AlliedModders - sv & var on CS:GO servers

AlliedModders (https://forums.alliedmods.net/index.php)

- Source Servers (SRCDS) (https://forums.alliedmods.net/forumdisplay.php?f=130)

- - sv & var on CS:GO servers (https://forums.alliedmods.net/showthread.php?t=303125)

sv & var on CS:GO servers

Hi,

We have a lot of CS:GO community servers and we have decided to migrate this night to better dedicate servers.
Now, we have higher sv and var on all servers :( (even if we have just one CSGO server started with 20 players on it)

All our servers here: https://meriletfou.fr/en/serveurs/
Last dedicate servers: https://www.soyoustart.com/fr/offres/150game2.xml
New dedicate servers: https://www.soyoustart.com/fr/offres/173game3.xml

our server.cfg:

Code:

//Testing commands

sm_cvar sv_unlag 1

sm_cvar bot_quota 10



host_info_show 2

host_players_show 2



//The amount of time players can join teams after a round has started

mp_join_grace_time 1000

mp_force_pick_time 60



//This should always be set, so you know it's not on

sv_cheats 0



sv_lan 0 



// rates

sm_cvar sv_maxupdaterate 128

sm_cvar sv_minupdaterate 128

sv_mincmdrate 128

sv_minrate 128000

sv_maxrate 0

sm_cvar sv_client_min_interp_ratio "0.001"

sm_cvar sv_client_max_interp_ratio "0.001"



// quality and network settings

sv_alternateticks 0

sv_consistency 0

sv_forcepreload 1

sv_force_transmit_players 0

sv_pure 0

sv_pure_trace 0

net_threaded_socket_burst_cap 4224000

net_threaded_socket_recovery_rate 4224000

net_threaded_socket_recovery_time 1

sm_cvar net_earliertempents 1

sm_cvar net_compresspackets 0

sm_cvar net_maxcleartime 0.001

sm_cvar net_minroutable 1000

sm_cvar net_splitrate 2



sv_maxusrcmdprocessticks 0

sv_client_cmdrate_difference 0

sv_clockcorrection_msecs 15



sm_cvar cl_clockdrift_max_ms 0

sm_cvar cl_clock_correction_adjustment_min_offset 0

sm_cvar cl_clock_correction_force_server_tick 0



sm_cvar mod_dont_load_vertices 1

sm_cvar sv_creationtickcheck 0

sm_cvar sv_playerperfhistorycount 1

sm_cvar sv_threaded_init 1



sm_cvar g_ragdoll_important_maxcount 1

sm_cvar g_ragdoll_maxcount 1

sm_cvar phys_enable_experimental_optimizations 1

sm_cvar prop_active_gib_limit 1

sm_cvar prop_active_gib_max_fade_time 1

sm_cvar sv_pushaway_clientside 2

sm_cvar sv_turbophysics 1

sm_cvar throttle_expensive_ai 1



sm_cvar sv_client_predict 1

sm_cvar sv_client_max_interp_ratio 1

sm_cvar sv_client_min_interp_ratio 1

sm_cvar sv_parallel_packentities 1

sm_cvar sv_parallel_sendsnapshot 1

sm_cvar sv_maxunlag 0.2



// preventing game spoilers settings

sv_allow_votes 1

sv_allow_wait_command 0

sv_dc_friends_reqd 0

sv_pausable 0



// communication settings

sv_alltalk 0

sv_deadtalk 0

sv_full_alltalk 0

sv_gameinstructor_disable 1

sv_voiceenable 0



// Friendlyfire

mp_friendlyfire 0

mp_solid_teammates 1

mp_teammates_are_enemies 1



mp_limitteams 0

mp_autoteambalance 0



mp_autokick 1

mp_do_warmup_period 0

sv_holiday_mode 0



mp_round_restart_delay 1

mp_match_restart_delay 1



//bots

sm_cvar bot_prefix "MERILETFOU.fr"



//**The following commands manage kicks and bans

exec banned_user.cfg

exec banned_ip.cfg

writeid

writeip

starting command:

Code:

SCREEN -dmS csgoshox ./srcds_run -game csgo -console -usercon -tickrate 128 -ip 188.165.81.232 +sv_setsteamaccount **** -port 27020 +game_type 0 +game_mode 1 +mapgroup mg_bomb +map de_dust2 -maxplayers_override 22 -nohltv +sv_pure 0

If you have any idea...

Thanks in advance.

Regards,
Marounda

Re: sv & var on CS:GO servers

that "server.cfg" is horrendous. Why don't you try something a little more default and see how that works out.

Re: sv & var on CS:GO servers

Quote:

Originally Posted by hamilton5 (Post 2562505)

that "server.cfg" is horrendous. Why don't you try something a little more default and see how that works out.

With the time and all reads i generated this one. But i'm not sure everything is still necessary...

Re: sv & var on CS:GO servers

server.cfg is ok, nothing to worry about.
the new hadware is no concern for performance.
if server settings are preserved, the only thing i'm thinking about is overselling.

Re: sv & var on CS:GO servers

Have you tried changing your governor to performance?
Have you tried changing your kernel to hz1000?

Re: sv & var on CS:GO servers

Hi,

Sorry for the late of my answer...

Quote:

Originally Posted by sphinx (Post 2562812)

server.cfg is ok, nothing to worry about.
the new hadware is no concern for performance.
if server settings are preserved, the only thing i'm thinking about is overselling.

Thank you for the answer, my sysadmin said me that we had a problem with an update of the kernel, so it's looks like better with the last updates now, even if i would like better performances, I think it's acceptable atm.

Quote:

Originally Posted by Maxximou5 (Post 2562826)

Have you tried changing your governor to performance?
Have you tried changing your kernel to hz1000?

Thanks for the answer too, 1000hz necessary with CS:S but it looks like 1000hz is not a good way, am i wrong ? Need I to choose a 1000hz kernel ?
I'll ask to my sysadmin if we are in performance or no...

Re: sv & var on CS:GO servers

There is not a benefit to using a different kernel hertz than the default. Previously it made a difference, but the high resolution timers are implemented in such a way that it doesn't affect things much any more in terms of game server performance.

Re: sv & var on CS:GO servers

Quote:

Originally Posted by sphinx (Post 2564563)

Ok, thank you to confirm this!

Re: sv & var on CS:GO servers

Quote:

Originally Posted by sphinx (Post 2564563)

I still find a performance gain even with a 6700k on Debian 9. I've got a 4790k as well doing the same. I'll see about doing some data logging and switching them back and see if there actually is any difference between the two. As per my last check there was a variance.

Re: sv & var on CS:GO servers

The switch from 4790k to 6700k is not an improvement at all in my opinion.
When you look at the single thread performance you see that 4790k still outrun the 6700k.

Source:https://www.cpubenchmark.net/singleThread.html