CRJ Prof. P3Dv5 Stutters & Performance

[Hans Hartmann 3694]

0.2 seconds sounds very short. I'm not sure if I have anything that's done at such a frequency but I can check. A video would be helpful.

[Hans Hartmann 3694]

42 minutes ago, Wise87 said:

I can add, when I'm in cold and dark my panning is a lot smoother thorough out the VC. Once I'm powered up I see what MediocreGorilla sees. Turning off the FO displays did not improve things. 

Yes, but that's normal because none of the displays is being refreshed while they don't have power. The draw process takes up most of the CPU time of anything in the cockpit (MSFS is much easier on the CPU there). And for 2.4.0.0 I already optimized all the non-essential displays to draw at a much lower frequency or only then changed.

[Wise87 63]

12 minutes ago, Hans Hartmann said:

Yes, but that's normal because none of the displays is being refreshed while they don't have power. The draw process takes up most of the CPU time of anything in the cockpit (MSFS is much easier on the CPU there). And for 2.4.0.0 I already optimized all the non-essential displays to draw at a much lower frequency or only then changed.

i hear you and I know your working on it. It’s just bizarre that only the CRJ gives me this issue when the QW,  FSLabs and PMDG products are silky smooth and they have way more displays. I hope you can narrow it down as I find the CRJ fun to fly.

[Hans Hartmann 3694]

I might have to bother you guys with some test DLLs to track things down. But for now I'm reviewing the source code to see if I can find anything. It's not necessarily a bug. For example I remember that there are some GDIPlus clipping functions that use an awful lot of computing time under certain circumstances. These are things that need to be checked.

[PeterZ KCLE EDDN 6]

I’d be happy to help and test drive dll’s Let me know if you need my email again, or if you want us to report in the beta tester forum. 

 

Schönes Wochenende,

Pete

[MediocreGorilla 2]

Hi Hans,

 

first of all thank you for digging into that, I highly appreciate how you care about your product!

 

 

Before watching the video down below, let me say that stutters ARE noticeable in the recorded footage, but much more intense when you experience it yourself with your hand moving the camera. That's why I have some numbers for you aswell and added the average FPS counter of P3D, which is able to measure FPS fluctuation in percent. And indeed it supports what my feeling tells me: Popping out the displays decreases the variation by almost 10%! The 6-7% variation with popped out displays is what I see on other aircraft aswell, so this is would be the desireable value.

 

 

If I can be helpful in any way, just let me know!

 

Kind regards

Johannes

[Hans Hartmann 3694]

Thanks for the video, Johannes.

 

I have one request to everybody with the stuttering problem: I received a report from a beta tester earlier that the stuttering would also occur when panning through the cockpit with the aircraft in cold & dark state. We need to know whether is the case for all of you or not. Please give it a try and let me know. Thank you

[Wise87 63]

1 hour ago, Hans Hartmann said:

Thanks for the video, Johannes.

 

I have one request to everybody with the stuttering problem: I received a report from a beta tester earlier that the stuttering would also occur when panning through the cockpit with the aircraft in cold & dark state. We need to know whether is the case for all of you or not. Please give it a try and let me know. Thank you

For me as mentioned a few post up, in Cold And Dark it's smooth for me. I can pan, cycle views and its smooth, no lag or hard stutters. Any other panel states it's like in the video above. 

[rmeier 4]

With the latest version (2.4), I've noticed a "click spot" taking up a large portion of the pilot's windscreen.

 

I've noticed that on my first two flights with the new version, if my mouse cursor is within the windscreen click spot, the sim will stutter, particularly bad if I'm using the zoom feature of the mouse.

 

Secondly, I don't recall this issue from v2.2, but, the aileron and rudder indicators do not show movement when testing the flight controls.

[PeterZ KCLE EDDN 6]

Cold and dark is super smooth for me as well. 

[Hoffie3000 96]

vor 2 Stunden , rmeier sagte:

rudder indicators do not show movement when testing the flight controls.

Strange, on my installation it works very well

[MediocreGorilla 2]

vor 9 Stunden , Hans Hartmann sagte:

Thanks for the video, Johannes.

 

I have one request to everybody with the stuttering problem: I received a report from a beta tester earlier that the stuttering would also occur when panning through the cockpit with the aircraft in cold & dark state. We need to know whether is the case for all of you or not. Please give it a try and let me know. Thank you

Cold and Dark is super smooth for me aswell!

[dan3138 1]

Hi @Hans Hartmann,

 

I just ran some tests:

 

Test#1: Immediate T/O and climb to cruise (Test time ~30mins)

1. Loaded up on the runway at EDDC with the aircraft ready for T/O - Everything works smooth.

2. Staring to roll - Starting to notice some minor stutters + FPS jumps.

3. On the climb - As the speed and altitude increases the stutters and FPS jumps become worse and worse.

 

Test#2: Sit on the runway ready for T/O without moving (Test time ~1hour and 30mins)

1. Loaded up on the runway at EDDC with the aircraft ready for T/O - Everything works smooth.

 

Hope this can give some kind lead.

 

Thanks for your hard work,

Daniel.

 

[cconesa 2]

On 3/12/2021 at 5:55 PM, Hans Hartmann said:

I have one request to everybody with the stuttering problem: I received a report from a beta tester earlier that the stuttering would also occur when panning through the cockpit with the aircraft in cold & dark state. We need to know whether is the case for all of you or not. Please give it a try and let me know. Thank you

 

Hello Hans,

 

I have just started from cold and dark and my frame rates are not running smoothly. I am getting an average of 16-25 fps when I would normally run 30 fps without a problem (they are capped at 30 fps). Didn't have this issue before.

 

I will try in a while with a turnaround state to see what happens then.

 

Regards,

 

Christian C./

[cconesa 2]

I further report that changing the start up situation does have a small effect to the FPS but stutters and highly volatile FPS count remain. I cannot remember this being an issue on v2.2, in fact, I remember the CRJ being quite benign on the FPS front.

 

Also, the reverse thrust does not really have much of an impact at slowing down the aircraft.

 

Regards,

 

Christian C./

[PeterZ KCLE EDDN 6]

On my system I get anywhere from 19-88 FPS in VC ready for takeoff (unlocked frames), but if I use either Nvidia or RTSS to limit frames at 30, that’s when the stutters become immensely pronounced.
 

I have my system tuned to where I can fly pretty much any study level sim into KORD at 30 FPS butter smooth. 

 

Cheers,

Pete
 

[Hans Hartmann 3694]

Good morning. I worked on profiling the displays over the weekend and found out some interesting things. So, I changed some settings, optimized the draw order and gave the whole thing to our beta testers. The comments were positive, so let's see if this does any good for you as well.

I'm mostly interested in the average framerate and the variation, so please use P3D's internal framerate display for testing. To add the average and variation to it, edit the [TextInfo.3] section in your Prepar3D.cfg to look like this:

[TextInfo.3]
FrameRate=1,1
LockedFrameRate=1,2
GpuMemory=1,3
AverageFrameRate=2,1

 

These are two gauge DLLs for testing, they go into <CRJ Installation Folder>\SimObjects\Airplanes\Aerosoft CRJ Pro Base\Panel. Don't forget to back up the original DLL.

P3D v4.5: ASCRJ Pro_FPS_P3D45.7z

P3D v5.1: ASCRJ Pro_FPS_P3D51.7z

 

I'm looking forward to the results

 

[MediocreGorilla 2]

Hi Hans,

 

thanks for your efforts! While what you did sounds very promising, I can't confirm if it's working yet. With the DLL in place, the plane is electrically dead, unresponsive, engines at idle, controls all full left and unresponsive.

I took the v5 DLL and put it where you told us to. Did I do something wrong? With the old DLL I backed up it works again.

 

Please login to display this image.

 

Kind regards

Johannes

[Hans Hartmann 3694]

Hmm. OK. Let me find out why this happens.

[Hans Hartmann 3694]

Noob error on my side. Please download and try again.

[PeterZ KCLE EDDN 6]

5 hours ago, Hans Hartmann said:

Good morning. I worked on profiling the displays over the weekend and found out some interesting things. So, I changed some settings, optimized the draw order and gave the whole thing to our beta testers. The comments were positive, so let's see if this does any good for you as well.

I'm mostly interested in the average framerate and the variation, so please use P3D's internal framerate display for testing. To add the average and variation to it, edit the [TextInfo.3] section in your Prepar3D.cfg to look like this:

[TextInfo.3]
FrameRate=1,1
LockedFrameRate=1,2
GpuMemory=1,3
AverageFrameRate=2,1

 

These are two gauge DLLs for testing, they go into <CRJ Installation Folder>\SimObjects\Airplanes\Aerosoft CRJ Pro Base\Panel. Don't forget to back up the original DLL.

P3D v4.5: ASCRJ Pro_FPS_P3D45.7z

P3D v5.1: ASCRJ Pro_FPS_P3D51.7z

 

I'm looking forward to the results

 

Tach Hans,

 

I have, I believe, an interesting observation with the new 5.1 dll. I just started a flight at the gate at Imaginesim's KATL, and what's happening now is - as long as I don't pan around in the cockpit, but while taxiing without panning it is much smoother - but as soon as I start panning, the headache stutter returns. If we can now get it more fluid during panning, this could be a winner Hans. I'll do more testing to confirm this.

 

Also, the stutters are much less pronounced without any FPS limiter usage (regardless of internal, RTSS, or NVIDIA), but return as soon as I lock them to e.g. 30 (half my monitor's refresh rate), or anything higher than 30. I noticed the VC textures are rather large, would you be willing to convert them down to a lower resolution for us to test? I do remember the TFDI 717 caused stutters on my system due to large vc textures, among other things that needed fixing. Just a thought.

 

Cheers,

Pete

[dan3138 1]

5 hours ago, Hans Hartmann said:

Good morning. I worked on profiling the displays over the weekend and found out some interesting things. So, I changed some settings, optimized the draw order and gave the whole thing to our beta testers. The comments were positive, so let's see if this does any good for you as well.

I'm mostly interested in the average framerate and the variation, so please use P3D's internal framerate display for testing. To add the average and variation to it, edit the [TextInfo.3] section in your Prepar3D.cfg to look like this:

[TextInfo.3]
FrameRate=1,1
LockedFrameRate=1,2
GpuMemory=1,3
AverageFrameRate=2,1

 

These are two gauge DLLs for testing, they go into <CRJ Installation Folder>\SimObjects\Airplanes\Aerosoft CRJ Pro Base\Panel. Don't forget to back up the original DLL.

P3D v4.5: ASCRJ Pro_FPS_P3D45.7z

P3D v5.1: ASCRJ Pro_FPS_P3D51.7z

 

I'm looking forward to the results

 

 

@Hans Hartmann,

 

Thank you for you efforts.

I will test it in the next days when I have the time.

Good luck with the MSFS release

 

Daniel

 

 

[Hans Hartmann 3694]

2 hours ago, PeterZ KCLE EDDN said:

Tach Hans,

 

I have, I believe, an interesting observation with the new 5.1 dll. I just started a flight at the gate at Imaginesim's KATL, and what's happening now is - as long as I don't pan around in the cockpit, but while taxiing without panning it is much smoother - but as soon as I start panning, the headache stutter returns. If we can now get it more fluid during panning, this could be a winner Hans. I'll do more testing to confirm this.

 

Also, the stutters are much less pronounced without any FPS limiter usage (regardless of internal, RTSS, or NVIDIA), but return as soon as I lock them to e.g. 30 (half my monitor's refresh rate), or anything higher than 30. I noticed the VC textures are rather large, would you be willing to convert them down to a lower resolution for us to test? I do remember the TFDI 717 caused stutters on my system due to large vc textures, among other things that needed fixing. Just a thought.

 

Cheers,

Pete

Hi Pete,

can you try the panning at a default airport? Part of it might be related to texture loading of the scenery.

 

Stefan sent me this link: https://www.prepar3d.com/forum/viewtopic.php?t=128626

It might be a common problem in P3D.

 

I can't help with the VC textures, because I know just about nothing about how to do that. But I can ask.

[PeterZ KCLE EDDN 6]

Hi Hans,

 

on my system I see no difference between default and add on scenery/airport. More importantly, no other study level  aircraft has that issue whatsoever. I looked at the P3D forum, but I’m not sure if that’s what we’re experiencing here. 
 

i just did another test, and it’s honestly not much better. I even manually removed the hud gauge and true glass to see if that was causing it, but to no avail. 
 

Cheers,

Pete 

[Hans Hartmann 3694]

Can you shoot a video of how panning looks for you right now? I know, there's already Johannes' video further up but that was shot with the older version.

Panning may stutter when you move the view for the first time or after some time, simply because textures need to be loaded (P3D 5.1 unloads textures pretty rapidly it seems - to save video memory - you remember the issues with out of memory errors when it was just released). But when you move back and forth, it should not stutter - or at least very little.