Just a question. How much could Asobo reduce certain settings in the background before any one would complain of lower performance. I am thinking of small changes of say 1-2% incrementally.
The same also applies to increasing before it is noticed.
Assuming the game seems to improve for no particular reason.
I donāt quite follow, do you mean increasing certain settings before people would notice lower performance?
Yes or reducing them
I use psxt and real traffic.
I know it learning the airport we are at and also learning the airport we fly to. I canāt say for sure if it works, I can try 
Also it needs to be real traffic reported from there to work 
1 Like
I remember many posts at launch of people complaining about ādreadful performanceā who were using 3-4 generation old midrange CPUs with GPUs of the same vintage that barely exceeded minimum recommended specs, but were trying to run 4K Ultra.
I met "recommendedā specs at launch and while there was a stuttering issue, my frame rate was fine until the Dec 2020 Sim Update (VR introduced) tanked my performance.
1 Like
And to reinforce that - I did some comparison testing on Xboxes - Series X vs Series S. the results are astonishing - flawless, fluid graphics on ramp, in air over complex landscapes in weather - on the Series S. stutterfest panning on ground or in air on the Series X. The difference in performance is huge. The series S has a much more limited radius in terms of detail, so for sure something has changed in the code which isnāt server related - and we see it on the Series X too.
I have no doubt about your test results. However, in the real world at large, itās not that simple.
Itās not a platform thing. There are people with both Series X and S who have flawless performance. There are others experiencing a low performance slide show. And there are others unable to even get to the main screen without a CTD.
Any assumptions about āan Xbox is an Xbox, theyāre all the sameā can pretty much be thrown out the window. YOU may have had this particular experience and youāre 100% in your observations. The next guy trying the same thing could potentially have a far different experience trying exactly the same thing.
I once again reiterate my often repeated quote - the most consistent thing about MSFS is how inconsistent the experience is from user to user.
4 Likes
I had thought that when I was tweaking the series X for weeks after SU9 - and neglecting the CTDs, though they are serious for sure - there is a difference in performance between S and X and it is not a marginal difference either. I did the comparison tests over a weekend having read somewhere that someone was seeing much smoother performance from their Series S.
I did not expect the results I got from the comparison
I think there is definitely something going on to do with object rendering, Honestly the difference in performance post SU9 is astonishing. The Series X struggles with cockpit panning on the ramp or at altitude over terrain in clouds - I found that the best Xbox Series X performance was at low level VFR - but even there panning around the cockpit is now a bit jagged.
On the S? Everything is super-smooth under all the conditions Iāve thrown at it. And there is no daily variation in the performance - at least not that I can seeā¦.
So Iām starting to think that a core problem with object rendering may turn out to be the root cause of the issues weāve been seeing in terms of stuttersā¦
But we still have different hardware in our PCs right. We will always have different hardwares.
The VR experience i know for sure is related to that the camera is moved alot more using that thing also needs to have much higher fps to feels smooth than without VR.
Thereās a bigger problem at play under the hood of the sim that has nothing to do with hardware or platform, as the results are seen on both XBox and PC.
On PC, folks with top tier 12th gen Intel CPUs and 3080+ tier cards are seeing slide shows while people with modest hardware are seeing excellent performance. But for others, itās the opposite. Same as your results testing on Series X and S.
While your results arenāt invalid by any means, theyāre not representative of the broader community who can have wildly varied performance and issues on even the same hardware (Series X vs Series X for example). At least not the conclusion that Series S runs MSFS faster than Series X.
True. But I donāt use VR in MSFS. They changed the rendering engine to make it work in VR. I was running 50+ fps without issue. The Dec 2020 update where VR was introduced, my FLAT SCREEN performance dropped from 50+ into the mid 30s with the same settings. Each subsequent update of the sim from that point until SU5 robbed me of more performance and I would have to lower my setting farther to keep the sim somewhat playable. I was barely able to get 20 fps by the time SU5 was released.
1 Like
But a simupdate not only adding one thing. Many complained about different graphical issues too at release that may have been fixed in the same update as VR introduced. And that fix may have increased load on our hardware.
I donāt know what exactly in that particular update caused the issue. I honesty donāt even care. I just know that particular update was the turning point for me where performance went from āHELL YEAHā to āā ā ā is going on with this piece of junk?ā. Each subsequent update (both Sim and World, as they all offered code udpates) chipped away at my performance a few frames here and there until I was struggling to play at 20 fps on the same hardware Iād previously been able to get 50+ on.
SU5 restored that performance, and overall itās been good since for the most part in terms of fps. Although SU9 suddenly cut my fps down by a full 10 fps that Iām unable to regain regardless how low I put my settings. Even running potato-grade settings doesnāt restore it and I still see myself CPU limited on a 5950X.
Thereās something terribly wrong with SU9 under the hood.
3 Likes
I think theyāre doing that now. I think thatās part of the problem. Instead of people not noticing anything we have people seeing changes that arenāt there while not seeing changes that are there.
They should just be concrete in telling us what changes they make and do it in writing so people should better be able to differentiate what may be due to changes Asobo made versus changes made on a users machine (settings, Windows updates, driver updates, etc)
Hope they find the issue in su10 then. I also want better FPS but i also not want them to decrease graphics to achieve that.
2 Likes
Preach brother.
1 Like
I have a Series S and it does run smooth and it does have a very limited TLoD. It has less to download and less to draw.
I think the code changes do effect the servers though. I think they are adding more information to the download (donāt have any proof at the moment) and so its logical that an increased TerrainLoD would require both more data downloaded as well as more data rendered/calculated. The Series X just tries to do a lot more. Iāve become glad I only got an S.
What I donāt know/understand is at what point does the improved wind/thermal and CFD model get its information and where are the calculations performed? Is this all just a local process model where the real live weather winds are irrelevant? Wouldnāt really work all that well if it were. If thereās an improved model that does more and requires more information (from the net) and then both calculations and downloaded data increase as TLoD increases, then I can understand why we see stuttering more now than before. Its Asoboās/Sebās improved airflow models. It must be. That is what they are working on, isnāt it?
2 Likes
I had planned to sell my Series S - instead it is now taking pride of place for my MSFS useā¦.
1 Like
Could be, but that isnāt a thing they should degrade either to increase fps. Seb said in a q&a that it didnāt had that much of an effect on performance. I often notice fps drops while moving camera and that feels more of a caching issue that we had pre off-screen caching setting.
Just upgraded my GPU (delivered tomorrow), but after this conversation perhaps it wonāt make much difference? I have a 5600 CPU, which to my understanding has excellent single thread performance.
Iām running at 1440 but perhaps I will increase my render settings to take advantage of my new GPU.
Anyone have any concrete steps to increase GPU utilization and unload CPU?