Sim Update 5 has landed and it is obvious that architecture has changed significantly.
Now MSFS tries to push the GPU to its limits. This is very stressful and dangerous.
When GPU is running at 100% for prolonged times, it heats up to 80C and beyond, draws a lot of power, increasing the risk of failure. Disbalance in power consumption may affect other components and the stability of the whole system in case power supply fails to handle the skewed load. Other components may receive less power and starve and thus become unstable too. Some power lines may appear weaker than the others etc. Capacitors here and there may have flaws that may reveal themselves under heavy load.
At this extreme state, any weakness in power supply chain and cooling may lead to crash and burn. Not only GPU is in danger but other components too.
I believe that stress and resulting overheat and power instability is the major cause of the crashes (CTDs)
For some of the users that report CTDs that appear and disappear, the case may be that the hardware has accumulated heat and wear after prolonged stress. So the CTDs will be more frequent. But after components cool down, the CTDs may become less frequent for some time.
So avoid stressing your hardware.
Limit the global FPS in driver settings to something between 30 to 60.
Monitor the temperatures.
Do not let the components work under stress for prolonged times.
Get the best power supply unit possible.
8 Likes
Modern computers should be able to run at 100% for years on end.
Also almost none of what you wrote is true.
66 Likes
with adequate cooling it is a non-issue, the card will run within its manufacturer assigned limits and will protect itself by downclocking and in extreme cases powering down. With sufficient cooling, cards will run within their temp limits without issue.
Set a 60FPS limit - No need for 60+ frames
For those with thermal issues:
- Increase case airflow
- Increase GPU fan curve
- Undervolt GPU
- Clean GPU cooler
Still reaching max temperature and causing card to downclock? (Should not be required)
- Reduce frame limit further
- Downclock GPU manually
7 Likes
not only overheat issues, but power supply issues are here too, affecting other components
That’s exactly what I wanted to write, the high temperatures will be one of the reasons for the many CTDs.
On my PC, I renewed the cooling paste for GPU and CPU a few months ago, the temperatures have been absolutely ok. Now after the SU5 I was shocked that they suddenly rise enormously.
Yesterday I even had a Windows bluescreen! Never before has there been such a thing.
When I think that many PCs will be very poorly cooled, I can well imagine that for these people the simulator may no longer start, because the temperatures are simply too high.
Well, someone can overclock too much, buy poor quality components or put in an inadequate power supply so I guess it is technically possible for software to cause hardware issues.
Though as you allude to, it isn’t the softwares fault. They didn’t put hardware together or set it up badly.
3 Likes
The problem is also that PC ressources get limited to Xbox specs. RAM and VRAM usage is just horrible:
This is what happens on ultra settings in VR on an extremely powerful HEDT PC. No wonder why CPU and GPU get stressed more because of absence of proper data caching. I now have to limit FPS to 20 to get GPU usage below 100%. Before SU5 i was able to reach ~40 fps with ultra settings in VR.
3 Likes
With a proper power supply this also is no issue.
Say you have a high-end computer that draws about 500 W at load. You should probably have a 750 W PSU in such a computer, so it’s running at 66.6% load. Any even OK power supply can handle that 12 hours a day for years.
3 Likes
the problem with power supply is that significant power draw via one power line may affect other power lines, including voltages, amperages etc.
I have never seen my hardware running this hot after the update.
My CPU sits between 75c-78c and GPU 75c.
I guess I will limit my fps to around 35fps.
Ryzen 5 5600x
3070 rtx
32gb RAM
850w PSU
nothing in this world in designed to run at full capacity for long time periods.
tons of burned video cards from mining farms are the example.
another example may be aircraft engines that are certified to maintain full takeoff thrust for only 5-10 minutes
My gpu is at the hottest it’s ever been , so I have had to ramp up the fans , and have ordered some higher rpm noctua fans
Plus i will lock FPS in game and nvidia control panel to 30 FPS
I do limit my FPS but for a very different reason: I get a much smoother experience, way less stutters, and at 30fps the feeling is butter.
I have not had the chance to expermiment with this in update #5, but I’ve had not a reason to, anyway.
2 Likes
This is one of the biggest misunderstandings about computer components, and its a big myth. Its not comparable to an aircraft engine for a few reasons I won’t go into here.
Lets take a graphics card for example, the materials and components are designed to cope with heat for an extremly long time. In-fact its likely something else on the card dies (non-heat related) before heat damage occurs.
Basics of physics. Metal expands and bends when its rapidly heated, and when cooled down too rapidly, will bend it. So making a GPU go from like 5c to 95c quickly a million times may cause wear eventually, (however the GPU fans are there to stop that happening) but consistently having a GPU at like 70c wouldn’t cause any wear as such.
Where you start getting into risky territory is is the temperatures go passed the safe manafacture limits for a GPU this is around 93c, this may start to cause issues with stuff like the lubricants used etc, when this starts to occur the GPU will throttle to reduce temperture to try and avoid damage.
What happens when your GPU goes above 93c for a long time?
- Plastics “soften”. Materials’ mechanical properties generally degrade as -they approach melting point or glass transition temperatures or …
- Chemical reaction rates increase with temperature, including oxidation from atmospheric Oxygen, impurity reactions etc.
- Components that rely on internal liquid content dry out. Notable examples are electrolytic capacitors and batteries.
- Lubricant properties change (eg motor bearings).
- Defects in materials are liable to form faster and propagate more rapidly and further. (This is a major factor in eg long term LED lifetimes).
But stuff like this only happens if tempertures reach dangerous levels, none of this will occur if you run the card at 85c for 10 years for example, thats fine.
Also cards have ways to throttle to stop them hitting damaging levels.
Running a card at 70 to 80C forever probably won’t cause any more wear than if the card was switched off in a plastic package. Well to be specific it will wear slightly more at 100% but its so tiny it doesn’t matter.
I should note there is some things which will wear if a card is 100% all the time, the obvious one is fan bearings, these are subject to friction so eventually running fast all the time causes more friction and they can break. (But I should mention this is still very rare, for example people have cards which are 10 years old and used daily and the fan bearings are still going! Technically the fans starting up and stopping also uses more friction than then running at 100% all the time)
16 Likes
Where exactly do I limit the frame rate?
1 Like
I have stumbled on 30FPS as the smoothest for my rig…Ryzen5900x, RTX3080ti, 32 Gb 3200 ram…..and I mean smooth, with all the sliders up to max…except for resolution at 90% as I’m running a 4K TV.
Tried 60…but 30 is smoother…….like silk.
1 Like
The OP is absolutely correct to bring this up.
After the 60GB+ update (with more to come as addons are updated), I had to change my GFX options again and then tested, my 3090 was running at 94% utilisation sat on the tarmac. This is a huge increase over what I used to get util wise (I was running aggressive GFX settings) and immediately sounded alarm bells.
I know some of you will think we can run at this high level for a long time, but is it worth the gamble? for me when a card like that is virtually irreplaceable due to stock issues / cost, I will be limiting my FPS out of the game.
Just worth a thought in the mind.
Thank you OP.
1 Like
Yes, MFS shouldn’t be maxxing your GPU in menus, but let’s make one thing very clear.
Your CPU/GPU overheating is very much an issue YOU need to sort out regardless. Whether it’s MFS or any other application.
Do people rely on Vsync functionality this much? So one day it turns itself off and your GPU is cooked?
Cool and maintain your systems properly guys.
1 Like
in nvidia control panel, there are global settings, there are 2 settings: max frames per second and background frames per second.
i’ve tried tuning these particularly for flight sim exe but this didn’t help. only global settings made change
1 Like
that is right but the reality is more severe. components may have defects and flaws that manifest themselves under prolonged heavy load.
another issue that most ignore: power consumption and distribution which affects other components in the system