Prepar3D: PMDG + FSUIPC = 1-second micro stutters

[RISCfuture]

Hello, I'm not sure which add-on is to blame here, but if I use both a PMDG aircraft (737 or 777) and FSUIPC, I get small pauses once a second. Use any other aircraft and the frame rate is smooth. Disable FSUIPC and the frame rate is smooth.

 

Notes:

• Re-generating Prepar3D.cfg has no effect; still happens with default settings
• Re-generating FSUIPC.ini has no effect; still happens with default settings
• Clearing shaders has no effect
• Autosave is disabled; and in any case, the stutter is regular at once a second

Wondering what else I can try to help address this problem.

[Pete Dowson]

5 hours ago, RISCfuture said:

Hello, I'm not sure which add-on is to blame here, but if I use both a PMDG aircraft (737 or 777) and FSUIPC, I get small pauses once a second. Use any other aircraft and the frame rate is smooth. Disable FSUIPC and the frame rate is smooth.

Those two aircraft are very heavy on FS resources, and if you have enabled the facility for FSUIPC to expose their cockpit values via offsets, FSUIPC is also involved. But the data involved is only sent when it changes -- which for some values could be much more frequently than one second, and for others a lot less frequently.

Some of the "normal" FS data FSUIPC provides is supplied at a maximum rate of once per second,  But again, most stuff doesn't have a minimum time limit like that.

Have you looked at the FSUIPC log, to see if there's anything there suggesting a problem, or some activity every second? Most of the time, when not servicing requests from applications and plug-ins, FSUIPC is doing nothnig but awaiting data arrival from SimConnect.

AI Traffic tables in FSUIPC are maintained every 1 second, normally, as is the Weather data. But those should be the same irrespective of whether you are using those aircraft or not. However, you could try one of these FSUIPC4.INI settings to stop the weather reading:

WeatherReadFactor=0
or
NoWeatherAtAll=Yes

and of course you can turn traffic off completely, to see if that does it.

But I really can't think of anything involving both those aircraft AND FSUIPC except that data reading already mentioned -- unless of course you have some other add-on which uses FSUIPC more when those aircraft are running -- perhaps a Lua plug-in? If you have other addons you might need to go through a process of elimination.

Pete

 

[RISCfuture]

No Lua plugins; I had "Lua log lvars" bound to a key but I removed it.

I added the two directives to the [General] section and still no luck. I also moved all AI traffic sliders to zero and the stuttering remains. One thing to note is the stuttering only happens when the game is unpaused; when it is paused, frame rates are consistently smooth.

The only other addons/DLLs I have running are PFChid, ActiveSky, and the PMDG DLLs. As I mentioned, using anything other than a PMDG aircraft, or disabling FSUIPC, makes the stuttering go away.

[RISCfuture]

The logfile looks very normal; in fact, very little is logged between startup and shutdown:

[snip]
      859 SimConnect_Open succeeded: waiting to check version okay
      859 Trying to use SimConnect Prepar3D
      859 Opened separate AI Traffic client okay
     9265 Running in "Lockheed Martin® Prepar3D® v3", Version: 3.4.18.19475 (SimConnect: 3.4.0.0)
     9265 Initialising SimConnect data requests now
     9265 FSUIPC Menu entry added
     9265 C:\Users\riscf\AppData\Local\Lockheed Martin\Prepar3D v3\Prepar3D_Default.fxml
     9265 D:\Program Files (x86)\Lockheed Martin\Prepar3D v3\SimObjects\Airplanes\IRIS Raptor Driver\Raptor.air
    27968 Weather Mode now = Theme
    29453 D:\Program Files (x86)\Lockheed Martin\Prepar3D v3\SimObjects\Airplanes\PMDG 737-800NGX WL\B737-800WL.air
    29453 C:\Users\riscf\OneDrive\Documents\Prepar3D v3 Files\AAL3228.fxml
    51265 User Aircraft ID 2 supplied, now being used
    51500 System time = 13/01/2017 19:44:56, Simulator time = 19:45:00 (01:45Z)
    51703 Aircraft="Boeing 737-823NGX American Airlines (OC) Winglets"
    56562 Starting everything now ...
    56562 Loaded PFChid.dll okay!
    56625 Setting external joystick scanner link ...
    56625   ... Okay!
    56625 ASN active function link set
    56625 Ready for ASN WX radar
    57734 Advanced Weather Interface Enabled
   159812 Sim stopped: average frame rate for last 109 secs = 26.4 fps
   159812    Max AI traffic was 53 aircraft (Deleted 0)
   176312 Joy 6 taking too long to respond (> 20 mSecs). Removed from scan
   424843 Sim stopped: average frame rate for last 246 secs = 30.7 fps
   424843    Max AI traffic was 52 aircraft (Deleted 0)
   762750 Sim stopped: average frame rate for last 333 secs = 31.5 fps
   762750    Max AI traffic was 55 aircraft (Deleted 0)
   798203 === Calling SimConnect_Close ...
   798390 === SimConnect_Close done!
   799406 System time = 13/01/2017 19:57:24, Simulator time = 19:56:26 (01:56Z)
   799406 *** FSUIPC log file being closed

 

[RISCfuture]

Made some helpful progress debugging this.

The stutters go away if I disable PFChid but leave FSUIPC enabled. So it's clearly related to my PFC C2. Looking at the PFChid.log file, the only thing that happens once a second is a LandingGear macro event being fired. My PFC.ini file has this macro mapped to C1126 (no event)... is it possible that this is what's causing the brief pauses? There are also brief pauses if I unmap it and let the default mapping (landing gear lever) take priority, but of course this also results in the landing gear lever moving in and out of the locked position once a second…

[RISCfuture]

OK, another helpful update. What I actually did was create a new PFC C2 macro file where I actually copied every single possible mapping from the CSV file and assigned them all to C1126 (no event). That fixed the stutters! So, given that, here's what we know:

• One of the PFChid default mappings must be "unmapped" in order to stop the stutters.
• It's not the landing gear or any of the other defaults I had already unmapped back when I had the stutters.

[Pete Dowson]

19 minutes ago, RISCfuture said:

OK, another helpful update. What I actually did was create a new PFC C2 macro file where I actually copied every single possible mapping from the CSV file and assigned them all to C1126 (no event). That fixed the stutters! So, given that, here's what we know:

• One of the PFChid default mappings must be "unmapped" in order to stop the stutters.
• It's not the landing gear or any of the other defaults I had already unmapped back when I had the stutters.

The only thing I can think of doing is to eliminate the others.

Assuming only one thing is doing it, use a binary division process for speed. Assign 1/2 to "no thing". If it isn't in that half, it must be in the other half. Either way, reenable or disable half the others (now down to a 1/4, and so on. For, say, 256 entries this gets you the culprit in 8 tries only.

Unfortunately that won't work if there's more than one culprit -- the resluts will be indecisive. So then the only way is all off, as you did then re-enable them one by one. When it occurs again, note that one and disable it again, then continue.

I don't have a C2 so I can't really help more directly. Certainly such stutters never noticeably occurred when I did have the C2, and yours is the first such report in all the years it has been around with my driver. Maybe it is because of a change in FSUIPC, responding to whichever command it turns out to be.

Pete