpt_Piett

[Pochi]

Regarding OLOD my understanding is that it increases the level of detail of Objects which is noticeable mainly on ground or close to it. Logic would be to increase it when the aircraft closes to ground while TLOD does the reverse.

Pierre

[GSalden]

12 minutes ago, Pochi said:

Regarding OLOD my understanding is that it increases the level of detail of Objects which is noticeable mainly on ground or close to it. Logic would be to increase it when the aircraft closes to ground while TLOD does the reverse.

Pierre

Custom objects from 3rd party developers like sky scrapers in Dubai react to the OLOD setting.

Mysejf I do not like a skyscraper appear out of nothing when I am just a few mls away ….in the app I have auto OLOD unchecked.

But its everybody’s choice.

[Noel]

6 hours ago, Cpt_Piett said:

Rendering PG-buildings is a very CPU-intensive task. This huge increase in PG level of detail increased the load significantly on almost all of the 12900KF's 16 P-core threads, increasing temperatures (occasionally hitting TjMax at 100 degrees celcius) and fan noise. After the rendering was complete though, the CPU load settled on a much lower level.

This must imply if one turns PG off as I have done recently I should be able to get away with much higher TLOD levels than I'm set at, yes? 

[Cpt_Piett]

@Pochi Yeah does seem that way. In the screenshots thread I posted some comparison images of a top-down view of JFK.

@GSalden I didn’t know that. But as I saw how little overall influence OLOD had in Rio with the metro airport, I really don’t think OLOD has much of a performance impact. Perhaps at those largest hubs.and like you say, it’s everybody’s choice.

@Noel yes definitely I would say so, based on my recent findings, as high TLOD in PG-cities absolutely wreaks havoc on the poor mainthread. I mean when overflying Manhattan, it has to render hundreds of buildings, if not thousands. The whole NY greater metropolitan area is just a huge PG-zoo, as can be seen in my video. Rendering all those buildings within a short time frame brought my 12900KF almost to its knees, caused several MSFSs CTDs, as well as two Windows freezes. PG is basically a CPU-killer! And after killing your CPU, it eats all your VRAM…

Edited March 19, 2024 by Cpt_Piett

[RobJC]

7 hours ago, Cpt_Piett said:

I've completed yet another rigorous test session, and to be honest I was surprised by what I found. Video first, then my observations.

Observations:

• TLOD has a huge impact on how PG-buildings are being rendered.
• TLOD affects terrain draw distance, PG level of detail as well as PG draw distance
• OLOD does not affect PG whatsoever

Rendering PG-buildings is a very CPU-intensive task. This huge increase in PG level of detail increased the load significantly on almost all of the 12900KF's 16 P-core threads, increasing temperatures (occasionally hitting TjMax at 100 degrees celcius) and fan noise. After the rendering was complete though, the CPU load settled on a much lower level.

The GPU load was not significantly affected by this. However, the impact on VRAM usage was massive, from 12.5GB to 20.5GB.

Next one up is OLOD 🤓

Btw I'm finally on 0.4.3.1-test. 

That was really useful. Thanks! 

[mmcmah]

12 hours ago, Reset XPDR said:

Further to @Cpt_Piett's response, yes the bigger the gap the more change can occur. Previously if you had set 1 or 2 FPS below target, you wouldn't have even crossed the FPS tolerance threshold and no upwards adjustments would occur but with this latest version it will work but much slower than if you gave it a couple more FPS to work up a greater difference from target FPS. In your case, 5 FPS would have done it, but 10 obviously works too.

I primarily use IFR mode since it seems to give a more controlled experience (where you can define more of the parameters). However, I wanted to see if VFR mode got rid of some unnecessary TLOD changes just to hit the TLOD Min at landing/ground value.

I looked at the tooltips, and both modes seem to want to achieve TLOD Min (or TLM+) on the ground, which made me doubt what would happen. I was trying to see if VFR mode would keep my TLOD high - even low and on the ground - if my FPS was high enough. What I found taking off from LSZH is that even though I had PSXT running, I had plenty of FPS and I had my TLOD Min showing before taking off in VFR mode.

Quickly after takeoff, my TLOD changed from 100 to 800 and it worked great. No stutters and no obvious issues with the settings, despite only getting about 1500-2000 ft AGL. I did a quick and tight circuit and started descending to see if the high TLOD values would remain. What I found is that the app tried to decrease my TLOD rapidly as well, as if it was trying to get to the TLOD min for when I was on the ground. This was despite having a good amount of FPS still. However, it was inducing big stutters during the descent, which was causing momentary changes in the FPS values. I ended up landing with a TLOD of 500 and AutoFPS was happy with the value on the ground.

From the log:

2024-03-19 08:56:47.257 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 70 FGAct: True Pri: TLOD TLOD: 100 OLOD: 500 AGL: 5 FPM: 65 Clouds: High
2024-03-19 08:56:56.260 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 70 FGAct: True Pri: FPS TLOD: 108 OLOD: 500 AGL: 104 FPM: 634 Clouds: High
2024-03-19 08:57:05.264 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 70 FGAct: True Pri: FPS TLOD: 180 OLOD: 500 AGL: 340 FPM: 2289 Clouds: High
2024-03-19 08:57:14.265 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 70 FGAct: True Pri: FPS TLOD: 252 OLOD: 500 AGL: 569 FPM: 1718 Clouds: High
2024-03-19 08:57:23.259 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 70 FGAct: True Pri: FPS TLOD: 324 OLOD: 500 AGL: 766 FPM: 1677 Clouds: High
2024-03-19 08:57:32.262 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 70 FGAct: True Pri: FPS TLOD: 388 OLOD: 500 AGL: 1102 FPM: 2682 Clouds: High
2024-03-19 08:57:41.260 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 70 FGAct: True Pri: FPS TLOD: 460 OLOD: 500 AGL: 1320 FPM: 2089 Clouds: High
2024-03-19 08:57:50.263 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 69 FGAct: True Pri: FPS TLOD: 532 OLOD: 500 AGL: 1598 FPM: 382 Clouds: High
2024-03-19 08:57:59.269 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 69 FGAct: True Pri: FPS TLOD: 604 OLOD: 500 AGL: 1574 FPM: 1136 Clouds: High
2024-03-19 08:58:08.268 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 62 FGAct: True Pri: FPS TLOD: 671 OLOD: 500 AGL: 1647 FPM: 268 Clouds: High
2024-03-19 08:58:17.272 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 67 FGAct: True Pri: FPS TLOD: 682 OLOD: 500 AGL: 1751 FPM: 38 Clouds: High
2024-03-19 08:58:26.261 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 63 FGAct: True Pri: FPS TLOD: 724 OLOD: 500 AGL: 1879 FPM: 663 Clouds: High
2024-03-19 08:58:35.269 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 64 FGAct: True Pri: FPS TLOD: 752 OLOD: 500 AGL: 1935 FPM: -144 Clouds: High
2024-03-19 08:58:44.263 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 64 FGAct: True Pri: FPS TLOD: 776 OLOD: 500 AGL: 1804 FPM: -294 Clouds: High
2024-03-19 08:58:53.272 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 63 FGAct: True Pri: FPS TLOD: 800 OLOD: 500 AGL: 1778 FPM: -602 Clouds: High
2024-03-19 08:59:02.270 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 65 FGAct: True Pri: FPS TLOD: 800 OLOD: 500 AGL: 1643 FPM: -718 Clouds: High
2024-03-19 08:59:11.262 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 64 FGAct: True Pri: FPS TLOD: 800 OLOD: 500 AGL: 1512 FPM: -1055 Clouds: High
2024-03-19 08:59:20.266 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 58 FGAct: True Pri: FPS TLOD: 800 OLOD: 500 AGL: 1356 FPM: -713 Clouds: High
2024-03-19 08:59:29.260 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 62 FGAct: True Pri: FPS TLOD: 793 OLOD: 500 AGL: 1246 FPM: -788 Clouds: High
2024-03-19 08:59:38.271 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 60 FGAct: True Pri: FPS TLOD: 795 OLOD: 500 AGL: 1176 FPM: -787 Clouds: High
2024-03-19 08:59:47.257 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 59 FGAct: True Pri: FPS TLOD: 798 OLOD: 500 AGL: 1146 FPM: -899 Clouds: High
2024-03-19 08:59:56.265 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 62 FGAct: True Pri: FPS TLOD: 794 OLOD: 500 AGL: 990 FPM: -1431 Clouds: High
2024-03-19 09:00:05.258 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 60 FGAct: True Pri: FPS TLOD: 793 OLOD: 500 AGL: 818 FPM: -1648 Clouds: High
2024-03-19 09:00:14.261 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 56 FGAct: True Pri: FPS TLOD: 791 OLOD: 500 AGL: 682 FPM: -803 Clouds: High
2024-03-19 09:00:23.258 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 49 FGAct: True Pri: FPS TLOD: 730 OLOD: 500 AGL: 552 FPM: -1119 Clouds: High
2024-03-19 09:00:32.270 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 44 FGAct: True Pri: FPS TLOD: 658 OLOD: 500 AGL: 383 FPM: -1181 Clouds: High
2024-03-19 09:00:41.262 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 45 FGAct: True Pri: FPS TLOD: 586 OLOD: 500 AGL: 213 FPM: -1043 Clouds: High
2024-03-19 09:00:50.260 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 47 FGAct: True Pri: TLOD TLOD: 530 OLOD: 500 AGL: 84 FPM: -747 Clouds: High
2024-03-19 09:00:59.260 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 51 FGAct: True Pri: TLOD TLOD: 530 OLOD: 500 AGL: 8 FPM: -522 Clouds: High
2024-03-19 09:01:08.258 [INF] [ MainWindow:UpdateLiveValues      ] FPS: 50 FGAct: True Pri: TLOD TLOD: 530 OLOD: 500 AGL: 4 FPM: -114 Clouds: High

 

 

[Cpt_Piett]

5 minutes ago, mmcmah said:

Quickly after takeoff, my TLOD changed from 100 to 800 and it worked great. No stutters and no obvious issues with the settings, despite only getting about 1500-2000 ft AGL. I did a quick and tight circuit and started descending to see if the high TLOD values would remain. What I found is that the app tried to decrease my TLOD rapidly as well, as if it was trying to get to the TLOD min for when I was on the ground. This was despite having a good amount of FPS still. However, it was inducing big stutters during the descent, which was causing momentary changes in the FPS values. I ended up landing with a TLOD of 500 and AutoFPS was happy with the value on the ground.

TLOD 100 to 800 before reaching 2000ft, that’s a huge change. Strange you didn’t notice anything then. I haven’t had much chance to test VFR yet, it looks interesting.