Performance Issues with Starship – Sporadic FPS Drops
-
Unfortunately I still has this issue randomly coming up even after 1.2 . It only happens to Starship and I believe it has to something with custom avionics.
Recently I upgraded to RTX5090 thinking may be related video memory issue obviously its not . I tried several long hauls routes without any issue. Then intermittent stutter came back - every second micro stutter. It usually happens enroute and all the way to the landing. I program FMS manually typing each waypoint using keyboard. The only thing come to my mind is something to do with waypoint programing.
-
I can confirm this but I can actually reproduce it. Flying from EETN/26 N0302F360 DCT ORTAX DCT LOGNA P31 KEDUX DCT EMGAR Z717 BODLA L733 LOLPA Z14 ERAKU DCT EDAZ/25.
Did it write and it happened twice. Also on 1.2.Funnily enough at the same time when entering briefly touching Russian airspace at Kaliningrad.
-
For me It only happens to BS aircraft. I don't have performance degradation with A2A or other add ons flying 3-5 hours blocks. Performance drop usually doesn't happens until after few flights. I bet my money on some kind of tear and wear code that overtime leaking memory or creates to many background processes that degrades performance
-
For me It only happens to BS aircraft. I don't have performance degradation with A2A or other add ons flying 3-5 hours blocks. Performance drop usually doesn't happens until after few flights. I bet my money on some kind of tear and wear code that overtime leaking memory or creates to many background processes that degrades performance
@sdflyer I would love to believe this too, since that would make it not terribly difficult to track down and solve, but many hours of testing and collaboration with the several people who have contacted me with performance issues in my aircraft have yet to yield a definitive answer. Each case has ended with some simulator or system setting solving the issue, or finding that the issue existed in default aircraft and waiting until the next sim update.
While the failure system is an obvious first place to look, it's also the easiest to test, and we have found no issues there. I also might add that the reason I use JS instead of WASM is because of its resilience to memory leaks, and other such issues. If you had any interest in disabling the failure system to test for yourself, I would be happy to walk you through it.
Starship, being by far my most complex aircraft, exhibited some early performance issues with my beta testers, which had similarly unrelated cures. Most often, a BIOS setting change made a world of difference for many users, and others found that a virus checker was the cause. Whenever I was able to examine performance degradation on specific flights, the symptoms were very specific. Every-other frame would be impacted, and limited to the "manipulators" thread of MSFS. We don't have any documentation on what exactly this means, but I was able to demonstrate that it has something to do with the cockpit interactions, which makes sense, given the name. Unfortunately, this would lead me to believe that it's a simulator level bug, and not necessarily something I can solve. Perhaps there is something that I'm doing to exacerbate it, but even deleting large sections of my interaction code did not have an effect.
Long story short... I have spent lots of time over the years trying to track down performance issues with the small number of users who experience this problem, and the solutions have always been on the system side thus far. This is not an excuse for me to stop looking, but it's merely what I have to report so far.
-
@sdflyer I would love to believe this too, since that would make it not terribly difficult to track down and solve, but many hours of testing and collaboration with the several people who have contacted me with performance issues in my aircraft have yet to yield a definitive answer. Each case has ended with some simulator or system setting solving the issue, or finding that the issue existed in default aircraft and waiting until the next sim update.
While the failure system is an obvious first place to look, it's also the easiest to test, and we have found no issues there. I also might add that the reason I use JS instead of WASM is because of its resilience to memory leaks, and other such issues. If you had any interest in disabling the failure system to test for yourself, I would be happy to walk you through it.
Starship, being by far my most complex aircraft, exhibited some early performance issues with my beta testers, which had similarly unrelated cures. Most often, a BIOS setting change made a world of difference for many users, and others found that a virus checker was the cause. Whenever I was able to examine performance degradation on specific flights, the symptoms were very specific. Every-other frame would be impacted, and limited to the "manipulators" thread of MSFS. We don't have any documentation on what exactly this means, but I was able to demonstrate that it has something to do with the cockpit interactions, which makes sense, given the name. Unfortunately, this would lead me to believe that it's a simulator level bug, and not necessarily something I can solve. Perhaps there is something that I'm doing to exacerbate it, but even deleting large sections of my interaction code did not have an effect.
Long story short... I have spent lots of time over the years trying to track down performance issues with the small number of users who experience this problem, and the solutions have always been on the system side thus far. This is not an excuse for me to stop looking, but it's merely what I have to report so far.
@Black-Square
I’m happy to help . I’m running i9 13900 , gtx5090 and 64G ram . I tried several add ons side by side to compare performance degradation. So far the only trouble I have is with BS that why I assume it has something to do with wear and tear. Perhaps some Aosobo API causing it . -
@sdflyer I would love to believe this too, since that would make it not terribly difficult to track down and solve, but many hours of testing and collaboration with the several people who have contacted me with performance issues in my aircraft have yet to yield a definitive answer. Each case has ended with some simulator or system setting solving the issue, or finding that the issue existed in default aircraft and waiting until the next sim update.
While the failure system is an obvious first place to look, it's also the easiest to test, and we have found no issues there. I also might add that the reason I use JS instead of WASM is because of its resilience to memory leaks, and other such issues. If you had any interest in disabling the failure system to test for yourself, I would be happy to walk you through it.
Starship, being by far my most complex aircraft, exhibited some early performance issues with my beta testers, which had similarly unrelated cures. Most often, a BIOS setting change made a world of difference for many users, and others found that a virus checker was the cause. Whenever I was able to examine performance degradation on specific flights, the symptoms were very specific. Every-other frame would be impacted, and limited to the "manipulators" thread of MSFS. We don't have any documentation on what exactly this means, but I was able to demonstrate that it has something to do with the cockpit interactions, which makes sense, given the name. Unfortunately, this would lead me to believe that it's a simulator level bug, and not necessarily something I can solve. Perhaps there is something that I'm doing to exacerbate it, but even deleting large sections of my interaction code did not have an effect.
Long story short... I have spent lots of time over the years trying to track down performance issues with the small number of users who experience this problem, and the solutions have always been on the system side thus far. This is not an excuse for me to stop looking, but it's merely what I have to report so far.
@Black-Square I have noticed this issue on Xbox and I believe for me it's pretty much been unique to the Starship(I have spent mannnnny hours in the Dukes).
It's different from the performance issues I see on longer flights with other similarly complex aircraft like the PMDG airliners and the Miltech MH60. On these you'll see occasional stutters, freezes and/or the glass cockpit going dark if you're having trouble(these are also well made and optimized aircraft so not saying that's every flight, just the symptoms if something is amiss). On the Starship it's more like the sim rate has very slowly decreased over time and eventually it's noticeable that I'm almost flying in slow motion.
For instance I used the Starship on a group flight and a couple hours in it became apparent that I wasn't keeping up with the group even though I was flying faster than them to try to catch back up
-
On my brand new machines first flight over 3 hours caused performance degradation. I was closely monitoring core temperature and CPU load. Comparing with Baron (pro) that runs core at 61 degrees, Starship after two hours temperature is reaching 68 degrees. With Baron CPU utilization around 14%, Starship 34%. I bet you money Starship code spanning some kind of recursive threads. Longer you go, higher CPU temperature and CPU utilization.
-
Thanks for the report. Agh, I'm so sorry that anyone has to deal with that. I've run Starship for 3-5 hour tests on every computer I have access to, and I've been unable to reproduce this. Yet, we've also done so many experiments, and found fixed things, and yet this seems to still exist somewhere out there. If you happen to be testing this again, my only recommendation is to screenshot the FPS dialog, so we can see if the issue still appears to be in the Manipulators thread, whatever that means.
-
Thanks for the report. Agh, I'm so sorry that anyone has to deal with that. I've run Starship for 3-5 hour tests on every computer I have access to, and I've been unable to reproduce this. Yet, we've also done so many experiments, and found fixed things, and yet this seems to still exist somewhere out there. If you happen to be testing this again, my only recommendation is to screenshot the FPS dialog, so we can see if the issue still appears to be in the Manipulators thread, whatever that means.
@Black-Square
I tested on brand new PC. So there can't be coincidence. I've been flying Baron Pro for few weeks. I fly mostly two hours hauls or 3-4 hours long hauls. Baron run smooth without any issues. I know some code is being reuse that originally pioneered in Duke and Starship to bring Baron and Bonanza up to pro. I'm guessing it something unique to Starship that causing it. I'm guessing navigation system or radar.Here are some data that may help:
My PC: Ryzen 9 9950X3D, 64G RAM, RTX 5090
My last Starship route: KHQM/06 N0256F170 DCT HQM V112 ILWAC T257 ONP V452 FMG T274 LIDAT/N0285F370 DCT BASIC DCT KLAS/26LMSFS 2024 settings:
image url)
image url)
image url)
