reported FPS Lost after DCS 2.7.7.14727 patch

[MattCri]

hope this doesnt become permanent

 

we had a similar situation when 2.5.6 rolled. People updating  from 2.5.5 complained of performance degradation

same huge thread on the forum, lots of testing, lots of discussion and then 2.5.6 was pushed to stable "as it" and perfomance never recovered.  

Edited November 3, 2021 by MattCri

[Sr.]

1 hour ago, BIGNEWY said:

Only setting I change was to increase my steamVR % form 54 to 60 as I am seeing good results. With my old 2080ti I was running at 54% in steamVR for higher frames. 

60? 

Prior to the patch, I could run SteamVR SS at 90/100 and get 45 fps (motion smoothing) and maintain 14-16ms GPU frametime.

Now, I can barely stay below 20ms frametime @ 70

[deep]

3 hours ago, BIGNEWY said:

they are identical now, it would be a pointless exercise, only differences would likely be in the saved games folder

The whole purpose of an open beta is to catch problems before software goes into release. There are significant fps problems with the beta, and yet now it's been pushed to release.

I may be wrong, but the only reason to do this seems to be to sell the new campaigns. Sorry to be so negative but this is really frustrating.

[Sr.]

I'm just guessing here but, perhaps the problem is the new reflective water. Would it be a safe guess to think whatever code is applied to make a mirror reflect is also used to make the water reflective.

We all know there's a slight fps drop with mirrors enabled (hence I never use them). Has the entire map (water) been turned into a giant mirror?

 

I think it's been asked but, is it possible to simply turn off the new, reflective nature of the water?

Edited November 3, 2021 by Sr.

[BIGNEWY]

59 minutes ago, deep said:

The whole purpose of an open beta is to catch problems before software goes into release. There are significant fps problems with the beta, and yet now it's been pushed to release.

I may be wrong, but the only reason to do this seems to be to sell the new campaigns. Sorry to be so negative but this is really frustrating.

We do not know what the cause is yet for some peoples fps loss, we know it does not affect everyone and they are using the same DCS versions as the people seeing problems. 

As frustrating as it is unless we get decent information / data from people who are seeing problems it is going to remain difficult to pin down. 

At this point this thread is just people saying yes they do have problems, but not actually sharing information that would help us or give us clues. 

We need to see

dcs.log
dxdiag
track replay showing an fps loss

thanks

58 minutes ago, Sr. said:

is it possible to simply turn off the new, reflective nature of the water?

Try medium or low water see if you notice a difference, and let me know. 

[Hiob]

5 minutes ago, Dzsekeb said:

Some people mentioned they had better results with higher settings than lower ones. I remember reading about shadows "flat only" being worse then "low" or "medium" for some people. I'm thinking maybe there are some specific combinations of settings that don't interact well with each other causing an unexpected fps loss. Might be worth experimenting with.

That are all wild guesses.

What some do not realize is, that the people that are actually raising their voice reporting fps-loss here are still „only“ in the double digits. That‘s hardly a representative sample size of all dcs-players, from whom the vast majority is silent. And of those, only a couple gives the asked data.

There is no way for us to know, how „big“ the problem is. Assume it‘s only a minority of say 10% that experience deteriorated fps. How are you supposed to pin it down without actual data?

You need to give ED something to work with.

And don‘t get me wrong! I don‘t suggest for a second, that the problem should not be dealt with, regardless of small the % of effected players are, or that your experiences aren‘t valid.

[Funflite]

Hi everyone;

 I still flying DCS (stable or open version) but not in VR since the new version of DCS...the 2.7...,definitely ED need to make this simulator running on multicores not on only 2.Everyone want to upgrade their stuff computing system to have the performance needed to run it decently and even that,it's desappointing a lot.DCS must evolve to this decade.You see a lot of mods to compensate the huge problem that ED will not take seriously.The main point is they have not a serious competitor to push them to upgrade this simulator.You begin to see some most viewed youtuber flying IL2 and some return to BMS...because even if you upgrade your computer at the max(max $$$$ too...),it still disapointing.A real upgrade of the core of DCS is needed.

[speed-of-heat]

35 minutes ago, Dzsekeb said:

Some people mentioned they had better results with higher settings than lower ones. I remember reading about shadows "flat only" being worse then "low" or "medium" for some people. I'm thinking maybe there are some specific combinations of settings that don't interact well with each other causing an unexpected fps loss. Might be worth experimenting with.

yes and they were wrong.

[MattCri]

well, just reuploading, im open to share any information necessary

DxDiag.txt dcs.log test patch 2810.trk

[Limaro]

Hello again,

before making the update to the newest stable, I've first created two new test-flights, and run the nvidia logging of performance.
With the new version I've played the tracks again and then also with cleared metashader and fxo.

Average FPS results in a nutshell:

Beirut
2.7.6                79.0 FPS
2.7.7                69.8 FPS
2.7.7 cleared: 67.8 FPS

Caucasus
2.7.6                84.7 FPS
2.7.7                75.1 FPS
2.7.7 cleared: 73.0 FPS

That the game with cleared cache files became even slower shocked me most. So, don't clear your cache after update...

Below see my details data containing:

- Both Trackfiles
- Screenshots of in-game settings
- Report Summary Excel
- For all versions (2.76, 2.7.7, 2.7.7 cleared)
    - Verbose NVidia reports
    - Logfiles
- The DxDiag (is the same as in my previous post.)

This time I was not very motivated to run a whole bunch of settings again. These are now just the once I am usually flying with.

2.7.6_vs_2.7.7_Analysis.zip

[Art-J]

OK, so for the sake of science, I reverted temporarily to 2.7.6.13436 OpenBeta (the last 2.7.6) and ran the fps testing procedure as follows:

1) disabling mods, deleting options.lua, logs and fxo + metashader files. The only custom soft runnig left is Reshade 4.9.1 (injecting SMAA and big framerate indicator);

2) launching DCS, setting up gfx options starting from "High" preset and lowering chosen parameters as I used to do before 2.7.7 - screenshot below shows these settings;

3) restarting DCS;

4) loading P-51D Caucasus takeoff instant mission, when in cockpit setting camera to front / back / left / right view, noting framerate. Doing the same in F2 external view, saving replay track;

5) loading F-86F Caucasus cold start instant mission, repeating the same front / back / left / right fps check in cockpit and in external view like above, saving replay track;

6) exiting DCS, saving logs, both replay tracks and options.lua file.

 

Then I updated again to 2.7.7.15038 and did all of the above in the same sequence.

RESULTS FOR THE MUSTANG:

2.7.6 cockpit: 74 / 71 / 72 / 64; external: 96 / 82 / 86 / 75

2.7.7 cockpit: 60 / 58 / 59 / 53; external: 74 / 65 / 68 / 61

RESULTS FOR THE SABRE:

2.7.6 cockpit: 80 / 76 / 75 / 84; external: 81 / 86 / 83 / 93

2.7.7 cockpit: 64 / 62 / 61 / 66; external: 64 / 68 / 66 / 73

 

All done in 2D, with NVidia 496.13 drivers.

Zipped logs, replays and options.lua files for each versions attached below, together with latest dxdiag.

2.7.6.13436 OB files.zip

2.7.7.15038 OB files.zip DxDiag.txt

Edited November 3, 2021 by Art-J

[Sr.]

3 hours ago, BIGNEWY said:

We do not know what the cause is yet for some peoples fps loss, we know it does not affect everyone and they are using the same DCS versions as the people seeing problems. 

As frustrating as it is unless we get decent information / data from people who are seeing problems it is going to remain difficult to pin down. 

At this point this thread is just people saying yes they do have problems, but not actually sharing information that would help us or give us clues. 

We need to see

dcs.log
dxdiag
track replay showing an fps loss

thanks

Try medium or low water see if you notice a difference, and let me know. 

Easily 3-7ms GPU frametime gain from high to low.

 

This is running SteamVR SS @ 50% ... which doesn't bode well for cockpit clarity from my usual 80-100%

 

Best improvement seems to come with MSAA OFF 
Even 2x loses 4-5ms

Edited November 3, 2021 by Sr.

[Boosterdog]

6 hours ago, Hiob said:

the people that are actually raising their voice reporting fps-loss here are still „only“ in the double digits. That‘s hardly a representative sample size of all dcs-players, from whom the vast majority is silent. And of those, only a couple gives the asked data.

There is no way for us to know, how „big“ the problem is. Assume it‘s only a minority of say 10% that experience deteriorated fps. How are you supposed to pin it down without actual data?

You need to give ED something to work with.

 

100% agree with the sole caveat that this forum as a whole is hardly a representive sample size of all dcs players either. But big or small, if you report a bug and are asked for data, please provide it. 

I do have to wonder if its something outside of DCS in the OS, botchedOC (my botched I hasten to add) or even the hardware. Ill soon be upgradng the unsexy part of my system (MB, Chip, fresh OS and RAM) and will be curious to see how and if that alters the metric on an unbutchered OS. 

[Sr.]

I've managed to recover some serious fps by setting water to Low, turning off MSAA and applying SOH 2.7.7 shader mod.

[Hiob]

3 hours ago, Sr. said:

I've managed to recover some serious fps by setting water to Low, turning off MSAA and applying SOH 2.7.7 shader mod.

That‘s kind of clear, isn’t it?

Water High->Med gave 5% fps, and MSAA 2x->off gave 15% in earlier tests already, before the latest updates were a thing.

[lefuneste01]

There is something more fondamental in fps issue, my rig is not used at its full power for some missions:

 

Edited November 4, 2021 by lefuneste01

[DmitriKozlowsky]

Yes me too. I am on latest 2.7.7.15038. FPS dropped on build 2.7.7.14727 by 15% then dropped another 10% on .15038.

Where I was getting above 140 on last 2.5.6. In 2.7.7.14727 went down to 100-120. In 2.7.7.15038 dropped to 78-90

The control test is Instant Action/FA-18C/Nevada/Cold & Dark and Instant Action/AV8B/Nevada/Cold & Dark

E5-1630v4 3.7-4ghz turbo boost, 64gb, RTX 2060Super FE(new replacement board), PCIe NVMe HP TurboDrive C:, WD Blue 1TB SataIII 6GB SSD D-drive, 2560X1440 @144hz DP 

Windows 10 Pro 64 21H1 OS Build 19043.1320

Nvidia RTX drivers: 472.39 Studio

I cannot use 496.XX Game Ready becouse those do not work with my DCC applications. 496.13 and 496.36 GRD have no impact on DCS performance vs. 472.39

Overall trajectory of DCS is downwards performance degradation with each build of DCS.

What is sad is that I am planning, when possible to upgrade RTX 2060Super to RTX 3060Ti (when available), and that would only five me a 15-20% increase. RTX 2060Super and RTX 3060 have identical DX12 performance. RTX 3070Ti only a 30% boost in DX11/DX12 . So even with 3070Ti, I wont see kind of FPS I saw in 2.5.5 and 2.5.6.

Maybe its the water, maybe its the clouds. 

Options: See attached

 

[Lange_666]

12 hours ago, BIGNEWY said:

We need to see

dcs.log
dxdiag
track replay showing an fps loss

Screenshots are taken at exact the same moment in the track i used to mesure.

Included in zip:

DXDiag -- FPS test stable and open beta tracks (same mission) -- DCS logs both stable and open beta -- screenshots of both settings (but they are the same).

Although i still wonder how one can see an fps loss on a track if it is played on another computer?
If i play someone else's track, it plays with my setting at my PC's performance. Just did a test with Limaro's track from a few posts above, the loss from "old"  nostable to open beta is clearly visible, with my tracks and his track. If i play whatever track from someone else the result is always the same, fps loss between both!!

I also wonder why not one single PC over at ED's is showing this problem let it be a low, medium or high end machine.

DCS Test.zip

[Hiob]

36 minutes ago, Lange_666 said:

 

Although i still wonder how one can see an fps loss on a track if it is played on another computer.

Because, that's not what it is about. Because to determine, why SOME people are experiencing deteriorating performance and others don't, there have to be a trigger.

Be it a certain combination of hardware or settings or other variables. It's about finding a pattern. This is how empiric analysis works.

Anecdotic reports are not a valid method on any kind of research.

[AdrianL]

46 minutes ago, Lange_666 said:

see an fps loss on a track

The track will contain your graphic settings as well as allows ED to test your scenario against various builds.

I have seen no performance loss. And for the 8 people I fly with, only 1 has seen performance loss. So it is not a common occurrence.

Edited November 4, 2021 by AdrianL

[Hiob]

51 minutes ago, Lange_666 said:

 

DCS Test.zip 1.71 MB · 1 download

@BIGNEWY This is nothing but another pitch of a wild guess/idea:

One thing to notice in the provided screenshots by Lange_666 is that there is obviously a difference in core usage of the cpu. May it be, that there has been updates/changes introduced (either by microsoft or by the game core) regarding core-scheduling? As far as I learned, the approach to core/thread-handling in win11 is supposed to be the culprit for the sub par performance of certain cpus. May it be, that a similar (or even connected) problem leads to the issue at hand?

Underusage of GPU and seemingly more load on cpu cores seem to a repeating pattern....

Edited November 4, 2021 by Hiob

[BIGNEWY]

2 minutes ago, Hiob said:

@BIGNEWY This is nothing but another pitch of a wild guess/idea:

One thing to notice in the provided screenshots by Lange_666 is that there is obviously a difference in core usage of the cpu. May it be, that there has been updates/changes introduced (either by microsoft or by the game core) regarding core-scheduling? As far as I learned, the approach to core/thread-handling in win11 is supposed to be the culprit for the sub par performance of certain cpus. May it be, that a similar (or even connected) problem leads to the issue at hand?

Underusage of GPU and seemingly more load on cpu cores seem to a repeating pattern....

 

certainly worth asking the team, I will let you know if I find anything. 

thanks

[Miro]

21 minutes ago, AdrianL said:

The track will contain your graphic settings as well as allows ED to test your scenario against various builds.

I have seen no performance loss. And for the 8 people I fly with, only 1 has seen performance loss. So it is not a common occurrence.

 

I suspect that everyone has drops, and only those people who are on the limit of performance can see, like me, all on ultra, sliders to the right except for shadows on the GTX 1080. I can see every frame that escapes me.

[Hiob]

19 minutes ago, BIGNEWY said:

certainly worth asking the team, I will let you know if I find anything.

Cool!

In the meantime, those who are effected by the fps loss, could try to lock the DCS-App to certain cores (via the task-manager). Best practice would be two cores on the same die.

There are tools, to determine which are the best cores on your cpu.

 

6 minutes ago, Miro said:

I suspect that everyone has drops, and only those people who are on the limit of performance can see, like me, all on ultra, sliders to the right except for shadows on the GTX 1080. I can see every frame that escapes me.

No, as provided earlier, with a controled repeatable benchmark scene (which I've provided), I get the exact same FPS as I did before the patch. Sorry.

Unless you have capped your fps by some means, everybody is on the edge of performance for his given settings....

Edited November 4, 2021 by Hiob

[Miro]

5 minutes ago, Hiob said:

Cool!

In the meantime, those who are effected by the fps loss, could try to lock the DCS-App to certain cores (via the task-manager). Best practice would be two cores on the same die.

There are tools, to determine which are the best cores on your cpu.

In my case, the graphics card clogs up, wear reaches 100% and frames drop down. In previous patches (the working ones) it didn't get clogged. If the DCS had changed graphically, it would have been understandable, but nothing has changed, except for water, the reduction of which to low does almost nothing.