Game Discussions Star Citizen Discussion Thread v12

Yeah, about 50 FPS, i'm limited by the GPU which is loaded at 98%.

s9W7FJv.jpg

Are you sure that's as low as you drop? Because you're definitely doing better than others with comparable kit.

I actually forgot to put your Ultra High settings in before, but it hasn't changed things much. The telemetry still gives an average FPS that's lower than the 50FPS you're seeing there.

xaIsUUx.png


Are you sure you don't dip lower still on foot in the major cities or comparable?

Either you're super lucky with your PC configuration (there does seem to be a lot of variability between comparable rigs), or you're just not playing in the same low performance areas as others seemingly.

Well, either that or the telemetry page is boned ;). (Which wouldn't bode hugely well for the game networking, as Turbulent are expanding their role there ;))
 
Are you sure that's as low as you drop? Because you're definitely doing better than others with comparable kit.

I actually forgot to put your Ultra High settings in before, but it hasn't changed things much. The telemetry still gives an average FPS that's lower than the 50FPS you're seeing there.

xaIsUUx.png


Are you sure you don't dip lower still on foot in the major cities or comparable?

Either you're super lucky with your PC configuration (there does seem to be a lot of variability between comparable rigs), or you're just not playing in the same low performance areas as others seemingly.

Well, either that or the telemetry page is boned ;). (Which wouldn't bode hugely well for the game networking, as Turbulent are expanding their role there ;))
My FPS as a comparison tend to be better than the heatmap would suggest they should be as well. I'm in the same range as Intrepid consistently despite having a poorer CPU score (i7 8700k@ 4.6Ghz) but the same GPU (RTX 2070 super). I do have 2x 2070's with an sli bridge but SC doesn't use the extra card. I also have 64GB of RAM...no idea if that makes any difference. 🤷‍♂️ I'm running at 1080p though which doesn't hammer the GPU at all, just the CPU. CPU usage is up at 65-90% where as GPU usage is as low as 30% regardless of location.

There are no ultra settings in SC...besides the fact the main graphic menu settings have little effect on performance or visuals anyway, the highest graphic settings are 'very high'...which mine have been set at since 2016. I do run a custom user.cfg file with various console commands that effect visuals, but only as optimisation or improvement. We still have no firm information from Ci¬G if adding that user.cfg file actually does anything at all since the addition of OCS...some backers say it does, some say it doesn't, Ci¬G have said nothing at all. I've been running the custom user.cfg since my time with evocati...back at a time when we had to run one to get the ETF test builds to work at all.

I've tried running with and without it but SC being SC and completely unoptimised with performance varying from patch to patch anyway...hard to give a definitive yay or nay :rolleyes:
 
Last edited:
There are no ultra settings in SC...besides the fact the main graphic menu settings have little effect on performance or visuals anyway, the highest graphic settings are 'very high'

Yeah was just going by Intrepid saying 'highest settings' there.

My FPS as a comparison tend to be better than the heatmap would suggest they should be as well. I'm in the same range as Intrepid consistently despite having a poorer CPU score (i7 8700k@ 4.6Ghz) but the same GPU (RTX 2070 super).

Yeah wonder if the whole thing’s bugged, or if it’s just deeply variable between comparable rigs. It’s not unusual to see guys on monster rigs rock up on reddit complaining of woeful performance. I’d guess there’s more at play than the metrics we’ve got anyway.

(It clearly needs an SSD entry ;))
 
Yeah was just going by Intrepid saying 'highest settings' there.



Yeah wonder if the whole thing’s bugged, or if it’s just deeply variable between comparable rigs. It’s not unusual to see guys on monster rigs rock up on reddit complaining of woeful performance. I’d guess there’s more at play than the metrics we’ve got anyway.

(It clearly needs an SSD entry ;))
It doesn't take into account what resolution folks are running SC at...you run it at 4K it'll run like crap no matter what you're running it on...some folks just refuse point blank to recognise any resolution under 4K these days. I can quite happily run everything I play at 1440p on the HDR monitors...but it all looks just fine and dandy at 1080p for me...old eyes I guess, my vision was so used to seeing games run on a cathode ray tube, the 30" HDR flat panels are just dandy :)
 
Last edited:
It doesn't take into account what resolution folks are running SC at...you run it at 4K it'll run like crap no matter what you're running it on...some folks just refuse point blank to recognise any resolution under 4K these days. I can quite happily run everything I play at 1440p on the HDR monitors...but it all looks just fine and dandy at 1080p for me...old eyes I guess, my vision was so used to seeing games run on a cathode ray tube, the 30" flat panels are just dandy :)

Oh it does actually cover that now. I’ve got Intrepid’s 1440p settings plugged in there.

I’m guessing that ram might be a useful one to have in there though. (It seems like a consistent thread amongst guys who don’t have major issues to have tons of ram). And genuinely the SSD thing. We know the hitching it smoothes out does up FPS averages too. (Plus it means buildings tend to appear conveniently when needed ;))

(PS I’m doing all this digging as someone who’s rarely given a crap about games being over 30fps for my own gaming ;). And I’m happily blundering around in EDO in VR in about the same, which is borderline foolish ;). But it is mad seeing top end gear only grinding out these kind of numbers in SC, with the content still ‘early days' etc. Also kinda intriguing to untangle what makes it fine for some and a fever dream for others ;))
 
Oh it does actually cover that now. I’ve got Intrepid’s 1440p settings plugged in there.

I’m guessing that ram might be a useful one to have in there though. (It seems like a consistent thread amongst guys who don’t have major issues to have tons of ram). And genuinely the SSD thing. We know the hitching it smoothes out does up FPS averages too. (Plus it means buildings tend to appear conveniently when needed ;))

(PS I’m doing all this digging as someone who’s rarely given a crap about games being over 30fps for my own gaming ;). And I’m happily blundering around in EDO in VR in about the same, which is borderline foolish ;). But it is mad seeing top end gear only grinding out these kind of numbers in SC, with the content still ‘early days' etc. Also kinda intriguing to untangle what makes it fine for some and a fever dream for others ;))

The truth is you do need to run it off an SSD, you try and run SC from a Spinny Disk and you will be in a whole world of hurt.

RAM, 16GB minimum, having 32GB can help smooth it out a little but doesn't make a huge difference.

The worst i have even seen it was during the Xeno Threat event, that was 30 FPS, still playable tho.

Source: https://www.youtube.com/watch?v=AoA4IrPTJYI
 
But what if it's not what we paid for ?
I paid for what was promised in the KS. Shall i list again the promises made back then ?
Or in short, where are my 100 systems, with proc gen planet surfaces etc. ? Where's the working MMO engine that supports thousands of players in the same location ?
Where are the capital ships and actual stations other than pilot ? Where's the reclaiming, refuelling ? Where's exploration, scientific experiments, plants growing and such (they sold a very expensive ship for that..) ?
Where's the trading, with offer/demand and production chains ?
Where are the actual NPC with an AI ? not place holders T-posing on chairs, or other place holders on a script loop that actually breaks down after a short while...
Where's Squadron 42 ? Some people backed the project just to have a replacement for Wing Commander, which is fair.
I could continue for a while...

I backed the project in hopes to have just a replacement for Jumpgate. This currently falls very far from what JG ever was, and the comparison really hurts SC. The only game loop currently in place is mining, and even then the subsequent offer/demand trade loop is still not there and is dreams.txt, 9 years in.

Plus the ever present issues, that will never be fixed:
The physics engine is a joke. Networking is the basic LAN party code that was given with CryEngine, and collapses completely whenever there are more than 20 people in the same location, also has terrible desync and no lag compensation, in 2021 ! Most ships over fighter size are filled with place holder assets that are just there for decoration. Lifts, ramps and ladders are still completely buggered (and somewhat lethal) to a point it's become a meme, and they have been like that forever. Player movement is horribly janky and too fast for what is supposed to be a simulation (plus it confuses the network code which struggles to follow...). Ship flight model is still a simplistic arcade thing with zero skill ramp, mav thrusters have infinite power, and said ships have the weight and density of a beach balloon (see side effects of them being blown up by wind, or flipped over by a vehicle or a cart). One player can crash the entire server by trying to do something the server doesnt like, which is quite a lot of things actually. The game runs like crap on a high end, watercooled 10900K/RTX3090 beast.


He did not. He's just saying that words have a meaning... Changing the meaning of words to try and support your argument is just as sterile as comparing a game to a piece of wood.
It has been hours since that post and it hasn't been challenged by the defenders. It's clear that the defenders only have the fun argument to defend their position now.
 
Are you sure that's as low as you drop? Because you're definitely doing better than others with comparable kit.

I actually forgot to put your Ultra High settings in before, but it hasn't changed things much. The telemetry still gives an average FPS that's lower than the 50FPS you're seeing there.

xaIsUUx.png


Are you sure you don't dip lower still on foot in the major cities or comparable?

Either you're super lucky with your PC configuration (there does seem to be a lot of variability between comparable rigs), or you're just not playing in the same low performance areas as others seemingly.

Well, either that or the telemetry page is boned ;). (Which wouldn't bode hugely well for the game networking, as Turbulent are expanding their role there ;))

Performance is also very server dependent. If you get on a "good" one, you'll have much better frames. A "bad" one, the opposite.

If every time you get on a "bad" one you server hop, you'll never really experience it.

Was playing last week in the PTU when I posted a video of the stations not loading in. Was on a server that only had 30 people on it. Most of those people were at Port O fighting. It was a really good experience (relatively). The frames were good, the usual desynch was not there, the aiming pips were working much better etc. Dont really know why it was, but seems plausible it was low population and a bunch of assets not actually loading on the server.
 
Performance is also very server dependent. If you get on a "good" one, you'll have much better frames. A "bad" one, the opposite.

If every time you get on a "bad" one you server hop, you'll never really experience it.

Was playing last week in the PTU when I posted a video of the stations not loading in. Was on a server that only had 30 people on it. Most of those people were at Port O fighting. It was a really good experience (relatively). The frames were good, the usual desynch was not there, the aiming pips were working much better etc. Dont really know why it was, but seems plausible it was low population and a bunch of assets not actually loading on the server.
Is there a way yet to quickly determine "I'm on a good server here, I'll stay" and/or "I'm on a bad server here, I'll hop"? I try to test out each patch as they go live and have always just accepted the servers as they come.
 
The worst i have even seen it was during the Xeno Threat event, that was 30 FPS, still playable tho.
It was down to about 15fps (*). Also PvP during that event was nigh impossible unless your target was stationary - server desync was about 2000-3000ms, when shooting at a moving target, by that time they would be somewhere else entirely. Absence of "rubber banding" explicitly points at the very primitive network code with zero lag interpolation (which to be fair, was not needed for LAN parties when CE was written) - player ships would just teleport around from place to place.
And that's for an event that's taken about 6 months of effort for the majority of CiG devs, to a point they had to push a lot of features down the line to make place for it, as it was entirely hand crafted, all this on backers money. During that event I did shoot a few torps at Idris piloted by brain dead AI, escorted by totally passive AI fighters with also zero tactical sense (they would just engage the nearest threat, leaving the bombers alone...), went back to base to refuel/repair only to find the refuel/repair bug which has plagued SC for about 6 years now. How many centuries to fix that one ? Seriously, this "event" has cost a few millions $ for having an NPC piñata spawn every hour or so. Korean MMO devs are laughing all the way back to their bank.

(*) i9 10900K, RTX 3090, 32GB RAM, running on SSD of course. Doesnt matter as this was the fps cap for everyone.

Is there a way yet to quickly determine "I'm on a good server here, I'll stay" and/or "I'm on a bad server here, I'll hop"? I try to test out each patch as they go live and have always just accepted the servers as they come.
Kind of, just after you spawn, the responsiveness of the lift buttons is a first hint. You can turn on server debug info on top right of the screen too (forgot the command..) it'll show packet loss and such, if you see any packet loss it means the server is dead or about to die. Then have a look in a specific direction (for instance at New Babbage look at the city from the windows) and check your fps - of course you have to know which is low and which is normal, so a few runs will be needed.
Then there's the ship spawn itself, from when you pick your ship on the terminal and when the terminal says it's ready - if it takes way too long, server is getting "tired"...
 
Last edited:
Kind of, just after you spawn, the responsiveness of the lift buttons is a first hint. You can turn on server debug info on top right of the screen too (forgot the command..) it'll show packet loss and such, if you see any packet loss it means the server is dead or about to die. Then have a look in a specific direction (for instance at New Babbage look at the city from the windows) and check your fps - of course you have to know which is low and which is normal, so a few runs will be needed.
Then there's the ship spawn itself, from when you pick your ship on the terminal and when the terminal says it's ready - if it takes way too long, server is getting "tired"...
Perfect, cheers (y)
 

Viajero

Volunteer Moderator
It has been hours since that post and it hasn't been challenged by the defenders. It's clear that the defenders only have the fun argument to defend their position now.

Well, patrons here sure have a bit more on their life plates than just responding to an already mildly peripheral topic about a video game development in a forum? Also on ocasion you may just not feel to continue an argument when all the points have been made and all that remains is to rehash and repeat them. An absence of response is just that, an absence of response.
 
Last edited:
The truth is you do need to run it off an SSD, you try and run SC from a Spinny Disk and you will be in a whole world of hurt.

RAM, 16GB minimum, having 32GB can help smooth it out a little but doesn't make a huge difference.

The worst i have even seen it was during the Xeno Threat event, that was 30 FPS, still playable tho.

Source: https://www.youtube.com/watch?v=AoA4IrPTJYI

Yeah this is where the bizarre variability comes in though.

Here's a guy on GTX 3080 / 5900X / 32GB / SSD averaging 16 FPS:

Source: https://www.youtube.com/watch?v=s4UFU9KEP4o&t=469s
 
Here's a guy on GTX 3080 / 5900X / 32GB / SSD averaging 16 FPS:
That was the same for everyone during Xenothreat.
Outside of this event, a high end rig will get 70+fps in space, down to about 30fps in Area 18. Again there seems to be a hard cap on frame rate in A18, showing the really poor optimization of the 3D engine. And before cultists would answer "but they have not optimized the engine yet ! It's alpha !" I would point out that CiG bought a complete engine back in 2011 so they would not have to do this at all as this whole process was supposedly done by Crytek prior to that, also it was the argument advanced by supporters of this engine choice ("but it would be too expensive to develop an in-house engine!" - guess what they're doing now...).
 
Well in fairness to Intrepid, their vid there does seem to show 2 Idrises and FPS hanging around the 20-30 range. (With one killer drop to 5FPS when getting too close to the remaining Idris)

Does seem like a somewhat miraculous outlier though ;)
Yeah i think it's likely related to these factors:
  • how many players are around
  • are the Idris active (the turrets seem to really tax the servers)
  • are people entering the area (streaming assets in = big drop in performance)
 
Yeah i think it's likely related to these factors:
  • how many players are around
  • are the Idris active (the turrets seem to really tax the servers)
  • are people entering the area (streaming assets in = big drop in performance)

Ay, it was interesting to see the steady attrition from 30FPS down towards 20FPS after arrival (he'd been off to re-arm).

Going by the chatter that group had been happy to server hop prior. So I guess in a server hopping world there's always going to be a game of 'Find a good one, leave when it starts to fill up and things deteriorate' ;)
 
Back
Top Bottom