Version 3.7.4 is not functioning properly

[zayixang]

After updating to the latest version 3.7.4, clicking the GSX icon in the game always shows “loading……” and it never finishes loading. The GSX interface fails to load properly.

In addition, I noticed that right-clicking the GSX icon in the taskbar at the bottom right has no response.

[Maverick3333]

same for me ;( and my MSFS2024 with PMDG 737-800 crashed several times !!!

[baramartin]

Same here, msfs2020 pmdg 737-800, crashes every time above 20k feet, ever since the 3.7.4 update. Dont even try to answer this with "it works on my machine" or "read the manual page 372829". We did that.

[kingnorris]

I'm not updating mine then. Yet another update that messes up something.

[zayixang]

I have uploaded all the log files here. Please help to look into the issue. Thank you.

[solentflyer]

Since updating to 3.7.4, when running MSFS24 and parked at KLAX Gate 23 in the PMDG 737-800 (PMDG 737 tutorial flight), couatl64_msfs2024.exe has been using virtually all of my 64GB RAM before my PC freezes due to a lack of virtual memory.

[Maverick3333]

@solentflyer. Yes since the update yesterday I tried 8 flights (EDMA-LGKO) 6 crashes around 10 minutes after launch during my preflight. I didn't even start GSX.But when I come to pushback, GSX connects automatically. I loved GSX but for nor I have to uninstall it!
I tried deleting caches, updated all driver etc,...not working and I have an AMD 9800d3 with RTX 5080, 64 gb RAM (everything installed brandnew) and it's my only game on this high end machine. I think it was the 3.7.4 update who let my Sim freezing so that I have to do a restart. Tomorrow I will check it with the PMDG 777-300 which causes me the last 2 weeks no problems, accept the warning "lower graphic settings" but never crashed. best regards

[virtuali]

Since updating to 3.7.4, when running MSFS24 and parked at KLAX Gate 23 in the PMDG 737-800 (PMDG 737 tutorial flight), couatl64_msfs2024.exe has been using virtually all of my 64GB RAM before my PC freezes due to a lack of virtual memory.

This is a known Simconnect bug, has been discovered recently and has been acknowledged by Asobo as well, it's a memory leak in the Simconnect_Open API, which we used to call if the simulator got stuck and didn't reply, and each time we called it, it leaked memory, so users were mislead thinking it was a "Couatl issue", when in fact it's a Simconnect bug.

Here's a link on MSFS devs forum, the leak has been discovered by AirManager author:

https://devsupport.flightsimulator.com/t/memory-leak-in-simconnect-open-function-sdk-1-4-5/17043

However, in version 3.7.4, we added a specific workaround for that so, even if MSFS got stuck, we won't even try to reconnect continuously anymore, so the problem caused by the Simconnect bug, which would happen with previous version, cannot happen in version 3.7.4

[virtuali]

I have uploaded all the log files here. Please help to look into the issue. Thank you.

You haven't added the most important log, the one from the couatl engine. The only useful one, the restart log, just shows MSFS exited after 5 minutes.

[solentflyer]

Thanks Umberto. In that case, do you have any idea why couatl 3.7.4 is using so much memory to the point that 64Gb RAM is no longer enough? It literally increases by 200Mb per second after entering the aircraft.

[virtuali]

In that case, do you have any idea why couatl 3.7.4 is using so much memory to the point that 64Gb RAM is no longer enough? It literally increases by 200Mb per second after entering the aircraft.

It doesn't do that here. It USED to to do that when:

- the simulator was quit (so the couatl engine *correctly* shut down), but didn't really quit and it was stuck in the Task Manager.

- the couatl restarter detected MSFS was still running, so it assumed the couatl .exe crashed, so it restarted it.

- the restarted couatl .exe tried to reconnect continuously, but since MSFS wasn't responding (because it didn't really quit, it was still stuck in Task Manager), the bug in the Simconnect client (which is still there), leaked memory each time the couatl .exe called Simconnect_open() over and over.

That was USED to happen until version 3.7.3. With version 3.7.4, we detect that case, and we stop trying to reconnect and quit immediately and, as an extra precaution, even if we fail to detect the case, after 10 minutes of failed reconnections, we just stop anyway. Also, the connection rate is much slower and, since the Simconnect is leaking about 160K each time it's called, even after 10 minutes (worse case), it couldn't take more than 30-40MB of RAM, then it will quit.

So, be sure you HAVE the latest version of the couatl .exe. Try the Offline installer.

[solentflyer]

I've just downloaded and installed the latest offline installer from mediafire.com (properties for couatl64_MSFS2024.exe is showing a date of 18 ‎December ‎2025 12:48:23) and it has made no difference i.e. couatl is still using over 40GB of RAM as shown in my last screenshot.

[virtuali]

I've just downloaded and installed the latest offline installer from mediafire.com (properties for couatl64_MSFS2024.exe is showing a date of 18 ‎December ‎2025 12:48:23) and it has made no difference i.e. couatl is still using over 40GB of RAM as shown in my last screenshot.

As I've said, it doesn't do that here, it takes just 300MB and it doesn't increase. Is this happening after some time ?

[solentflyer]

Thanks Maverick3333, seems like you're experiencing the same issue as me. I suspect the sim is freezing as it's running out of virtual memory, which is crazy on a 64GB machine.

[solentflyer]

No it happens as soon as I load a flight.