Lowering FPS, reducing tickrate - will it have any effect's in-game?

[Mike]

Hi,

I run two 24 slot servers on my Core2Duo machine, however the FPS and tickrate bumps around a lot. Will "limiting" the fps_max variable to 33.33, reducing the rates to the same and by doing that reducing the tickrate have any negative effect in-game, like poor hitregistering and choke etc..?
I've installed a custom kernel and since i did that it's slightly more stable, but i sometimes get chokespikes, and due to me not having a other computer to test it on i dont know if its my internet connection or if its the server.

Thanks,
Mike (former michael_sj123)

[CentralFrag]

Well anytime you lower the rates and the frame rates of course the registration will decrease but I mean people played that way for years so I mean its nothing that will be unplayable. My best advice is just to play around with it and see what works best for your machine and your setup.

[click4dylan]

better off to limit both -tickrate (with an enabler) and the fps. i run two servers and 99% of the people can't notice a difference between 40 ticks and 66. if reducing choke is what you want, increase sv_maxrate to 50000 and set your rate to 50000 and ensure your fps is equal to the updaterate or higher

[BehaartesEtwas]

there is no difference in reducing the tickrate with the -tickrate option (if available) or with the sv_maxupdate/cmdrate commands and/or fps_max. I suspect fps_max 33.33 will be the best option in case your server cannot maintain stable higher fps. what kernel did you install, and what are your cpu specs (post output of cat /proc/cpuinfo)?

[Mike]

Code:

processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 15
model name      : Intel(R) Core(TM)2 Duo CPU     E4600  @ 2.40GHz
stepping        : 13
cpu MHz         : 2394.070
cache size      : 2048 KB
physical id     : 0
siblings        : 2
core id         : 0
cpu cores       : 2
apicid          : 0
initial apicid  : 0
fpu             : yes
fpu_exception   : yes
cpuid level     : 10
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov                                                                                                                                                              pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant                                                                                                                                                             _tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl es                                                                                                                                                             t tm2 ssse3 cx16 xtpr pdcm lahf_lm dts
bogomips        : 4788.14
clflush size    : 64
cache_alignment : 64
address sizes   : 36 bits physical, 48 bits virtual
power management:

processor       : 1
vendor_id       : GenuineIntel
cpu family      : 6
model           : 15
model name      : Intel(R) Core(TM)2 Duo CPU     E4600  @ 2.40GHz
stepping        : 13
cpu MHz         : 2394.070
cache size      : 2048 KB
physical id     : 0
siblings        : 2
core id         : 1
cpu cores       : 2
apicid          : 1
initial apicid  : 1
fpu             : yes
fpu_exception   : yes
cpuid level     : 10
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov                                                                                                                                                              pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant                                                                                                                                                             _tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl es                                                                                                                                                             t tm2 ssse3 cx16 xtpr pdcm lahf_lm dts
bogomips        : 4787.98
clflush size    : 64
cache_alignment : 64
address sizes   : 36 bits physical, 48 bits virtual
power management:

i followed your guide and compiled a kernel.

edit: here's some fpsmeter measurements:
#2
#1

taken when they was both full.

[BehaartesEtwas]

I would keep the fps at 66, that looks fine. keep in mind, that round restarts, player connects etc always will produce some fps drops. I don't think you really notice those fps variations ingame.

[Mike]

Ok. However, what about the choke spikes? I can see the lerp going yellow and then the "sv" drops to 20 and below sometimes and then it haults the game for a second, and then it resumes with 60 choke and then the choke goes down.

[BehaartesEtwas]

choke has nothing to do with fps. make sure all net settings (both server and client side) are correct and that you are not using any other special settings (in fact, valve defaults are pretty much optimal - change only what you really understand). if that doesn't help, your problem might be connection related.

[Mike]

PHP Code:

sv_mincmdrate "30"
sv_maxcmdrate "66"
sv_minupdaterate "30"
sv_maxupdaterate "66"
sv_minrate "30000"
sv_maxrate "60000"
sv_client_max_interp_ratio "3"
sv_client_min_interp_ratio "2"
sv_client_cmdrate_difference "20" 


My netsettings

[BehaartesEtwas]

that looks fine. make sure you dont have other disturbing settings. i.e. try with valve default (but server name and passwords) both on server and client side.

[Mike]

Everything set to default, only roundtimes changed. No difference. Client is set to default too. I also tried unloading all the sm plugins and that made no difference either. But i think like, when the lagspike is happening, the "SV" in netgraph drops to below 33 and the lerp turns yellow. And then it goes back to the normal.

[BehaartesEtwas]

ok, maybe the fps meter underestimates the fps drops (in some scenarios the stats command is more unlikely to be executed during a low-fps frame). did you try any kind of Linux optimizations? If not, you find a howto in my signature...

[Mike]

As stated, i did everything in your guide.

[Mike]

Just bumping this because problem did not resolve itself...

[BehaartesEtwas]

maybe your cpu isn't sufficient... I don't know.

what kernel do you actually have? and did you check your servers are really running with elevated priority (FIFO or RR scheduling)?