I feel like this whole discussion is derailing.
I dont think that anyone said that OC or UV will cause issues in MSFS.
This all started to make aware that it CAN cause issues and be the source of CTDs.
So if someone has an UV or OC (incl. RAM, eg. XMP) it might be worth to temporarily deactivate the OC / UV to see if it solves the issues with MSFS.
If it does not solve it, it wasn’t what caused it and you can turn it back on.
If it solves it, you can reactivate it but just clock it back a bit until you find the sweetspot where you get most out of your hardware while running MSFS stable.
In my case, I can run a quite aggressive UV but had to reduce the RAM Speed from the XMP profile (6400) to 6200, while leaving all other XMP timings.
A neglectable reduction with a huge effect to stability.
For my GPU I can not add the +10% powerlimit wirhout risking a “DX Device hung” from time to time.
But i still get stable 60 FPS in 2D and 30FPS in VR with high to ultra settings, TLOD 200 and OLOD 150, BATC with 100% traffic (AIG models) and Fenix, Inibuild or FBW on complex 3rd party airports.
So for me, slightly reducing the XMP profile and removing OC from the GPU helped a lot and I did not have a CTD in a very long time.
But if you dont have any CTD, then you dont have to touch any settings, of course.
I think it would help if everyone can stop trying to be the “Smartest Guy in the Room” for a few, take a step back and remember the basic points, which are all well-covered up-thread.
The sim is a very complex beast, and many of us are running quite complex pieces of code (e.g., aircraft and sceneries) inside the sim itself. Hardware comes in all varieties and capabilities, and even the same brand and model of hardware can vary in quality and capability, especially when pushed towards its limits. If the sim is having stability issues, a good first troubleshooting step is to revert your hardware and software settings to defaults and begin testing methodically to see what may be triggering your problems.
As for me, a modest -10 all-cores PBO undervolt gives me stable sim performance and low temperatures. Five hours into a 12 hour long-haul, things are doing great.
I run my sim capped at 60 fps, 1440p, TAA, TLOD and OLOD both at 200. I get very nice visuals, system temps and total utilization stays low, giving plenty of headroom - both thermally and in terms of processor cycles - to handle sudden demand or transients. My CPU runs up to ~5.25Ghz on all-cores during a typical sim session and maintains that more or less indefinitely. Absolutely no need for any overclock beyond standard boost settings - I’m not rendering complex animations or processing videos where every percent of extra speed gives me a few dozen extra hours of processing over the course of a year. My sim rig is essentially a grownup’s toy and what I want from this toy is stable, consistent performance with my favorite hobby (flight sim) and whatever other games I play as the mood strikes.
Same thing for my GPU - 16GB covers my needs currently, and with the sim running at 1440p and a second 1080p monitor for charts, ATC, etc., my card is absolutely ideal. Would I like more VRAM? Well, if Jensen spent less on leather jackets and devoted that energy to his original core customer base, we’d have it across the board, but it is what it is. We’ll see what a 50 Super series looks like this winter; maybe I’ll upgrade or maybe not. But in terms of raw pixel-pushing throughput, I have zero need for it. The card handles everything I throw at it, and even when I run it at 98%+ utilization, temps peak in the low-60s C. Why bother undervolting or overclocking that? It’s just asking for trouble.
Anyway, my two cents.
2 Likes
With performance like that, I wouldn’t be overclocking either. 
Very nice!
1 Like
I think we’re on the same page. I’d also suggest setting the CPU Loadline Calibration Control - I went with Mode 3.
This setting helps prevent a significant Vdrop under CPU load, which in our case happens during gaming. If the Vdrop is too high, CPU might not get "enough power’, which can lead to system instability or even CTDs. Generally, on MSI motherboards, the best results are achieved with LLC set to 3 or 4.
3 Likes
I learned about LLC from a YouTuber a few months ago. I also run mine at the ‘Medium’ setting.
The reason for changing it from the default (which allows a lot of Vdroop) simply made sense.
2 Likes
Been out of hand a bit working on OCing mysystem and can report that SU4 loves OCing my Ryzen 7950. I managed to get a 5% bump using the ASUS idiots guide to OC and verified that it runs OK.
This 5647 MHz translated to 45-60 FPS at TLOD of 300 and mostly high settings. I am using TAA and getting crisp cockpits in the 787 in 2D… VR is VR.
In sum, all working well with SU4 for me pushing the hardware.
1 Like
In SU4 testing I’ve seen some temp spikes in some scenarios.
Looking at your screen shot, I’m just wondering if you applied any undervolting at all? I would be a little nervous about those 90 degree levels over time (although the safety mechanisms of PBO should prevent overheating by shutting down at a certain point, if you haven’t overridden them)
Your motherboard BIOS should have “Curve Optimizer” settings, usually found in AMD Overclocking > PBO > Curve Optimizer > and to reduce voltage, you would start with All Core, Negative 10 or Negative 15. This can result in a roughly 8 to 10% decrease in temperatures, which really helps with those spikes.
1 Like
No I did not apply the Undervolting as i was trying to do the simplest thing first. This is just their OC AI so to speak. I would like to apply both if I could. I guess this thing just runs at these hot temperatures. Scares me too, but seems to be a new world.
What I noticed regarding SU4 was that doing this erased any stuttering and my FPS is running between 40-60 under all sorts of graphics stress… I do have a 5080. I tried the ASUS AI OC for that and that definately ran hot with fans blazing and did not do as much for performance. MSFS is CPU bound it seems.
Well, several people gave me the same advice and I tried it at -10. The computer became unstable and I had to go back to the old settings. I was able to run some benchmarks while it was at -10 and remarkably it ran at 5775Mhz but the temp was still 95-96C.
I tried to run MSFS and it crashed. After I restored the old BIOS settings MSFS would not open and would not repair so I am trying uninstalling and reinstalling.
No sour grapes as I knew this was on the edge. At least the computer seemed to go back to the old BIOS settings and I am glad I did not try to do this in Armoury Crate or some other software. At least BIOS changes seem reversible.
This really is what this thread is all about. Getting people to understand that any system mods you make COULD cause instability in any game. It just happens that flight sims seem to have always been the ones to push the envelope.
So, it is not about stopping people from trying mods. It is about getting people to back up their bios BEFORE they make any changes. Then, if they get any issues it is very simple and quick to restore.
It also enables users to understand their situation and to not go blaming Asobo without good cause.
You tried something, it didn’t work, and you fell back. Fairly straight forward, and you know that it was not caused by Asobo.
3 Likes
Yes, unfortunately I do have still a bone to pick with MS since the standard repair and reinstall of MSFS 2024 does not work and gives me the 0x80070032 nonspecific error.
I once did a clean reinstall I think after SU2 came out but I cannot find the detailed instructions. Anyway, all part of the game. It was worth trying… it did run at almost 5800 Mhz for 15 minutes 
Reinstall gaming services.
1 Like
Keep in mind that you need to look at the effective core clocks with sleep states (and other items) being factored in. If you run the HWINFO64 utility it will show you the effective clocks which are not quite the same as what your pics above are showing.
1 Like
Yes, nothing is simple but there is a repair tool:
Xbox Support .
Back to baseline. But in fact the thing is running better that before at 60FPS down the ILS into SFO with the default settings. Just goes to show it is a good idea to take a laxative every once in a while…
Thanks 
That repair tool doesn’t always work - I’ve checked it myself while helping other users, which is why I suggested reinstalling Gaming Services instead. Anyway, if it did the trick for you, great.
Fly safe.
Thanks, whatever you got me back on track and I am back to baseline.
But I decided to run that same Cinebench Benchmark without the undervolting and I get the picture that this unit does like to run fast but HOT. I have to look up that other HWINFO64 program for comparison.
I would suggest you run the Prime95 stress test just to be sure none of your cores are failing.
Thanks this is run after I dropped the UnderVolting. Comparable on the other software above.
Well, i am certainly getting an education here! That one also seems widely used in the various instruction YouTube Videos…. I have used another called OCCT. OCCT I recall will run up to an hour on the free version. The Cine one seems to give you 1/2h.
I don’t use the monitoring software you mentioned. I’ve been running HWiNFO64 for years.
FYI, if you want the most accurate, real CPU temperature, install Ryzen Master. It’s an official AMD tool, so its readings are the most reliable.
And about overclocking - use tools like Prime95 or OCCT only as a rough reference to stabilize the system. Remember that during these stress tests the temperatures will be much higher than what you’ll ever see while gaming. What really matters for you is the temperature during actual gameplay, because stress-test temps are always significantly higher.
When it comes to overclocking in general, I’d really recommend using a forum that’s actually meant for that for example overclock.net rather than the MSFS forum, which isn’t a great place for this kind of discussion.
1 Like
