Exiting Lift 100% CPU usage | FPS drops to 20-25 fps

Just before I post this to the fault tracker was wondering if anyone else has encountered this issue.

Prelude
I run the game at 1440p / 60 FPS as I record /stream so my CPU usage sits at around 15% // GPU sits at around 30-50%
Game runs beautifully bar this one issue and if the bug occurs it kills my session completely as I have to recompile the shader/graphics table to fix the issue.
I'll put the obligatory "every other game I have runs without any issues" also

Issue
For a while now, RANDOMLY, when exiting concourse to go to the loading bay, when the lift doors open my GPU usage spikes to 100% but frame rates drop to 20-30 fps,
I'll stand just outside of the lift and after 2-3 mins it clears and FPS / GPU usage return to nominal figures but the game doesn't feel smooth anymore.
I board my ship to cockpit and FPS drops again from 60 to around 25, almost as if some kind of texture /shader wasn't loaded when transitioning from concourse to loading bay.

On testing, if this bug occurs, it then affects other areas of gameplay in particular on foot CZ, on foot settlement missions, massive fps drops (down to 20-25) and constant stuttering
After persevering and completing 5 CZ's, decided to call it a day and return to the station to collect my bonds as it was starting to give me a headache / feeling nauseous.

I then googled a bit and decided to test the fix of another issue, which was to rebuild the shaders / cache' etc by deleting the graphics folder in appdata/frontier developments.

Relogged into the game after deleting the graphics folder, obviously it recompiled the shaders at login etc, I then enlisted for the same on foot CZ, arrived at the CZ and it was smooth, a few little FPS drops here and there (5 fps) etc but felt really smooth in comparison to what it was after the lift bug.

This was also present on Windows 10

Things I've done to try and remedy this
• Fresh install of Windows 10 and 11
Fresh install of EDO
• DDU clean of Nvidia Driver
• Virus Scans
• Running a cleanboot option with zero added peripherals
• Using an older Nvidia driver 451.71

Background info / specs etc.
OS
: Windows 11 (Fully up to date)
Drivers: Fully up to date via mobo site
CPU: 5900x (running stock PBO)
GPU: RTX 3070 (running stock)
MOBO: X570 Auros Elite
RAM: 32gb (2x16gb 3600mhz)
Monitor: 27" 1440p 144hz 1ms
BIOS: Latest Update inc AMD Fix for fTPM
Gsync off
HAGS off


Graphic settings
ULTRA preset
Upscaling: Normal
Fullscreen
Vsync off ingame (On via NVCP)
FPS Limiter off ingame (Capped to 60 FPS via NVCP)

Appreciate any input or more so if anyone else has encountered this issue.
 
Last edited:
I've run into the same bog and in different stations. I'm not sure what it means but it's always in the lifts area starting at the threshold of the lift/hanger boundary all the way to the lift doors...
...but it's not "all the time".

o7
BIZZ

edit: I have nearly the same system you do only on a laptop.
 
when first starting the game from the main page and you get either the wireframe of either a man spinning or ship spinning
how long on avg does it spin before the game starts this should be around 10 seconds or lower ideally
if it takes much longer your problem is partly connection related imo...……………..
sometimes this is just the servers bogged down due to demand and basically jack :poop: you can do about iT
the other thing I would recommend is a periodic deleting of the gpuworktable.xml in your products folder just to clear the cache of any klingons
now let simmer and add in a little poorley optimized transistion areas and your set for an interesting play session or lack of one
good luck an o7
(last but not least sometimes the latest driver is just to recent for the poor devs to keep up with so try older versions 466.77 being very stable )
 
Last edited:
Can't duplicate this on my end, but it sounds like VRAM contention, which is uncommon with 8GiB parts, but not unheard of at ultra.
 
Just before I post this to the fault tracker was wondering if anyone else has encountered this issue.

Prelude
I run the game at 1440p / 60 FPS as I record /stream so my CPU usage sits at around 15% // GPU sits at around 30-50%
Game runs beautifully bar this one issue and if the bug occurs it kills my session completely as I have to recompile the shader/graphics table to fix the issue.
I'll put the obligatory "every other game I have runs without any issues" also

Issue
For a while now, RANDOMLY, when exiting concourse to go to the loading bay, when the lift doors open my GPU usage spikes to 100% but frame rates drop to 20-30 fps,
I'll stand just outside of the lift and after 2-3 mins it clears and FPS / GPU usage return to nominal figures but the game doesn't feel smooth anymore.
I board my ship to cockpit and FPS drops again from 60 to around 25, almost as if some kind of texture /shader wasn't loaded when transitioning from concourse to loading bay.

On testing, if this bug occurs, it then affects other areas of gameplay in particular on foot CZ, on foot settlement missions, massive fps drops (down to 20-25) and constant stuttering
After persevering and completing 5 CZ's, decided to call it a day and return to the station to collect my bonds as it was starting to give me a headache / feeling nauseous.

I then googled a bit and decided to test the fix of another issue, which was to rebuild the shaders / cache' etc by deleting the graphics folder in appdata/frontier developments.

Relogged into the game after deleting the graphics folder, obviously it recompiled the shaders at login etc, I then enlisted for the same on foot CZ, arrived at the CZ and it was smooth, a few little FPS drops here and there (5 fps) etc but felt really smooth in comparison to what it was after the lift bug.

This was also present on Windows 10

Things I've done to try and remedy this
• Fresh install of Windows 10 and 11
Fresh install of EDO
• DDU clean of Nvidia Driver
• Virus Scans
• Running a cleanboot option with zero added peripherals
• Using an older Nvidia driver 451.71

Background info / specs etc.
OS
: Windows 11 (Fully up to date)
Drivers: Fully up to date via mobo site
CPU: 5900x (running stock PBO)
GPU: RTX 3070 (running stock)
MOBO: X570 Auros Elite
RAM: 32gb (2x16gb 3600mhz)
Monitor: 27" 1440p 144hz 1ms
BIOS: Latest Update inc AMD Fix for fTPM
Gsync off
HAGS off


Graphic settings
ULTRA preset
Upscaling: Normal
Fullscreen
Vsync off ingame (On via NVCP)
FPS Limiter off ingame (Capped to 60 FPS via NVCP)

Appreciate any input or more so if anyone else has encountered this issue.
Although listed as a VR issue, several of the confirmations are from flatscreen users experiencing the same problem - the report is Acknowledged by FDev and can be voted for here:

 
I experience this a little bit, but not as bad as you OP. Especially noticable on my fleet carrier when entering the hangar. Mostly its just a drop for a second or two though.
 
Cheers for the replies,

Forgot to mention that prior to the latest EDO update, I wasn't experiencing this issue.

Thanks @Arioch
I've added my notes to the tracker.

@Morbad
you might be onto something there tbh, as noticing my VRAM usage hitting the top end of what my card can handle. (7.5gb process), don't ever recall it being that high in previous performance tests, though could be wrong.

I've tuned down the preset to HIGH, will test tonight.
 
Something interesting to note regarding the VRAM etc

Sitting at this station I noticed my GPU usage was 90% + (still capped at 60 , so shouldn't go above 50% normally)
I hadn't disembarked yet, went to grab a drink whilst auto-docking came back to see these stats
Quit the game, deleted appdata/graphics folder
Reloaded and was stting at 41% usage now.

Untitled-2.png


 
Last edited:
usage in stations is always higher than expected, thats been the case for years even in horizons. when i would set up VR i always tested while in station, theory is getting it to work there they will work everywhere.

i have noticed, the GPU gets more work to do dependant one where the station is. if the station is near to asteroids then for some reason while in the station it makes a difference. maybe the game is still drawing them even though we cannot see them from inside the station.

i play at 4k on a 2080ti, its nearly maxed out 95 or 96% when in a station. low 35% to 40% when in space
and of course 99 to 100 when in station on foot in some places on foot in station fps drops to 49
 
Last edited:
Adding to this,
Still encountering the issue, even with preset High / Medium / Low

One thing that did remedy it almost immediately was to open the Galaxy Map , wait 10 secs, close galaxy Map.
Odd.
 
OP, you have a 5900x which takes a max ram of 3200mhz., but your system ram is running at 3600mhz., which I assume is XMP driven.

Nothing wrong with running ram that high, but unless you overclock the cpu to handle the increase, its useless.

I found that lowering the ram in bios to 3200mhz. to match the max ram the cpu can handle without overclocking the cpu, seems to give me somewhat better FPS and lower cpu usage.

I use a AMD 5700x. I set my upscaling to normal with smaa, and supersampling to .5 which gives me higher FPS in stations, but I also run at 5120x2880 at 120hz., medium settings and downscale (VSR), as opposed to upscaling. I just wish there was an "off" selection for normal upscaling. I'm also using a Resizable Bar of 8GB and 16x bus from the gpu.

I can maintain 60-70 FPS in ground/space stations and 110-120 FPS in space with gpu usages in the 60-70% max range.
 
Last edited:
@Paul Eddington Thanks for the reply, you'll have probably noticed a better response at 3200 due to lower CAS Latency

Understand you're trying to help and thank you for that but not sure what you mean by overlocking your CPU, at no point have I ever needed to overclock a CPU to fully utilise RAM speeds that have passed QVL of a motherboard.

To note though, I've have already tried downclocking with manual timings to 3200mhz, in fact, last nights session was ran on 3200mhz CL 16.
The issue occurred twice during a 3 hour playtime.

Thanks for your setup tips also, but I have zero issue with performance 99% of the time, steady 60FPS on Ultra 1440p until this lift bug/transition occurs.
(to repeat this happens on High/Med and Low Pre-sets too) so I'm not entirely convinced now it's VRAM now either, on client side.

Again this wasn't happening prior to the last patch so I'll leave with the tinkering for now, seeing as there is an active fault-tracker listed for this issue already.

o7
 
Can you induce the issue by repeatedly using the lift to/from the concourse, or does it appear to be entirely random?

Appears random but also only appears to be the bigger stations.
Outposts / Planetary stations I've yet to encounter it
(although I did have big drops on one planetary station when looking directly at the bar, face away from bar, normal figures)

Good shout though, I'll revisit one of the stations I had issue on and attempt this.
Will advise in due course.
 
Still happening in Update #12

See video (will render to 1440p once YT has done it's bizz)
Skip to 02:50 if you want to skip the preliminary testing

Source: https://www.youtube.com/watch?v=-zTepOUZdc8


Attempted to induce the issue by repeatedly using the lift. Nothing on this occasion but in the same area during the same video managed to capture the same issue but this time it happened when boarding my ship.

Frame rates drop, GPU usage spikes
@Morbad
 
Seems like it loads the interior of the ship when you first board it then doesn't stop rendering it until you exit to main menu.
 
When this bug occurs, I even tried to switch to 640×480 resolution and still I had 30 FPS only :-D

Other thing, I've noticed there's a really high usage of Frame Buffer (probably caching or something,) normally I have 10--15 %, here you can see around 70 :-D
front_buffer_02.jpg
 
I don't think this is a bug, just a loading screen.

When I enter the concourse lift to go to the hangar there is a significant spike in CPU load, presumably as the hangar is being loaded, which abates soon after entering the hangar. The hangar also has higher 'frame buffer' utilization (which should more appropriately be called memory controller load) than most other areas, but that isn't particularly unusual itself; it just implies the hangar is a memory intensive area.

Source: https://www.youtube.com/watch?v=jO59_SYdiQo
 
Last edited:
Cpu spike when getting out of the lift is texture streaming at work.
The higher you set texture quality, the more video ram will be used during that process (and so the cpu spike will be longer).
It seems to me that the last steps in texture quality don't change texture quality that much. But the ram usage gets pretty big (i'd think the game is loading more textures as cache).
 
Back
Top Bottom