#dedicated-servers

1 messages Β· Page 226 of 1

proper magnet
#

does the name matter then?

vocal jungle
#

Satis-one, Satis-two, Satis-three

#

yes for your sanity later

proper magnet
#

perfect.

ivory hare
#

but uhh how does one make a dedicated server

vocal jungle
#

@ivory hare have you read the wiki first

proper magnet
#

now that that is done, it should just work?

vocal jungle
#

if you did it correcly

#

you should be able to connect from the external ip

proper magnet
#

for address do I now put my PC ro router IP?

vocal jungle
#

are you hosting the server from your pc or another pc on the network

paper ridge
proper magnet
#

from my pc

vocal jungle
#

and you pointed the forward port to your computer

proper magnet
#

yeah

#

yes

vocal jungle
#

so you should beable to connect via your local ip or you excternal ip

proper magnet
#

does external mean my routers IP?

vocal jungle
#

no

#

your router has 2 ip a internal so like 192.168.1.1 and a external 98.58.165.45

#

you want the external

#

you can get that from the router or from googling what my ip

proper magnet
#

ok. the option to authenticate does not prompt

vocal jungle
#

did you click on the setup tab

proper magnet
#

setup?

vocal jungle
#

sec

ivory hare
#

so do i just download the dedicated server thing and open it? im confused on itr

vocal jungle
#

when you click on the server settings tab at the top

#

did you read the wiki

paper ridge
#

we cant understand anything from the wiki

#

thats why we ask here

vocal jungle
#

where do you get lost

smoky tangleBOT
#
Reaction Roles - Message Information
Message ID

946137320569061546

Channel ID

902621736602861588

Guild ID

370472939054956546

Message Link

#dedicated-servers message

paper ridge
#

how this thing works

wha.PNG
vocal jungle
#

you install via epic or steam

paper ridge
#

epic

vocal jungle
#

whats it do when you run it

paper ridge
#

runs some commands

#

and stop

ivory hare
#

ok i was downloading the steamcmd thing and i opened it to update it and it says steamcmd needs 250mb of storage to update but there is like 42gb of storage on that drive

vocal jungle
#

is steam cmd on that drive

ivory hare
#

ye

vocal jungle
#

you sure

ivory hare
#

ye

vocal jungle
#

what is the full path to steamcmd ?

ivory hare
#

thispc/Windows(c:)

paper ridge
#

what does this mean

wh.PNG
ivory hare
#

so what should i do

vocal jungle
#

do you only have a c drive ?

ivory hare
#

no i have a c, d, and e drive

vocal jungle
#

so is it C:\steamcmd ?

ivory hare
#

yes

#

i didnt put it in a folder though, i just put it in my c drive

vocal jungle
#

so its c:

ivory hare
#

ye

vocal jungle
#

with a steamcmd.exe in c:\

ivory hare
vocal jungle
#

you cant have it in the root of c

ivory hare
#

so what do i do with it

vocal jungle
#

make a folder called steamcmd

ivory hare
#

does it matter that the steamcmd thing says old

vocal jungle
#

install steamcmd in the folder you made

ivory hare
#

so just delete it and reinstall it or can i just move it

vocal jungle
#

you might beable to move it

#

but i am not 100% on that

ivory hare
#

ok i just reinstalled it

#

its workin now

vocal jungle
#

cool

ivory hare
#

help

#

i just put in the start servers.bat thing in my dedicated server folder and when i go to edit it it says that my pc cant run it

#

never mind

vocal jungle
#

make sue to edit it to match you setup

#

or you could endup with more then one install

ivory hare
#

im not sure how to install ports

#

or use them

vocal jungle
#

you need to forward port 15777,15000,7777 from your router to the pc hosting the server

ivory hare
#

?

#

how

vocal jungle
#

you have a mic ?

ivory hare
#

yes

vocal jungle
#

sent you a invited

#

in you dms

ivory hare
#

@vocal jungle

unreal charm
#

Bump. Would be nice to gather some statistic to find out if it's a common linux issue or just don't host on ubuntu or something. So far we have just three people: me, #dedicated-servers message and the poster from QA site.

cloud flint
#

Soooo, after a lot of fiddeling with the VM with adding more storage to it (because I only had 6GB and that was not enough for the validation because it needs at least enough for a new installation) I finally was able to do the validation process. After I started it, the first line was Update state (0x5) verifying install, progress: 34.81 (2250421785 / 6464719048).

After that it started immediately with downloading an update (or just the server again because something was faulty). After that it ran through the veryfying with no problems. I also saw that the starting service includes an steamcmd verify but I think it just didn't do it because it didn't have enough space left.

I won't be able to test it out today but I'll just hope that I'll be able to join the server tomorrow without restarting it.

uncut copper
#

Question, is there anyway to prevent items on conveys from disappearing at certain points and then returning later down the conveyor?

kind phoenix
#

does the linux Steam CMD version of Satisfactory server store save files differently? I read the wiki, but ~/.config/Epic/FactoryGame/Saved/SaveGames/server is not what I'm seeing. I do have satisfactory/1690800/FactoryGame/Saved/ but it only contains Config and Logs folders. I have no idea where my save game file is.

orchid vault
#

yes server saves local

kind phoenix
#

where? game only has access to one folder above the satisfactory folder

hallow flax
#

is this chat for all dedicated servers or just dedicated servers that are on the experimental version?

orchid vault
#

somewhere on the hard drive.. try searching

#

this is room for dedicated servers

kind phoenix
# orchid vault somewhere on the hard drive.. try searching

I have done this, and I've even searched folders higher up on the folder tree. I have searched all the folders that this game has access to. I've gone so far as to move the install folder, and install again and start a new server to confirm there are no .sav files generated by my steam CMD install.

#

I've searched the log files, which reference a local save, but do not reference the location, nor do they throw an error saying that they can't write to disk.

orchid vault
#

its in spot where configs and logs go or where u installed the server

#

or search should find the files if u do it right

hallow flax
#

when I look at the console for the server I often see "TimeStamp expired" as well as other yellow text. When this happens, its tied to when ever anyone on the server is getting lag. (Not trying to explain that to anyone here, just pointing out the observations I have made) This didnt happen much when we first started the server, any way to fix this?

orchid vault
#

its desyncing

#

and waiting for slow data

kind phoenix
orchid vault
#

make sure network quality setting is maxxed on server and all clients playing on the server

#

good luck heavy

hallow flax
#

also, I should probably mention that the first 9ish times I closed the server, I didnt do it in a safe way. (I properly close it now) I just pressed X on the console. Will that have any lasting effects on the server? And can I fix that.

orchid vault
#

hittin x is like anything on ur computer

#

it closes without saving

kind phoenix
#

Well then. used the server.savegame command in console and got this:

Saving game to ''
LogServerConnection: Saving game to ''
LogGame: World Serialization (save): 0.170 seconds
LogGame: Compression: 0.239 seconds
LogGame: Write To Disk: 0.001 seconds
LogGame: Total Save Time took 0.415 seconds

vocal jungle
#

you tying to find your save there Heavy

kind phoenix
#

yes. Lost my server with about a month of play time, and want to make sure it doesn't happen again.

vocal jungle
#

windows or linux server

kind phoenix
#

Its running inside a games hosting docker, AMP.

vocal jungle
#

hmm

orchid vault
#

look in amp to see where its storing the stuff

kind phoenix
vocal jungle
#

amps saves in /home/amp/.config/Epic/FactoryGame/Saved/SaveGames/server

orchid vault
#

server only has 2 choices

kind phoenix
#

even for the steamCMD version?

vocal jungle
#

yes

#

amp only uses steamcmd

#

for satisfactory

kind phoenix
#

Because i can find the FactoryGame/Saved/ folders, but they are not inside .config/Epic

vocal jungle
#

I use Amp as my game manager

kind phoenix
vocal jungle
#

i done it in both

kind phoenix
#

Mine is in a docker which only has access to the folder /mnt/user/appdata/amp

vocal jungle
#

is your docker self loaded or you useing something like proxmox or unraid

kind phoenix
#

unraid

vocal jungle
#

one sec

kind phoenix
#

heck. I think the files are being written into the container not to appdata.

vocal jungle
#

could be i am looking

kind phoenix
#

I'm able to restart amp and my saves are fine.

#

but I did an update to amp, and lost my save

vocal jungle
#

do you have krusader installed in unraid

kind phoenix
#

I do.

vocal jungle
#

are you useing it to browse the dir trees

kind phoenix
#

ish.

#

I'm using a shortcut to the appdata folder

vocal jungle
#

hmm

#

i found my docker save in /home/amp/.config/Epic/FactoryGame/Saved/SaveGames/server

#

for standalone amp

kind phoenix
#

ah!

vocal jungle
#

in the sf docker i just setup i found the save in

kind phoenix
#

I don't have it mapped out

vocal jungle
#

/media/appdata/satisfactory/.config/Epic/FactoryGame/Saved/SaveGames/server

#

but that was just a sf docker not a amp docker

#

i think i am running diffrent versions of the satisfactory server just from helping people

vocal jungle
#

so @kind phoenix did you get it all sorted then

kind phoenix
vocal jungle
#

well you can upload a save using the editor and unlock what you had befor

#

so your not starting 100% over

kind phoenix
stoic stream
#

which they still haven't fixed

rich palm
white brook
#

I know the minimum is 6GB, anyone get away with 4GB? I have an old laptop with an i7, thinking about trying to run a server for 2 people on it

cloud flint
#

I can't see all the blocks of the train lines. My Server is in my local network and the quality is set to 3 (Ultra) on the server and on my client. The tick rate is between 29 and 30.
Is this a known issue?

cloud flint
paper ridge
#

if Im running a ds from my own pc, it has to be turned on to keep the server also online ?

dawn vine
#

Imagine if you turned your tap off, the water would stop right? Same principle.

paper ridge
#

thanks

vocal jungle
# cloud flint

its a known issue it was supposed to be fixed in last patch but i think its still being worked on

warped spear
#

i want to setup a dedicated server, and i was able to setup the server, claim it and create a savegame, but everytime anyone tries to connect to it they instantly disconnect with the mesage "Connection to Host lost". Any Ideas?

nimble sapphire
#

did you forward the ports?

warped spear
#

no, wich ports do i need to forward?

#

its a dedicated deboan machine btw, im not hosting it from my home

nimble sapphire
#

oh

#

hmm, I dunno in that case

#

check the logs, post it here

vocal jungle
#

15777,7777,15000

#

are you hosting from home ?

#

sorry missed that last cooment

#

if you can connect remotely to set it up then other people should beable to connect in less your connected to the server via a vpn

warped spear
#

this is everything the serverconsole says when someone wants to connect to it: [2022.02.24-21.03.05:869][773]LogNet: NotifyAcceptingConnection accepted from: ipadress

#

i never needed to forward ports or somehting for any other gameservers on this machine btw

#

where can i find the client logs?

vocal jungle
#

are any of the ports used by another game/service

warped spear
#

currently i stopped everything

dawn vine
#

I don't think it would bind if there was other services using same port

vocal jungle
#

it shouldnt but who knows sometimes

#

did you allow the ports in UFW

dawn vine
#

I don't think server would see the connection acceptance if firewall was blocking πŸ˜„

#

I'd wage UFW would be disabled on his machine or allow all rule is set

vocal jungle
#

you could if say port 15777 was allowed and 7777/15000 where blocked

dawn vine
#

Eh, that would be intentional to allow one and not the other though right? Sure he could have forgot but very slim chance on that

vocal jungle
#

who know most people dont notice there double nat'd so i ask my question based on the lowest comon denomantor

dawn vine
#

oh and debian by default doesn't have ufw

#

At least in my experience Debian 9, 10 and 11.

#

I doubt it's port/network related since he said it's on a dedi.

#

Unless host's firewall (like OVH's Game) needs configuring, but I don't think he'd see a connection attempt at all if that was the case.

vocal jungle
#

it not a issue at all i am able to connect to his server no problem

dawn vine
#

@warped spear What does the client experience/show when connecting to your server using port 15777?

#

Yeah, I thought it may have been a client-side issue

vocal jungle
#

i am on it i placeed the hub now i am eating all the nut and berrys

#

tick rate is 30 says open nat- seems fine on all fronts

warped spear
#

i am out of ideas, we cannot connect to it from 3 different places, i even disabled my complete firewall...

vocal jungle
#

can you connect

warped spear
#

ingame it goes to the loading screen and just kicks us back to the main menu, with "connection has been lost"

#

i can configure it in the server manager though

dawn vine
#

Are you able to connect to any other server?

vocal jungle
#

what he said

warped spear
#

dunno i never tried

#

are there public servers i could test it on?

vocal jungle
#

are you trying to connect with EA or experamental

dawn vine
#

FYI - the server management is pretty buggy generally. For me sometimes it takes like 3-4+ "Retry" attempts before it shows the join icon.

warped spear
#

early acees

#

i need experimental, right

vocal jungle
#

i have a server you can try let me make sure its up

#

sent you the ip

#

no EA is the correct one

#

@dawn vine can you try my server quick i dont want to say its a him issue with out a second connection success

nimble sapphire
#

Maybe its a permission issue? ie when people connect I think it saves the player data, maybe it cant create that?

vocal jungle
#

no i was able to connect no issue to his server

dawn vine
vocal jungle
#

no worries i going to assume my server is still good i have had people on it in the past

#

i think he is going to reinstall his game

#

we think we solved his issue woot woot

#

good old case of uninstalling mod loader and not removing the mods first i have been stung by that bee befor

dawn vine
#

The sooner DS work with mods the happier I will be

vocal jungle
#

its coming it slow but its coming

orchid vault
#

sooner mods work on dedis .. the slower dev will go

ebon bone
#

Anyone want to start a new factory with me

frosty gale
#

Hello! I am looking for some assistance with getting buddy on my dedicated server. Is this a good place to come for that help?

orchid vault
#

this is the place

frosty gale
#

Ok so, I've made it so far as to be able to actually join my own dedicated server. I am not using a third party service such as gtxgaming or low.ms. Everything is local (as I understand it). So my question is how do I get a friend on my dedicated server? I have tried every rendition of my server info I can think of to include my public IP address. We've used the name of my server, the name of my play session, tried to join through friends list but nothing is working. One fix I was told about, I have to open the ports through my routers settings. Haven't started figuring that part out yet. Thanks again for any help.

orchid vault
#

fix the ports on the wiki page

#

firewalls

#

defender or any bs software

frosty gale
#

Ok, I will look into these. Thank you for the help! Enjoy your night.

orchid vault
#

u 2 good luck

next cedar
#

<@&370483737957236737> [2022.02.25-11.51.31:913][ 0]LogNiagara: Warning: NiagaraSystem /Game/FactoryGame/VFX/World/AmbientLife/Flies/NS_Fireflies_01.NS_Fireflies_01 IsReadyToRunInternal() failed due to missing SystemScript. Spawn[<none>] Update[<none>]
[2022.02.25-11.51.31:913][ 0]LogNiagara: Warning: NiagaraSystem /Game/FactoryGame/VFX/World/AmbientLife/Bats/NS_Bats_01.NS_Bats_01 IsReadyToRunInternal() failed due to missing SystemScript. Spawn[<none>] Update[<none>]
[2022.02.25-11.51.31:914][ 0]LogNiagara: Warning: NiagaraSystem /Game/FactoryGame/VFX/Misc/Event/Christmas/NS_Fireworks_02.NS_Fireworks_02 IsReadyToRunInternal() failed due to missing SystemScript. Spawn[<none>] Update[<none>]
[2022.02.25-11.51.31:915][ 0]LogNiagara: Warning: NiagaraSystem /Game/FactoryGame/VFX/Misc/Event/Christmas/NS_Fireworks_01.NS_Fireworks_01 IsReadyToRunInternal() failed due to missing SystemScript. Spawn[<none>] Update[<none>]
[2022.02.25-11.51.31:915][ 0]LogNiagara: Warning: NiagaraSystem /Game/FactoryGame/VFX/Misc/Event/Christmas/NS_Fireworks_03.NS_Fireworks_03 IsReadyToRunInternal() failed due to missing SystemScript. Spawn[<none>] Update[<none>]
[2022.02.25-11.51.32:035][ 0]GameAnalyticsModule: Display: GameAnalyticsModule has started!
[2022.02.25-11.51.32:036][ 0]LogStreaming: Warning: Failed to read file '../../../FactoryGame/Saved/SaveGames/GameAnalytics.sav' error.
[2022.02.25-11.51.32:254][ 0]LogAudioCaptureCore: Display: No Audio Capture implementations found. Audio input will be silent.
deticated server errror HELP ME! πŸ™ πŸ™ πŸ™

#

@marsh dawn help PLS

dawn vine
#

So entitled

#

Good luck πŸ‘

uncut rampart
#

Hello, I would like to know how to create a dedicated server on satisfactory, thank you for your help!

sick bay
#

Hi ! How update my dedicated server ? My server version is 177909 but my game version is 179103..

#

I use Satisfactory Dedicated Server from Epic Games

dawn vine
dawn vine
nimble sapphire
#

you can always force an update by running the verification tag

#

as far as I understand, both the epic and steam versions have it but i dont know how it works on the epic version

dawn vine
#

No one does because who hosts servers using Epic's platform lul

uncut rampart
nimble sapphire
#

I just run a docker on my unraid server...so easy

sick bay
#

Where is Verification tag ?

nimble sapphire
#

dunno, Ive never used epic

#

look for its instructions, there is usually a flag you can add

dawn vine
vocal jungle
#

stop server then in epic library click the ... and verify

#

i think that will do it

#

it might update if you close the server and relaunch

vocal jungle
uncut rampart
#

i look on yt a video with steamcmd but i have satisfactory on epic game

vocal jungle
#

doesnt matter you dont need to own it on steam to use the steamcmd

uncut rampart
#

oh ok

#

mb

#

how does it work ?

vocal jungle
#

what do you mean

uncut rampart
#

how to use the steamcmd

#

where i can install

vocal jungle
#

do you have a diffrent system to run it on or you going to run it on the same system you play from

uncut rampart
#

that is to say ?

vocal jungle
#

there are 2 methods hosting it on the same system and hosting it on a dedicated system

#

which are you doing

uncut rampart
#

I don't know just I want to play with my friends, so something not too complicated to do and quickly

vocal jungle
#

then just play and share your session id with your friends

#

you dont need a dedicated server for that

#

inless you want the game to keep running while your not there

uncut rampart
#

yes but the host must be connected to the server to play

#

yeah

#

i want that

vocal jungle
#

you know it need to stay running on your system if you dont have a second system to host right

uncut rampart
#

ok

vocal jungle
#

do you have a mic

uncut rampart
#

yeah but im french

vocal jungle
#

like you can type english but can hear it ?

uncut rampart
#

so it's easier for me to talk by message

#

ok

#

its good

#

where you want go

lilac trench
#

Hey,

i am currently setup the dedicated server on an Ubuntu system. All works fine (done via SteamCMD).

Each day at 5:00am the server is automatically rebooting via crontab 0 4 * * * /sbin/shutdown -r +5 to ensure updates are installed automatically.

It seems the Satisfactory server process does not recognize the shut down (via reboot command) of the host system in any way.

If I am online in the game at the reboot time then I figured out that the Satisfactory server process does not save the game before host system is shutdown. Is there a way to force "save now" from the host system into the Satisfactory server?

On Linux, the server process can be gracefully shut down by sending SIGTERM to the server process or, if you have the server controlled by systemd, with the command systemctl stop satisfactory, where "satisfactory" is the name of your unit file. Note that the service file can specify the KillSignal= attribute in the Service block, but SIGTERM is the default if not set.

I tested systemctl stop satisfactory manually to check if a save game was created after using this command. But no save game was created.

What happens if the Satisfactory server process is killed during game saving? Is it possible that the game save is broken after that? I am a little bit concered about that.

(Note: I am a beginner on Linux systems. So I hope my questions dont look to stupid πŸ˜‰ )

Thanks for help!

lilac trench
#

In Addition:

If I understand the wiki correctly then if the host system shuts down it kills all processes with a kill signal. All systemd controlled services will stopped with a kill signal, too. So why the Satisfactory server does not recognize it correctly.

I can define KillSignal=SIGTERM in the service file but this have no effect (SIGTERM is the default if not set via the attribute).

I am a little bit confused.

cloud flint
# lilac trench Hey, i am currently setup the dedicated server on an Ubuntu system. All works f...

I am also a beginner on linux but I hope I can help you a little bit at least. The first thing is, that at least my dedicated server is restarting automatically every 24h with no extra script. Also I am running the server as a service https://satisfactory.fandom.com/wiki/Dedicated_servers/Running_as_a_Service. In this service is defined, that it checks for updates before every start. I think that would be the most elegant solution instead of using you own crontab. Regarding the saves, do you have any saves or isn't the server saving at all. There is a Console Command to create a savegame server.SaveGame saveName you can find it in here https://satisfactory.fandom.com/wiki/Dedicated_servers#Console_commands. But normally it should make autosaves and create a save before a soft restart or soft stop.

orchid vault
#

its not restarting its reloading the session same as the game does after 24hours

#

its to keep the sanity of the engine

smoky tangleBOT
#
Reaction Roles - Message Information
Message ID

935994011473293322

Channel ID

902621736602861588

Guild ID

370472939054956546

Message Link

#dedicated-servers message

unreal charm
#

If you check the logs you will see that server does recognize the first SIGTERM but throws and error in (hopefully) the end of the stopping process and stays in that state

#

On a second thought perhaps you don't actually have this problem. Can you check the logs (e.g. via journalctl -eu satisfactory) to see if you have "LogOnline: Error: OSS: Leaderboards Interface Requested" in them when you try to stop the server?

Regardless, I don't think saving the game is part of graceful shutdown sequence
https://questions.satisfactorygame.com/post/6187120f831c85205235b402
https://questions.satisfactorygame.com/post/618d6778831c85205235d0d1 (the commenter here seems to be wrong and the kick doesn't result in a save)

lilac trench
unreal charm
dull berry
#

Hello everyone, would someone be so kind as to explain to me how to create a dedicated server? For information I am on epic game and I have downloaded "Satisfactory Dedicated Server" (help please ;__; )

unreal charm
dull berry
#

No, actually, I would like to create a server...

unreal charm
dull berry
unreal charm
# dull berry Sorry, I didn't understand. In this case I don't have a separate computer so I w...

I would advice against that since you already have multiplayer on your pc without a dedicated server, and I'm fairly sure that running both server and the game on your pc at the same time will consume more cpu and memory. I personally rent a VPS for $12 per month (clovux KVM-8G, their kvm plans are clunky since they seem to setup them by hand, and they only host in France(I think), but I guess that's why they are cheaper than others)

But anyway, you downloaded "Satisfactory Dedicated Server" on epic games. I don't use epic games, but can't you just launch that like you do with everything else you download on epic?

dull berry
#

when I run "Satisfactory dedicated server" it opens a command window a bit like Powershell, I don't really know what to do with it next...

unreal charm
#

If you want other people to join your server you probably need to setup port forwarding so port 15777 (also preferably port 7777 and port 15000) on your pc is reachable from the Internet. Also do read the wiki https://satisfactory.fandom.com/wiki/Dedicated_servers . Usage with epic game on windows seems pretty much the same as with steam on windows

dull berry
#

The problem is that when I want to create a server, I put my ip address (as many people say they do), and I put the port 15777 (I think it's the one to put) so a server is created but it shows me "this server seems to be offline"

unreal charm
#

In my exeprience once you see "server appears to be offline" it will always appear to be offline even after the server is fixed until you restart the client

dull berry
unreal charm
#

yes

dull berry
#

I just did it and it still says "the server seems to be offline".

unreal charm
dull berry
#

Thanks for your help, I think I did it. Now I just have to hope that I don't have any problems with the backups :))

dire oar
#

can somebody help me setup my server?

unreal charm
dire oar
#

steam windows

unreal charm
# dire oar steam windows

Pretty much nothing to setup there aside from port forwarding if you want other people to join. And port forwarding setup is different for ever router and every ISP

dire oar
#

ugh

#

isnt there a thing on steam delighted server i can download when i open it nothing happends except a cmd opens up

#

dedicated *

unreal charm
dire oar
#

how do i see the server adress tho\

frosty gale
#

but how do you open the server from therew

unreal charm
dire oar
#

if my firend wants to join doi just give them my ip?

#

or hamachi?

#

wait is it possible with hamachi?

unreal charm
#

ooh i forgot things like hamachi exist

#

anyways - individual for every case. First, your firend needs to be able to ping that IP one way or another. Then ensure the ports are open

#

(localhost doesn't count for "able to ping")

dire oar
#

how do i do it with hamachi

unreal charm
#

well hamachi probably assigns you an ip on a hamachi network. If your firend can ping that ip he should be able to put it into the server manager

#

if windows firewall pops up about blocking some features - tell it to allow. Don't forget that you need to be the first person on your server if you want to own it (localhost is handy for that)

dire oar
#

so i start a network and then we use that ip?

unreal charm
#

Probably, Your firend also probably needs to join your hamachi network. But don't ask me, i don't know how to hamachi

dire oar
#

ok

dire oar
unreal charm
dire oar
#

ok

sleek kraken
#

dedicated servers are for friends to join not random players right?

lilac trench
# unreal charm `journalctl -u satisfactory` optionally `-e` to scroll to the end or `-f` for it...

Here is the log. I have marked the positions with actions (manual or automated ones) beginning with --.

It seems on normal reboot it shuts down the Satisfactory server process gracefully (see for example timestamp Feb 26 04:05:01 to Feb 26 04:05:36):

Feb 24 04:05:02 satisfactory systemd[1]: Stopping Satisfactory dedicated server...
Feb 24 04:05:02 satisfactory systemd[1]: satisfactory.service: Succeeded.
Feb 24 04:05:02 satisfactory systemd[1]: Stopped Satisfactory dedicated server.
-- Reboot --
Feb 24 04:05:13 satisfactory systemd[1]: Starting Satisfactory dedicated server...
Feb 24 04:05:35 satisfactory systemd[1]: Started Satisfactory dedicated server.

But no save games are created.

Conclusion: Bug?

(And thanks for where to find the log πŸ™‚ )

#

(Is there a way to connect via host system console/shell to the Satisfactory server console to call commands? So I can use this as a workaround to inform players online about reboot and force to create a save game?)

unreal charm
unreal charm
# lilac trench Here is the log. I have marked the positions with actions (manual or automated o...

I think it's not a bug but rather a missing feature
Upvote these or maybe we need a new one
https://questions.satisfactorygame.com/post/6187120f831c85205235b402
https://questions.satisfactorygame.com/post/618d6778831c85205235d0d1

Also in your case I think you can just disable automatic system reboot since very few updates actually need one. Otherwise I have an idea: instead of timing saves to the restarts you can time restarts to the saves - instead of a simple shutdown run a script that will monitor save folder and wait up to 1.5xautosave_interval for a new save to be created / autosave updated

unreal charm
#

@lilac trench I'm very interested in how your server stops when you tell it to stop. Mine doesn't do it. Can you please run an experiment for me? Wait for the server to laod the game and then systemctl stop it. Does it stop within 5 seconds or does it hang around for 90 seconds? I'd like to see your logs for that run.
What version of ubuntu are you using? (cat /etc/os-release) Is it a headless server?
Are you using satisfactory.service file from the wiki or a modified one?
What are your "Pause game when no players are connected" and "Save the game when a player disconnects" settings? Any changes to the config files?
Did you do a steamclient.so symlink?

stoic stream
#

I run it as a service set up though NM

#

nssm

lilac trench
lilac trench
# unreal charm I don't think there is at the moment. Which is a shame.

Are you using satisfactory.service file from the wiki or a modified one?
From the wiki.

[Unit]
Description=Satisfactory dedicated server
Wants=network-online.target
After=syslog.target network.target nss-lookup.target network-online.target

[Service]
Environment="LD_LIBRARY_PATH=./linux64"
ExecStartPre=/usr/games/steamcmd +force_install_dir "/home/steam/SatisfactoryDedicatedServer" +login anonymous +app_update 1690800 validate +quit
ExecStart=/home/steam/SatisfactoryDedicatedServer/FactoryServer.sh
User=steam
Group=steam
StandardOutput=append:/var/log/satisfactory.log
StandardError=append:/var/log/satisfactory.err
Restart=on-failure
WorkingDirectory=/home/steam/SatisfactoryDedicatedServer

[Install]
WantedBy=multi-user.target

What are your "Pause game when no players are connected" and "Save the game when a player disconnects" settings?
Both activated.

Any changes to the config files?
No. Only disable automatic send reports.

Did you do a steamclient.so symlink?
Not as I know.

I installed it like this with a new host system user (not root) with sudo rights.

sudo add-apt-repository multiverse
sudo dpkg --add-architecture i386
sudo apt update
sudo apt install lib32gcc-s1 steamcmd 
steamcmd +force_install_dir ~/SatisfactoryDedicatedServer +login anonymous +app_update 1690800 -beta public validate +quit

Wait for the server to laod the game and then systemctl stop it.
I loose the connection after some seconds.

unreal charm
#

Can you please share the log from a loaded game state up to a to systemctl stop's "Stopped Satisfactory dedicated server."?
I'm not sure if all of your logs are in the journalctl or some of them are actually in /var/log/satisfactory.err and /var/log/satisfactory.log but I'd like all of them.

unreal charm
#

What is your starting biome?

lilac trench
unreal charm
#

I just realised the post on QA site is actually about server running on windows and I can in fact reproduce the issue on my windows PC.
Which really doesnt help, wth

tawny mountain
#

is there a way to get the server using multiple cores?

orchid vault
#

it does

tawny mountain
#

MIne isnt

orchid vault
#

it cant just spread out all the work

#

most work is serial and has to be done in order

#

so it loads 1 core and uses threads across rest

tawny mountain
#

So it basically single threaded thanks.

orchid vault
#

no

#

its multithreaded but it cant just split it all up

#

cause the shit has to be done in order

#

it uses a bunch of worker threads

#

so its not 1

#

have a good one

tawny mountain
orchid vault
#

u can disagree but u would be wrong

#

devs explained it more then 1 time on streams and other things

#

try the wiki or videos or streams

#

just dont ping me i dont want to know

tawny mountain
#

Just make separate Conveyer belt Systems there own Thread just a way to aproach this Problem.

#

@golden roost I dont know who your programmer is so I pinged you sry.

quaint storm
#

anyone having an issue with programable splitters on a server crashing your game?

orchid vault
#

check the bug site

quaint storm
#

yea i found only about 8 threads for the same issue

crystal eagle
#

They’ve been broken since the start of update 5 πŸ˜“

harsh silo
#

Is there a way to run an already existing world on a dedicated server instead of creating a new one when first connecting?

harsh silo
#

pls jace_smile

past basin
#

Hello there. Can i get some help? I bought a dedicated server and i can join it with no problem, but my friend cant connect " Your connection to the host has been lost". Is there anything i/he can do about it?

calm saffron
calm saffron
#

I've been seeing memory leaks in my dedicated server, where it starts at around 5.2GiB, and just keeps going up until it reaches the 8GiB limit and crashes πŸ˜• Is this a known issue?

modest glen
#

anyone here with experience in satisfactory dedicated servers able to tell me how stable they are when running one?

late coral
#

The dedi server is quite stable, just follow recommend system requirements.

It can use up to 8G memory, so good advice is to have a bit more available.
Due to the nature of games, dedicated servers included, it is more preferable to have a higher clockspeed over having more cores. Have an SSD.

So basically run it on a gaming rig, minus the GPU.

dawn vine
#

Clock speed is so meaningless

#

You can have low clock speed high IPC single core, high multicore CPU results that absolutely ruin low core count CPU

#

You want high single core performance, core count is meaningless provided it's single core perf is high.

#

AMDs Zen 3 is a great example, high core, high single core perf.

#

I'd ignore what Master said a lot of it reads false. Sorry chief.

bright robin
#

You read it false, but he is right. IPC does make a difference, but people usually infer that recent CPU are better than old ones, so the IPC talk (which heavily depend of the load type, so it's complicated) does not help. And yes, having lot of core does not exclude good single core perf. But all of that is noise. The info "higher clockspeed over more cores" is a good simple way to put it πŸ™‚

#

Concretely, an Intel 12100 with 16Go DDR4 and a value SSD (500Mo/s for read & write is enough) would make a very good dedi server for a budget price! But it's hard to have this sort of hardware in hosted environment unfortunately.

vocal jungle
#

I am running 3 server on
Ubuntu 20.04.3 LTS | 31GB RAM
Intel Xeon X5675 (16C\16T) only a 3.06(3.46Boost) and i can have around 4 - 10 people on each server

#

i found memory made more a difference with SF the clock speed

unreal charm
vocal jungle
#

yea

#

when i only allocated 6 gigs to the VM then i had tick issues

unreal charm
#

Good, good. They say clockspeed over cores but they don't really specify in what proprotions or minimal frequency per core. If 3GHz is enough it opens up a lot of vps options

vocal jungle
#

and no ssds in the sever

#

that the server

unreal charm
#

@vocal jungle when you gracefully stop a server (e.g. via console) does the server process stop? Like in 5-10 seconds as opposed to hanging around until another SIGTERM or a systemd kill

vocal jungle
#

then the vm has 8 cores and 32 gigs with 200g of storage

#

i use Amp

unreal charm
#

nvm then

vocal jungle
#

but no it does not gracefully stop its a known issue

unreal charm
vocal jungle
#

its still so new i just treat it like it is a WIP

#

like there no real api no rcon no real command set

#

its very barebones so i dont treat it like a fully flushed software suite and i can live with that

unreal charm
#

It seems like graceful stop is supposed to be an implemented feature tho

vocal jungle
#

having about 6 diffrent methods of using the dedicated server currently install i am well versed in most issues

#

just cause something is implmented doesnt mean it works lol

unreal charm
#

Wiki is written like it works. I'd expect to find a qa post about it but the one I did find is 4 month old with 3 upvotes (1 mine) and is really hard to find for some reason

#

Either people just don't care or it's not a common issue.~~ (or qa site is not that good)~~

vocal jungle
#

i have see people post in here but if they dont follow up at the QA site is just sits there

dawn vine
lyric niche
#

Hey, Im having a bunch of issue using the epic version of the dedicated server tool. Would anyone have a moment to help?

vocal jungle
#

what trouble

lyric niche
#

I am opening the dedicated server tool but it does not seem to be running correctly. I get error messages involving a steam ID or environment

#

I can send screenshots if that would help!

lyric niche
#

I have also tried it through steamcmd and i get the same sort of issues

orchid vault
#

warnings are just that

#

errors are labelled ERROR

#

your server is most likely working fine

lyric niche
#

when I open the game and put in my IP it says the server is offline

subtle sundial
#

those old xeons only have 6C12T

#

even dual socket that's 12C24T

#

f u dyno

vocal jungle
#

please dont mansplan my server hard ware to me

orchid vault
#

hahaha

subtle sundial
#

or your hardware is being misrepresented by your provider

vocal jungle
#

except i am not, it lists in Amp as (16C\16T) becasue it is a VM and that the number of cores i have given the VM

orchid vault
#

its wrong lol

subtle sundial
#

oh i see, you're conjuring cores out of nowhere lmao

orchid vault
#

just look up ur cpu lol its 6core

subtle sundial
vocal jungle
#

i dont need to read it

subtle sundial
#

magic VM cores is not going to make your thing more performant. it has the opposite effect

vocal jungle
#

i know the specs of the cpu and your not incorrect no did i say you where wrong

#

err i did not say your wrong

subtle sundial
#

so what's the deal here, im confused

#

you have a dual socket setup, and you've assigned 16 out of 24 threads to a VM?

vocal jungle
#

i understand its 6/12t and it in a r710 - to make a total of 12cores 24 theads

orchid vault
#

exactly its not 16c/16t that was the point

vocal jungle
#

the vm has 16 threads allocated to it but AMP/Linux sees it as 16cores

subtle sundial
#

weird
but you know what it is, so why say 16 cores?

vocal jungle
#

and i copied the stats off the amp setup page and didnt correct the terms not thinking some retentive dick you care about the terminology

#

i am not misleading i just need think it was a bfd

subtle sundial
#

my initial intention was to warn you, since i did not know you or your setup/level of tech knowledge
because server providers can be very bait-and-switch

vocal jungle
#

i dont know why it says 16 cores i didnt write amp or linux or exsi ask them why the count theads as cors

#

cores

#

fair enough and that info is important for people to have, and thanks

unborn path
vocal jungle
#

just may get all the info before you say someone is wrong, when it could be a type-o or claim there misrepresenting there hardware

#

its the AMP game hosting manager it use to be McMyAdmin then it was rebranded as AMP

#

not it supports many different gameserves

#

the crap part about server providers is there selling 4 slot servers with 2 gig of ram thinking that is enough

frosty gale
#

asd

tawdry fjord
#

Hello, does anyone have experience with Satisfactory servers? I want to connect to my Dedi server, but when I join I just float around and it works. Until it crashes!

spark raven
#

anyone know if the items on the conveyer belts on servers are just not properly displayed or really not working as fast as they should? xD (on servers)

late coral
#

I can already hear someone saying " yes, but boost clock"

Yeah, well, after the CPU reaches a certain temperature, you only get the base clock.

dawn vine
# late coral I may have oversimplified. Still you want higher single core performance over ma...

Who's buying a threadripper to host a Satisfactory server though? You can make the argument for Xeon servers sure, but most Panel hosts use Core I-series or Ryzen chips and if you buy "game ready" Dedi's from a lot of places (OVH is a good example) they provide high single core performance chips.

Most people who host from home aren't going out of their way to buy expensive hardware to run a dedi for satisfactory, and if they are they likely know about the hardware nuances. For the people who just want to run a server from home will just deal with what they have.

spark raven
#

lmao my server uses a threadripper but shared lol

late coral
#

Most people will indeed not buy a Threadripper, I merely used it as an example.

spark raven
#

jep, totally agree - it's stupid though that only few games and stuff like this use more than one core

late coral
#

Still holds true with Core-I as well. You do not need the top of the line high core count CPUs, you need the ones with the highest base clock.

late coral
dawn vine
#

High Single Core performance isn't dictated by core count

late coral
#

The 12 core Threadripper (as per example) has a higher single core performance over the 64 core one

dawn vine
#

bro I have a 5950x. It has one of the best Single Core performance chips available

#

It has 16 cores and 32 threads

#

High core count βœ…
High single core performance βœ…

late coral
#

It's a server processor.... 16 cores is low core count....

dawn vine
#

What does server processor mean?

#

My OVH dedi has a Ryzen 3 3600X in it

#

Consumer chip, used in a datacentre

#

Is this not a "server processor"?

late coral
#

For it's intended purpose: running a game.

dawn vine
#

Isn't that what we're discussing though?

#

We're not talking about running an Oracle Database

#

We're talking about running Satisfactory

late coral
#

Yeah, for which you want, within a processor family, the one CPU model that has the highest single core performance, not incidently that is in most cases the model with less cores.

#

As the game server runs continuously, the CPU will always be at thermal limits, and thus running base clock

dawn vine
#

tbh not sure why you even bought up Threadrippers. It's so far from your own point idk what you're arguing anymore.

#

cpubenchmark DOT net/singleThread.html

late coral
#

I brought up the Threadripper as an example cause there's just 5 models, which makes demonstrating the point a lot clearer

dawn vine
#

No "server processors" in that list for a while

late coral
#

It's a nice extreme example

#

My point holds true for any processor family

dawn vine
#

State your point/argument again.

ivory sage
#

Unfortunately, Satisfactory's server runs best on a 2-4 core CPU with the latest architecture, which has a higher IPC than a processor with 128 cores and 256 threads.

late coral
#

The ones with high clockspeed, and thus high single core performance, have less cores.

dawn vine
#

Talk about extreme examples ^

dawn vine
late coral
#

Within a processor family it does

dawn vine
#

Sure you could make a trendline, but you cannot objectively state that

#

There's way too many nuances

ivory sage
#

In the past experience, when operating from a server that accommodates 16 people to tier 8, when the system was operated with Xeon E5-2620v2 at the beginning of server operation, a very serious lag occurred in tier 4 game state.

dawn vine
#

Provided the Single Core performance is high tier, the core count nor clock speed plays any factor.

ivory sage
#

When I felt operational problems and put in Ryzen 5600g, from tier 7 to around 7 megabytes of save file size, about 12 people could play the game with a little bit of lag.

dawn vine
#

You could have some magical CPU which has 500 billion cores, but it's single core performance is fucking fantastic which Satisfactory would run just fine. It'd use the cores it needed and the single core performance would be good enough to support whatever it needs to do

late coral
#

You can't compare a Zen3 with a Core I based on clock speed, true.

You CAN compare, and readily state that a single core , OF THE SAME ARCHITECTURE, the one with the higher clockspeed has the highest IPC.

That's is what I'm saying.

dawn vine
#

I'm speaking broadly

ivory sage
#

I'm not really making any comparisons between Intel and AMD either. It's just real data from actually running a multiplayer server.

dawn vine
#

Real data from how many samples though

#

Let's not be too confident lol

late coral
#

I'm talking real world CPUs, where there is plenty data that shows high single core performance is to be preferred for games.

dawn vine
#

I agree with that statement

#

We're making progress.

late coral
#

Indeed you are.

dawn vine
#

However, that statement is a far cry away from your previous stance

late coral
#

In which way?

dawn vine
#

MasterDiverM β€” Today at 15:10
The ones with high clockspeed, and thus high single core performance, have less cores.
Original statement ^

#

I'm talking real world CPUs, where there is plenty data that shows high single core performance is to be preferred for games.
This ^ statement didn't factor in core count or clock speed

late coral
#

Oh, boy you are stubborn...

dawn vine
#

You made my argument

#

High clock speed doesn't dictate high single core performance.

#

High Single Core performance isn't dictated by core count

#

We agree, games (like Satisfactory) prefer high single core performance.

late coral
#

It does within a CPU architecture.

A Zen3 core, running 3gHz has a higher IPC, performance etc, then the same Zen3 core running 2gHz.

dawn vine
#

How is this in anyway comparable?

#

Who's artificially limiting their CPUs?

#

Also. Just for reference. The Zen3 cores have all very similar single core performance results

late coral
#

Not artificially, production defects have a significant influence on the max clockspeed

#

Same performance in a benchmark?

dawn vine
#

Within tiny tolerances. Also this is SO far from your original claim.

#

Single Core performance chart for Zen3 (highlighted)

#

Notice something?

#

Lower core count have worse single core performance

#

Funny that

#

Your example btw

late coral
#

Yes, it's a benchmark, so no thermal limits have been accounted for

dawn vine
#

A Zen3 core, running 3gHz has a higher IPC, performance etc, then the same Zen3 core running 2gHz.

#

Doesn't matter. Sample sizes would mitigate issues with peoples setups

#

This isn't one sample for each chip lol

#

5950x -> Samples: 2560*

late coral
#

Thermal limits have no correlation with anyone's setup

dawn vine
#

Okay so why mention it?

#

5600x -> Samples: 5140*

#

Fact of the matter is, there's that many samples any defects or differences in chips would be smoothed out in the data.

late coral
#

Well, maybe because you can only run the CPU at base clock when you reach it's thermal limit?

dawn vine
#

You could have a Zen3 chip that runs like 5% of it's expected performance; won't change data too much if at all due to sample size

dawn vine
#

Or in the inverse, you could win the silicon lottery and get some chip that boosts most of the time without any thermal issues

ivory sage
#

Looking at the situation now, I'm just sorry why we have to run the server while considering CPU architecture, IPC, and Core Clock. Basically, if the server side supports multicore properly, this debate will not be very meaningful.

late coral
#

You definitely can win the silicon lottery, but you can only assume the average.

dawn vine
dawn vine
#

Like I said the amount of samples will create a MEAN average; ignoring top and bottom

late coral
#

Yeah, but you are using the wrong data to make an assumption.

dawn vine
#

kinda like a bell curve

#

Show me data that proves you right

#

because just an FYI, we've moved from your original claim

#

The ones with high clockspeed, and thus high single core performance, have less cores.

#

Are you happy to concede this is not true?

late coral
#

CPU cores, even when they do nothing, use energy, and thus create heat, which takes away from the thermal budget of the cores that do have a task.

dawn vine
#

haven't we just agreed that this works inversely?

#

You can have shit CPU lottery, you can have great CPU lottery

#

or you can fit right in the normal where most people will see expected results from X chip

late coral
#

Nope, that is due to the manufacturer selecting the best silicon for their "higher" placed models.

dawn vine
#

Then it's a different chip

#

you can't compare chips like that and say "I won the silicon lottery because I bought a 5800x and you lose the silicon lottery because you bought the 5600x"

#

If it has a different name on the chip, then we should not say they are the same chip, fundamentally sure, practicality they're sold at different prices, they yield different result, they may have different feature sets

#

I've never said "Wow, our 5600x and 5800x's are the same chip."

#

Just another pivot check, we're far from the original claim you made.

late coral
#

Lets compare the Ryzen 5000 series 5950X and 5800X

compare.jpg
dawn vine
#

And the real world yields are?

#

Samples are there

#

5950x has better Single Core performance

#

5950x has more cores

late coral
#

that's not even relevant, what matters is the TDP

dawn vine
#

5950x has slower clock speed

late coral
#

you are just showing another benchmark, that does not take in account the thermal limit

dawn vine
#

Bro listen. I do not care about TDP. I'm focusing on your original claim.

#

let me reiterate

#

The ones with high clockspeed, and thus high single core performance, have less cores.

late coral
#

it's all about TDP

dawn vine
#

Then why is that not what you said AT ALL until now?

#

FYI, this is called shifting the goal posts

#

First time you mentioned TDP

#

First time you mentioned thermal

late coral
#

you want a high sustained clock, in which a CPU with a lower core count has an advantage, it has fewer cores that generate heat.

dawn vine
#

Original claim

#

State your new claim because I feel I've adequately proved your original claim wrong.

#

oh and just an FYI, both of those chips (5800x and 5950x) have the same TDP

#

Your example remember

late coral
#

my original claim is still true, you only show some benchmarks that have been performed using the boost clock, yeah sure, the boost clock is all but the same

dawn vine
#

You keep trying to get this in there. You state it like it changes the facts.

late coral
#

so the result of the benchmark will reflect this, the performance at max boost clockspeed

dawn vine
#

State your (new) claim.

#

Be as concise as humanly possible

late coral
#

it's still my old claim, 5950X and 5800X have the same architecture, the cores are thus identical in design, and in performance at the same clockspeed.

dawn vine
#

let me just reiterate, correct me where I'm wrong. Your new claim is specific to the 5950x and 5800x; where you state the performance would be identical at the same clockspeed?

late coral
#

but you will also notice that the 16core 5950X has a lower base clock (when you reach the thermal limit, or TDP), than the 8 core 5800X.

dawn vine
#

Can we step through this conversation one by one please.

late coral
#

which also means that when the CPU can only run at it's base clock, the single core performance of the 5800X is higher than that of the 5950X

dawn vine
#

let me just reiterate, correct me where I'm wrong. Your new claim is specific to the 5950x and 5800x; where you state the performance would be identical at the same clockspeed?

#

I need a yes or a no and why

late coral
#

no it's not specific for these cpu's. It holds true within any cpu architecture family

dawn vine
#

Okay, I will ask the same question again. Please be as concise as possible.

What is your claim IN FULL?

#

I want us to be on the same page and right now I cannot do so because we're so far removed from the original claim.

late coral
#

it's true for the Threadripper Pro, It's true for the 12th Gen Core-I, etc

dawn vine
#

Concisely, what is your claim?

late coral
#

I think I have made myself clear multiple times?

dawn vine
#

no; hence why I'm asking for you to clarify in full so we can start a new conversation going forward.

late coral
#

My original claim, BTW, it holds true within any architecture family.

Knipsel.JPG
dawn vine
#

Okay

#

Perfect

#

Please note NO WHERE in that screenshot you provided states anything to do with power budget. So I expect going forward you do not bring it up.

#

I disagree with that claim.

#

Reason: Single Core Performance is NOT tied to Clock Frequency OR Core Count

#

Sub reason: it's made up of other factors, major one being IPC (Instructions per clock)

late coral
#

I also showed a nice extreme example, the Threadripper Pro, showing that within this family of CPU's the base clock is higher with less cores.
And this has always been true: Same architecture, the single core performance is proportional to the clockspeed.
Also true: You can't run a CPU at boost clock forever, it will throttle down to base clock (per core in some models) when the thermal limit is reached.

dawn vine
#

Okay let me address each point by point

late coral
late coral
#

5950X does not have a higher IPC than a 5800X

#

same is true for the Threadripper, the IPC per core of the 64core is the same as the 16core

dawn vine
#

I also showed a nice extreme example, the Threadripper Pro, showing that within this family of CPU's the base clock is higher with less cores.
We agreed this comparison doesn't make sense in reality. Threadrippers aren't mostly used to run game servers nor play games. Some people do, more power to them, the vast majority do not.

Amos --
What does server processor mean?
My OVH dedi has a Ryzen 3 3600X in it
Consumer chip, used in a datacentre
Is this not a "server processor"?
MasterDiverM β€” Today at 15:04
For it's intended purpose: running a game.

And this has always been true: Same architecture, the single core performance is proportional to the clockspeed.

This is not true. Using my example from earlier, the 5950X has a lower clockspeed, higher single core performance than the 5800X.

Also true: You can't run a CPU at boost clock forever, it will throttle down to base clock (per core in some models) when the thermal limit is reached.

This point is irrelevant. Taking a huge average of thousands of samples will provided expected result with X chip and will follow.

Note... You snook thermals back into conversation after having nothing to do with original claim.

#

What part of let me address your points first was misunderstood by you?

late coral
#

Snook in thermals? thermals have everything to do with my original claim

dawn vine
#

Yet they're stated in exactly 0 places (original claim)

#

Please note NO WHERE in that screenshot you provided states anything to do with power budget. So I expect going forward you do not bring it up.

dawn vine
late coral
#

this point 2?
"This is not true. Using my example from earlier, the 5950X has a lower clockspeed, higher single core performance than the 5800X."

dawn vine
#

Yes

late coral
#

of course it is, the benchmark ran at boost clock. the 5950X has a higher boost clock than the 5800X

dawn vine
#

Why did you use those CPUs as an example? They play into my argument.

late coral
#

basicly you prove my point: higher clock higher IPC

dawn vine
#

No lOL

#

That was not your original claim

late coral
#

same clock same IPC

#

it is

dawn vine
#

No

#

it was not

late coral
#

yes

dawn vine
#

if I took some shitty Pentium from 1922 (before they existed) and overclocked it to 50 trillion gigahertz whatever that would be. It would likely be the best CPU in the world.

#

Right? The single core would be nuts

late coral
#

yes

#

it would

dawn vine
#

But that's out of the realm of reality. The reality is. Core Frequency is NOT the result of good Single Core Performance chips.

#

I should say, not the only result/factor - as per my original claim

late coral
#

things I did not say, but you should infer:

  1. This is true within a processor family
  2. It depends on the architecture
dawn vine
#

No. You cannot pull the "infer" card

#

I said like 4 times to concisely state your point

late coral
#

yes I just did

dawn vine
#

You refused each time referring back to an obscure point

#

You lost any chance to use "infer" cards

#

it depends on architecture. Sure. So what? Doesn't prove your point at all.

late coral
#

you infer you can win this argument by determining I can't assume you should know about thermals and architecture?

dawn vine
#

Neither can I? We're not discussing whether one of us knows the designs more than the other are we?

#

We're discussing the base principle that Clock Speed = Single Core Performance

#

Fyi your original claim mentioned nothing about architecture either

late coral
dawn vine
#

You may have starting bring in thermals, but when I said to restate your claim you posted that ^

#

Which doesn't include Thermals OR Architecture

late coral
dawn vine
#

bro

#

Do you know what "restate" means?

#

Yes or no

late coral
#

well, I do, it appears you do not

#

I have restated (and expanded upon) my original claim several times

dawn vine
#

You chose to state your original claim after I gave you plenty of chances to modify your statement and provide a more concise, and thorough statement

#

The fact you didn't take that opportunity is your own fault

late coral
dawn vine
#

No you cannot expand on it after I explicitly asked you to do that in the restatement

#

I told you plenty of times I'm arguing your claim

#

And your ORIGINAL CLAIM

#

is WRONG

late coral
#

my original claim is true.

dawn vine
#

Yes you did. Then because the conversation was getting cloudy I asked you to compile your thoughts and restate it CONCISELY

#

You refuse to take that opportunity

late coral
#

I expanded upon the reasons, and limits of my claim

#

You keep bringing up the same false arguments, showing benchmarks that have been performed using a clockspeed that cannot be sustained.

dawn vine
#

See if you said correlated I'd be more inclined to agree (Frequency = Single Core Performance)

late coral
#

Is English your first language?

dawn vine
#

yes

late coral
dawn vine
#

Because we both know you didn't say as such

late coral
#

I made the correlation by stating that games benefit from single core performance, thus a higher base clockspeed (which can be sustained for longer, if not indefinitely), which is correlated to the core count, as more cores equates to a lower base clock as a general rule of thumb.

#

Because of thermal limits

#

higher clockspeed = more heat generated (per core)

dawn vine
#

It's such an irrelevant fact when considering CPU choices

#

CPUs generate heat. This is true. Okay cool, does it run faster though?

#

I use "faster" as a generic term that's all encompassing

late coral
#

the more heat per package it runs slower

dawn vine
#

This is not true.

late coral
#

why?

dawn vine
#

Sorry, let me restate.

#

"slower" than what?

#

in my statement of "faster", I was referring to whatever they had currently for comparison

late coral
#

well, let me explain: you have chosen a socket, all CPU's that fit this socket have the same package (more or less) right?

dawn vine
#

Define package

late coral
#

that package has a maximum dissipation of, let's say, 100W

dawn vine
#

package = IHS?

late coral
#

Package: the thing that carries the silicon, that enables the silicon to interface with the socket

dawn vine
#

So the PCB?

late coral
#

The IHS is part of the package

dawn vine
#

or the Die?

#

Package = PCB, die, IHS?

late coral
#

the whole thing

dawn vine
#

Okay cool

#

Continue with the analogy

late coral
#

so you have 100W to divide between all cores within said package right?

dawn vine
#

Just as a quick one, would you agree. Package = CPU?

late coral
#

you may call it CPU

dawn vine
#

Okay I'll refer to it as that from now on

late coral
#

for all it's intended purposes they are the same

dawn vine
late coral
#

true

#

All activated cores generate heat, agree?

dawn vine
#

AI = Artificial Intelligence? Can you provide an example of these cores if the acronym is correct.

late coral
#

all, typos I will correct πŸ˜‰

dawn vine
#

Okay, I would agree with a caveat. I don't know what happens if there's literally no work happening at CPU level, the CPU will still cycle at a frequency, how much heat that would generate vs a full workload at that frequency - I don't know the answer.

#

If you're asking for absolutes. Does a CPU that is powered generate any amount of heat? Yes.

late coral
#

You would agree, cores that have been deactivated (by having their power lines on die cut) do not generate heat?

dawn vine
#

No power, no heat. I agree.

late coral
#

so a 8 core cpu has per core more thermal budget than a 16 core cpu, right?

dawn vine
#

No. The thermal budget is the same. If it's rated for ~100W, a 1 core CPU still only has 100Ws of thermal room to use vs 8.

#

But obviously 1 core could be doing a lot of work in the 8 core CPU above and would have a portion of that thermal room available

#

Sorry - just seen the "per core" part. Yes. I believe that's reflected in the above statement(s)

late coral
#

ok

#

lets say a core uses 2W doing nothing (for this hypothetical cpu)

dawn vine
#

Sure

late coral
#

so the 16 core cpu has left 68W and the 8 core 84W to be used by cores doing work

dawn vine
#

I'm following

late coral
#

we also can establish that an increase in clockspeed means a exponential increase in heat generated?

dawn vine
#

Remove exponential. It's linear; you stated 2w per hypothetical core. Graphed would be a line

late coral
#

Each time the clock ticks you're charging or discharging a bunch of capacitors. The energy for charging a capacitor is:

E = 1/2CV^2
Where C is the capacitance and V is the voltage to which it was charged.

If your frequency is f[Hz], then you have f cycles per second, and your power is:

P = fE = 1/2CV^2f
That is why the power goes up at linearly with frequency.

You can see that it goes up quadratically with voltage. Because of that, you always want to run at the lowest voltage possible. However, if you want to raise the frequency you also have to raise the voltage, because higher frequencies require higher operating voltages, so the voltage rises linearly with the frequency.

For this reason, the power rises like f^3 (or like V^3).

Now, when you increase the number of cores, you're basically increasing the capacitance C. This is independent of the voltage and of the frequency, so the power rises linearly with C. That is why it is more power efficient to increase the number of cores that it is to increase the frequency.

Why do you need to increase the voltage to increase the frequency? Well, the voltage of a capacitor changes according to:

dV/dt = I/C
where I is the current. So, the higher the current, the faster you can charge the transistor's gate capacitance to its "on" voltage (the "on" voltage doesn't depend on the operating voltage), and the faster you can switch the transistor on. The current rises linearly with the operating voltage. That's why you need to increase the voltage to increase the frequency.

dawn vine
#

oh fuck me you're adding a lot to this hypothetical CPU

#

Why didn't you start with this?

late coral
#

for completeness we do need to add a lot

dawn vine
#

In your hypothetical model

#

fyi - hope you like it - paint pro

late coral
#

"You can see that it goes up quadratically with voltage. Because of that, you always want to run at the lowest voltage possible. However, if you want to raise the frequency you also have to raise the voltage, because higher frequencies require higher operating voltages, so the voltage rises linearly with the frequency."

#

This is the important bit, we need to raise the voltage to raise the clockspeed

dawn vine
#

Problem is, now you're out of the realm of this conversation. I'm not an electrical engineer.

late coral
#

remember, no silicon lottery in this hypothetical CPU

#

I happen to be an electrical engineer

dawn vine
#

I just have to agree with you, with no possibility of pushback

#

This is going to be a very boring conversation moving forward

late coral
#

which CPU can sustain the highest clock speed, knowing this?

dawn vine
#

I can't answer because I don't understand the provided context above.

#

It would be naΓ―ve of me to just say yes going forward. As I said, boring conversation.

#

We can have a conversation back on the main point if you want; but you've drove it into a wall - a wall I'm not going to learn Electrical Engineering to climb over

late coral
#

I can do a laymans explanation

dawn vine
#

But now you're appealing to authority

#

Again...

#

I just have to agree

#

You could feed me a load of bullshit, I can't pushback because I wouldn't know.

late coral
#

the 16 core had left 68W
or 4,25W per core
the 8 core had 84W left
or 10,5W per core

dawn vine
#

I just explained a moment ago there's no point continuing down this road since it would be me watching you type with no resistance.

late coral
#

this is not that hard to understand, I hope

#

You see that the 8 core has over 2x the power budget of the 16 core

#

and this is why less cores equates to a higher clockspeed

#

this is high-school fysics

#

analogy:
2 cars same model, one having a 100HP engine and the second a 200HP engine, same displacement, same everything. which car has the highest top speed, and by how much you guess?

dawn vine
#

In this crafted scenario, I'd be inclined to agree with you. Thermal headroom = higher boost = better single core.
However, this doesn't disprove anything of what I said.

5800x - 5950x example.
5800x has higher thermal room - less cores - same TDP - lower boost freq
5950x has higher single core - more cores - same TDP - higher boost freq

Though, the 5950x cannot maintain the boost freq forever - and if it can it has a higher single core performance on a high core CPU which fits against your original claim.

late coral
#

I've tried to explain that you can't assume the boost clock

dawn vine
#

Okay, ignore the boost clock then

#

5950x has lower base clock - higher single core perf - more cores - same TDP - less thermal headroom

late coral
#

that is what the benchmark tells you

dawn vine
#

But you just told me to ignore boost

late coral
#

yes, but that also means you can't use the benchmark score, which is measured at this boost clock speed

dawn vine
#

I'm lost. Why can I not use boost clock and if I can well see my above claim

late coral
#

and at the same time, you can see, the score is lower, for the CPU with the lower boost clock

#

the benchmark only tells you what the performance is, at the boost clock speed

dawn vine
#

I'd argue the 0.3GHz delta isn't enough to move single core performance to the point it disagree with you

late coral
#

I had to scroll back a bit:

#

But lets divide the score by the boostclock

#

you get something funny

#

5990X scored 3500 SCP (Single core Perf)
5800X scored 3450 SCP

#

you posted this benchmark result

#

divided by the respective boostclock we get:
5950X: 3500/4.9= 714.3 SCP/GHz
5800X: 3450/4.7= 742,6 SCP/GHz

#

funny result isn't it?

#

I think we can't trust the benchmark, in my humble opinion

dawn vine
#

gimme 20 mins just gotta run I will respond though lol

late coral
#

gonna prepare dinner, so might be more than 20mins before I respond πŸ˜‰

dawn vine
unreal mauve
late coral
#

Except when you're talking the exact same core-die as is the case with 5950X and 5800X.

unreal mauve
#

They're not exactly the same bruh

late coral
#

5950 probably scored lower because of scheduler overhead, 5950 has 2 core dies identical to the 5800's one core die.
Interface complex is identical as well

#

Just do a quick Google search...

unreal mauve
#

πŸ€¦β€β™€οΈ

#

"just google it bro"

late coral
#

Yep, "they are exactly the same bruh" is an even worse argument than "go Google it."

unreal mauve
#

there's this magical thing called binning

#

sure, in theory the CCDs in both the 5950X and 5800X are the same; they are not in reality

late coral
#

There are physical defects that influence performance

#

I KNOW

unreal mauve
#

if you're going to try to argue "this benchmark is invalid because [dumb math] you're making a huge mistake"

#

especially if said benchmark is an average of every result submitted by the users

#

it's going to naturally have flaws and biases that you can point towards that don't involve math

dawn vine
#

You're stealing our conversation 😑

#

We were making some nice progress

unreal mauve
#

no, no. I'm just here waiting for the "well, you're a woman you obviously don't know tech" or something dumb like that

#

because I have yet to get it myself

dawn vine
#

I'm joking about the convo theft btw. More the merrier regardless of what opinion you hold πŸ˜„

#

heeeeeeeeeey now

#

I don't think Master would make that argument, you made that argument yourself

#

c'mon now

#

Have some self-respect

unreal mauve
#

oh shit, you're right

late coral
#

As a matter of fact, the best hydraulics engineer I ever worked with was female

unreal mauve
#

i must go yell at myself now

dawn vine
#

tbh you said that like you wanted the conversation to derail into that

unreal mauve
#

it's mostly car people that make that argument anyway

dawn vine
#

Really unnecessary

unreal mauve
#

it was a joke

#

but back on topic; if you want to argue against a benchmark go for userbenchmark

dawn vine
#

yeah for sure 100% agree ^

#

If I was lifting benchmark results from userbenchmark I'd be for sure providing shitty data

#

But Passmark I hold quite high respect with benchmarking companies

unreal mauve
#

going from being the one everyone loves and trusts to getting banned from several subreddits is impresseive

late coral
#

Any benchmark....
Most will not divulge how they come to the results they present

unreal mauve
#

passmark does

#

all their formulas are available on their forum

#

and, on the topic of dumb math arguing against passmark

#

I benchmarked an E6600 and a Q6600

late coral
#

Still the IPC of a 5950 and 5800 are identical when looking at a single core

dawn vine
#

And I believe that shows in Passmark

unreal mauve
#

the multi-thread result of the Q6600 was almost exactly 2x the result of the E6600

dawn vine
#

There's a delta of 50 in Single Core perf

unreal mauve
#

within margin of error

dawn vine
#

^

late coral
dawn vine
#

let me see what freq I get on my 5950x when stressing sc with prime95

unreal mauve
#

expected result for those two CPUs given the E6600 is 100% just half a Q6600

dawn vine
#

I don't think I'll see 4.9

unreal mauve
#

I just scored 3276 for single-thread with my 5950X

#

but I have a bunch of stuff running so take that with a grain of salt

late coral
#

Ah, yes, any user submitted data is influenced by exactly this

unreal mauve
#

you actually have to hit an upload button

#

which I do not plan on doing

late coral
#

You didn't, but countless others did

dawn vine
#

yes but this would be smoothed out in samples size

unreal mauve
#

that's the downside of crowd sourced performance metrics

dawn vine
#

idk why you're so concerned with what some people do with their chips. After 1k samples it doesn't matter

#

People with 5950x aren't ALL going to be using 90% of the CPU and leaving 10% for the benchmark

#

Same with any other user and CPU combo

#

All of it will be factored into high sample size scores

#

Those scores will be reliable

unreal mauve
#

3543 after closing everything

late coral
#

Well, the interconnection complex is going to be busier with 2 CPU complexes attached

dawn vine
#

Funny this btw. My 5950x is stressing 1 core at 4.7GHz Stable, same as 5800X

#

And the SCP is within 50 points delta

late coral
#

They are remarkably the same

dawn vine
#

IDD

late coral
#

Until you start stressing enough cores

dawn vine
#

Yes but the numbers we're referring to is Single Core

#

So that's a pretty fair result right there

#

And reflects reality AND the passmark benchmark

late coral
#

Run multiple instances bound each to their own core.
Reality is that a game uses 1 core to the max, but the other cores are busy too

dawn vine
#

IDD

late coral
#

You quickly run out of thermal budget to sustain a higher clockspeed

dawn vine
#

So your argument is... Single Core Performance is important, provided it's actually sustainable on Multiple Cores, though Multicore itself isn't too important?

late coral
#

Because of the total thermal budget simply being much larger per core (our 8 Vs 16 core CPU)

dawn vine
#

So am I accurate with the reframing?

unreal mauve
#

you know, I could also bring an actual 5800X into the conversation; my roommate has one

#

that definitely requires getting up though

late coral
dawn vine
late coral
#

And remember, for games, as these are very single core performance dependent

unreal mauve
#

not all games are single-core dependent

dawn vine
#

yeah but I just helped reframe it and that isn't your entire argument

late coral
#

True, but Satisfactory is

unreal mauve
#

most newer games actually arent; devs are pushing multi-threading whenever possible

dawn vine
#

Your argument is... These games are single core performance dependant, on multiple cores, but not all cores.

#

lol

#

Which we were never arguing at the start

#

this is an entire new argument

#

And I don't disagree with

unreal mauve
#

that argument is actually part of why userbenchmark is banned from several subreddits (along with chaning their formulas to support their own point of view leading to some fun comparisons)

late coral
dawn vine
#

tbh I don't know many games that are true multicore. Most if not all I've played have most of the logic on the main tick thread/simulation thread, and everything else can be passed off to additional threads as small workloads.

unreal mauve
#

the only game I actually payed any attention to started transitioning to multi-core support by offloading sound processing to a second core

#

dunno where it went after that; i stopped caring for that game

#

Unreal does support multiple threads

dawn vine
#

I believe most modern game engines will support multiple threads, but it's up to the developer to utilise it

unreal mauve
#

there's a launch option to utilize all available threads

#

I noticed not change in performance or thread usage when using it

dawn vine
#

And in practice, it's hard to move simulation/order sensitive stuff from the main thread due to risk of out of order execution (this is assumption/speculation)

#

Thus, making is effectively heavy single core performance with hints of multicore workload πŸ˜„

dawn vine
unreal mauve
#

i should also mention that work on multiple threads is easily observed when playing Satisfactory

late coral
#

yes, but the main simulation thread is single core

dawn vine
#

oh yah, most games will use multiple cores, but almost all work will be done in main loop - single thread

unreal mauve
#

do you have documentation that backs that claim up?

dawn vine
#

Almost every single game

late coral
#

thus limited by clock-speed on any given CPU

unreal mauve
#

leaving out some information there

late coral
unreal mauve
#

clock speed isn't the only indicator of performance still

dawn vine
#

Satisfactory, Dying Light 2, Minecraft, New World, Hell Let Loose, Deceit, Dyson Sphere Program, Counter-Strike

unreal mauve
dawn vine
#

there's a few, multicore games with heavy main-loop single core load

late coral
unreal mauve
#

Factory tick parallelization: Spreading out the factory tick over more cores if available. This can considerably improve factory performance, depending on the CPU

#

the only thing I could find on short notice related to multithreading

late coral
dawn vine
#

eh good luck with this argument lol. I've tried having it. The new, restated claim which I agree with is:

Single Core Performance is important, provided it's actually sustainable on Multiple Cores, though Multicore itself isn't too important.

unreal mauve
#

there are situations where that claim is true and situations where it's false

if you have between 8+ cores sure; being able to sustain higher performance on actually necessary amount of cores will be important.

Anything less than 8 cores and that argument makes zero sense.

vale torrent
#

I'm running a server with linuxGSM and it seems to want to reboot, but will only actually do it if I'm online and playing? Like I got the 60 second warning and logged off, couple minutes later logged back in and 'server will reboot in 30 seconds'

unreal mauve
#

Does it eventually kick you off then reboot the server if you stay connected?

dawn vine
vale torrent
#

It did kick me off and reboot, it's back up now.

dawn vine
#

Have a look at the logs. I've had my server online now for like 100 hours -- since last update

vale torrent
#

I didn't trigger the restart though

unreal mauve
#

setting up a dedicated server in this game is an interesting experience

spark raven
#

anyone experienced a bug where a train station is appearing randomly on the map

#

calculator is showing me this...

unknown.png
#

ingame...

unknown.png
#

lol it's a common bug as i just found out on the satis question website... that's stupid... :/ no fixes provided so far

fossil hemlock
#

I can connect to my server using the internal ip but not with the external, and i cant find anything else useful searching up key words in the discord

#

and my friends cant connect using the external ip either

orchid vault
#

fix your ports and forwarding an d firewalls and such.. wiki page has tons of info

fossil hemlock
#

all the ports are forwarded

#

i can try again i guess

frosty gale
#

Hi, how do I play a dedicated server?

orchid vault
#

try the wiki link in here

frosty gale
#

where is the link

orchid vault
#

look up

quick mist
#

check the pins

frosty gale
#

what pins

orchid vault
#

pinned messages

#

its the thing that looks like a pin

#

or just read header of the room

quick mist
frosty gale
#

Is there an easy way to set the server IP up with my domain name on my CloudFlare DNS, thx

orchid vault
#

setup server properly and setup ur dns stuff so it points to the server ip

frosty gale
#

Would I need something like SRV records cause it's not a static IP

orchid vault
#

u need to keep the dns shit updated

#

the net is full of how to set those up

frosty gale
#

I have an auto updater

#

Just need to know what the correct SRV is

orchid vault
crystal eagle
#

Anybody experiencing major train signal failure either they stop or crash due to the trains not obeying the signals, is it us in our dedicated server only this happening or happening to everyone? before the latest patch everything was 99% perfectly working, I've had to shut down the server due to all we do is fix damn trains all the time!!!

orchid vault
echo barn
#

Is there a way to configurate the RAM usage of an Dedicated Server?

orchid vault
#

assign ram in linux

#

but not with the server it uses what it needs

echo barn
#

Ah ok

#

THX

dusky smelt
#

My server keeps crashing after some time, any suggestions what I should look at?

#

Its a fairly mature server

#

I haven't crashed while in game, but after a few hours? maybe a day, it crashes

modest glen
#

are experimental servers run off your own pc?

alpine skiff
#

I've had this server running since the experimental release and after I loaded a new save into it my friend is getting this error message when connecting to the IP. Has anyone else experienced this?

unknown.png
formal laurel
tropic wraith
#

train stations, trains, trucks

#

and you will end up there if game decides to zero your position (or truck you are in), under map unfortunately

unreal charm
unreal charm
chilly flume
#

how do you make a dedi server

orchid vault
#

read the topic or pins

fossil hemlock
#

can someone do a vc with me sometime or something? the wiki is all over the place and i just need heeelp

orchid vault
#

ask your question with some info if possible

stoic stream
#

note to self:. do not put tick rate above 60

chrome torrent
crystal eagle
#

I hope the next patch will fix it as I have my server shutdown it’s unplayable we have over 40 trains all we do is fix trains problems cannot play the game as intended

#

Most of the factories shutdown due to no trains coming in or out

orchid vault
#

it should work fine but you might have to do little fiddling with signals like moving slightly closer/away from junction/station

#

and as you narrow areas down u can post screenshots in that room and people will try to help u get it fixed up

stoic stream
#

nice bug on dedicated server, all my conveyor belts are invisible, awesome

sudden cliff
#

@stoic stream above . maybe use a non dedicated server?

stoic stream
#

Just weird bugs.

sonic mural
#

Hello, I've got a couple of questions around dedicated servers and saves/performance
According to the wiki the save file for my dedi server should be in %LOCALAPPDATA%\FactoryGame\Saved\SaveGames\server

#

And my "normal/singleplayer" should be in %LOCALAPPDATA%\FactoryGame\Saved\SaveGames

#

But I don't have a subfolder called server at all