Performance Issues with Starship – Sporadic FPS Drops

MD82

I believe C-State Control needs to be enabled?

jmarkows

The BIOS pic has me wondering: the people with FPS issues, could there be thermal throttling at play? My newest build has a 14th gen i9, and for those who follow tech is the Intel cpu that was famously self-destructing on launch. I came in late, but had some issues with stability that were resulting in my CPU asking for (and getting!) upwards of 500A.

I have since become keenly aware of CPU voltages and temperatures lol, but with some BIOS configuration it's been drastically reduced.

Black Square

@rafgath Sorry, FMS No. 1.

rafgath

@Black-Square Another flight and another FPS drop (maybe because it's exactly the same flight every time). Please see the video link below. Luckily, I didn't have to click too many circuit breakers, as everything returned to normal after turning ND No. 1 off and back on.

https://youtu.be/4aDTWlVdmPM

Black Square

Very interesting! That's admitted a little strange, since ND 2 runs the exact same code... This might take a little more experimentation. I will look through things later and see what I can come up with, but on your next flight, try configuring ND 2 to show exactly the same as what ND 1 is showing, and see if the problem is exhibited by both units. Thank you for the help!

Wael

I’m having a very similar issue.

After about an hour of flying, I get a major FPS drop — from 160 down to around 40 — and GPU usage drops to 0%–30%.
When I press ESC and pause the flight, GPU usage returns to normal, but as soon as I resume, the performance drops again.
Everything runs perfectly at the start of the flight — this only happens mid-flight.

Specs:
• CPU: AMD Ryzen 7 9800X3D
• GPU: RTX 5080

rafgath

@Black-Square Okay, I'm uploading another video showing the FPS drop. As you can see, only disabling ND No. 1 helped. Turning it back on doesn’t cause the drop either. But if we look at all three videos I uploaded, they’re all from the same route, and the issue generally always occurs around the same point — just before reaching the ROGAL point. Maybe it’s a coincidence, maybe not.

https://youtu.be/mEYfGcpgg50

Black Square

Thanks for showing exactly what I have been asking to see! I finally have a chance to spend some time on this later tonight. I will let you know what I find. At least you can continue your flight to landing after pulling one circuit breaker. I know that's far from ideal, but I'm glad there is some solution that at least gives you the satisfaction of a smooth landing while I work on the problem.

Did you try the BIOS change suggested above? Specifically enabling C-State Control? It would be interesting to see if that could change the outcome, now that we've collected some data.

rafgath

@Black-Square Yes, I enabled C-State Control, but it's still the same. At least I can continue the flight by turning the circuit breaker off and on . Thanks for looking into the issue.

edit: I discovered something new — even if I turn off ND No. 1 and turn it back on, the FPS starts dropping again after a few dozen seconds. So I have to turn off the breaker again, but once I turn it back on, the FPS drops again after several dozen seconds. Unfortunately, I don’t have a recording of it because something went wrong and it didn’t record.

I also discovered that the FPS starts dropping for the first time between 55 -60 minutes of flight.

MD82

For me, enabling C-State was like Gandalf tapping my PC with his cane. AB SO LUTE night and day. Thank you for this suggestion. I was going crazy with the cruise stutters with all hardware loads under 70%.