i'm tring to run a nexuiz server on a server box were is only 1 cs server and same webhosting running
but just have a look of this coppy'd out of top:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
15402 crash 15 0 97780 44m 1060 S 37.8 4.4 1:39.65 nexuiz-linux-68
20537 crash 15 0 104m 94m 8396 S 0.7 9.4 99:33.90 hlds_i686
(rest of the out put removed for securty reason and because its in relavant )
for the record the Cs server is hlds_i686 and the other 1 is the nexuiz server.
you are not gowing to tell me dat a 2 slots nexuiz server on standaart setings uses more Cpu dan a 12 slot boosted ( 1000 fps ) server of counter-strike1.6.

so i'm dowing smarting wrong,
or the progammers have to have a big look @ the server side because if this progamma is gowing to get big it maby a advise to have a small but stabbel running server.
because of the hosting company's willing to run server.

The server console:
./nexuiz-linux-686-dedicated
^7Nexuiz Linux 13:23:06 Feb 13 2006
^7Trying to load library... "libz.so.1" - loaded.
^7Compressed files support enabled
^7Added packfile data/data20060214.pk3 (2537 files)
^7Console initialized.
^7Playing registered version.
^7"maxplayers" set to "2"
^7couldn't exec autoexec.cfg
^7Warning: could not find cvar exit_cfg; when expanding alias chmap
^7 exec $exit_cfg; exec game_reset.cfg; exec maps/$1.mapcfg
^7
^7exec <filename> : execute a script file
^7Server using port 26000
^7Server listening on address 0.0.0.0:26000
^7Loaded maps/nexdm01.ent

Greetings Crashzor

Crashzor wrote:you are not gowing to tell me dat a 2 slots nexuiz server on standaart setings uses more Cpu dan a 12 slot boosted ( 1000 fps ) server of counter-strike1.6.

Yes, i am
An empty Nexuiz 1.5.1 server (development version) takes 17% CPU time on my system (a 2.6GHz P4 that is). I guess the cs server is "cheating" and just checking whether there are player when empty. The Nexuiz server doesn't make a difference whether it's empty or not. That means a rather high CPU usage when it's empty, but it also means that it won't increase much when there are players.

its linux is far is i know you can resave cpu use
its running there on a p4 2.4 ghz i think and i think you need to do sameting about it.
but dat is just me i mean its lags on 2.4ghz and no its not the internet 100mbit most be good for 2 slots .
wel a boosted cs server don't lags.
there is sameting wrong here i think the server version can be mutch litler.
wou be great for ppl running it from data centers.
maby releas a server lite version were more of the cpu use wil be done by the client's

Kadaver, are you insane? My 2 slot server uses 2.8 % or something (3GHz P4). And thats with 0.03 ticrate. I'd check, but I forgot to restart the server after the powerout today. I'll let you know when I've checked for sure.

Crashzor wrote:maby releas a server lite version were more of the cpu use wil be done by the client's

Not possible atm, since the client can't execute any quakec gamecode.

tChr wrote:Kadaver, are you insane?

No, not that i'm aware of...

My server here (normal ffa, 6 players, mostly used for testing now as its dsl) uses around 1-5 % cpu when empty. The computer is a amd sempron 2200+ with 512 mb ram. And i really would not expect more from it
Strange that you two have so much more..

KadaverJack wrote:Not possible atm, since the client can't execute any quakec gamecode.

Look i'm dowing this from a users point of view and dan its realy easy.
I don't give shit about the source code i just want it to run on less cpu.
i understand you're trobbel but work on it i also know you're not a wizzard waving you're stik and the cpu use is less but you know what i mean right ?
al also notice the client is only using like 15 % cpu grafish full on a amd 3500+

KadaverJack wrote:No, not that i'm aware of...

1 of the point of begin insane is not notice it you're self

Hmm i do not get why our servers show such different cpu usages.. 1-5% seems fine and it was mostly around 2-3 so maybe nexuiz is having issues with something on you servers? libraries or what ever.. almost 40% is just insane.. although i'm sure KadaverJack is not
Did one of you try to strace nexuiz?

hmm, on our server (Athlon xp2000+, 512 MB RAM), which is running ladder+dreckloch+a t2 server, we have an average cpu use of 20%....

CPU time is a pretty relative value

TOP shows different values than PS, so this is not really a accurate way of telling CPU usage anyway

CPU usage is pretty random for me. it can be 2% and also 35%. a empty server has between 15-25%. I'm pretty sure this thing is heavily bugged and no1 traced it down to the issue yet ^^ (since only the people who cannot do anything about it get the problems, while the developers never have these problems)

I mean, hey, all developers claimed to have fluent gameplay when the choppy movement-fix was not there yet, so I think they are somehow insensible for such things xD

my server hardware is amp sempron 2200+, 512 mb ram, suse 9.3, broken hardware and drivers which cause this high CPU usage of course (developers explanation xD)

perhaps someone could give the people with high CPU usage a special server version that is profiling all the time and logging it to see whether soem functions in the server start going crazy sometimes ?

GreEn`mArine wrote:I mean, hey, all developers claimed to have fluent gameplay when the choppy movement-fix was not there yet, so I think they are somehow insensible for such things xD

It just means the developers never had those problems and if you can't get the same problems you have a damn hard time to solve them..
Guess you know the feeling as you program yourself

GreEn`mArine wrote:TOP shows different values than PS, so this is not really a accurate way of telling CPU usage anyway

Read 'man ps' and 'man top' and you'll see why...
If you don't find it, here's a hint:

man top wrote: k: %CPU -- CPU usage
The task's share of the elapsed CPU time since the last screen update, expressed as
a percentage of total CPU time. In a true SMP environment, if 'Irix mode' is Off,
top will operate in number of CPUs. You toggle 'Irix/Solaris' modes with the 'I'
interactive command.

man ps wrote:CPU usage is currently expressed as the percentage of time spent running during the entire
lifetime of a process. This is not ideal, and it does not conform to the standards that ps
otherwise conforms to. CPU usage is unlikely to add up to exactly 100%.

to skip al the crap there is no way to fix it.

Would you please give (the right) people some time before making such staggering posts??

esteel wrote:

GreEn`mArine wrote:I mean, hey, all developers claimed to have fluent gameplay when the choppy movement-fix was not there yet, so I think they are somehow insensible for such things xD

It just means the developers never had those problems and if you can't get the same problems you have a damn hard time to solve them..
Guess you know the feeling as you program yourself

Yeah I know what you mean, however, from v1.0 until v1.2 every (!) internet server WAS choppy (for everyone), just more or less (sometimes hardly noticable) strong. And somehow this should have been noticable by developers a bit earlier, but this isn't the point of this discussion anyway so I better shut up now .

@Crashzor: Skip all the crap ? Lol, you're pretty polite, eh. However, this thing won't be fixed in a minute, I think it will take a while until the real problem can be found.

I am still thinking that some benchmark-loggin--dedicated-server-versions for people who do have CPU problems would help, but this is best decided by LordHavoc himself

it realy simpel i notice the developers talking around it but i don't FUCKING CARE how the system works, i just want it to work right.
So basicly its i was just look for a awaser like:
1. this progamma is not workt out totaly so yeah its normal
or
2. you can fix it downing this this and this.
so just keep it simpel spare me al the "useless" crap.
if i want to know who the system works i gowing to read the source
and basicly using skip the crap to keep it simpel.
it was a simpel you're fucket and were not gowing to help you( just buy a faster cpu ).
So i don't want the total techincal sidetoer ( i'm not a toerist ) so just get to the point and say yes or no.
keep it simpel so us humans also understand the awnser whitout looking up how the total progamma works.
Greeting
Crashzor

now isnt he nice

Crashzor wrote:it realy simpel i notice the developers talking around it but i don't FUCKING CARE how the system works, i just want it to work right.
2. you can fix it downing this this and this.
so just keep it simpel spare me al the "useless" crap.
if i want to know who the system works i gowing to read the source
and basicly using skip the crap to keep it simpel.
it was a simpel you're fucket and were not gowing to help you( just buy a faster cpu ).
So i don't want the total techincal sidetoer ( i'm not a toerist ) so just get to the point and say yes or no.
keep it simpel so us humans also understand the awnser whitout looking up how the total progamma works.
Greeting
Crashzor

- talking about how the system works is necessary to find a fix

- no one said that they're not going to help you

- You agreed not to post any abusive, obscene, vulgar, slanderous, hateful, threatening, sexually-oriented or any other material that may violate any applicable laws.

Crashzor i had i very bad day yesterday so the reply from me was actually very polite and i rewrote it several times
But i want to make clear to you that the kind of attitute you are showing right now will get you nothing.. This 'Hey i have a problem, i have no idea about it but i need it fixed yesterday' thing is just annoying. People are reading about your problem and everyone that has posted a reply took some time to help.. either with more information, thinking about it or just by finding a solution. As far as i see none of them in directly involved with the engine itself so we need some time till this problem reaches the right persons.

If you read the posts in here you should have noticed that different people have this high cpu usage problem while others do not have it. So i guess the only correct anser right now is 'its this way on your computer currently so you'll have to live with it'.
You can however try one of the newer engines (servers) from here: http://icculus.org/twilight/darkplaces/files/ for example this one http://icculus.org/twilight/darkplaces/ ... 4beta1.zip (currently the latest version). But to be honest i have no idea if it will help. It might even give problems with users that use the 'normal' Nexuiz 1.5. There was some talk about that but i did not follow it too much.. So please test it in case you download and use it.
Once people find time to take a closer look at the problem there might be better answers. Or you can try to go the #DarkPlaces on irc.anynet.org and directly talk to the people that made the engine. BUT PLEASE BE MORE POLITE THERE!!!!!

Oke for the record i maby got a litle to pissed of of this problem you're right about dat maby i coud spare a fuw i don't give a shit etc.
but stil i'm used to forums were there is relativly i same anwaser or just a can't bedone right now or just a were working on:
i'm the persoon dat is more happy whit a short post were working on it dat whit a totaly story about what is maby wrong.
te second problem is its not my box so i can't remove think randomly because its the box of a hosting company and if i remove like the webserver the owner wil get a litle bit pissed, because people play good money for there webhosting.
next time i wil look out to not go over the top whit posting,
anyway thnx for the time.

Ive got my ladder server up again.. so lemme see:

Code: Select all

ps:
nexuiz    2260 28.5  4.3 96800 45364 pts/2   S+   15:11   3:08 ./nexuiz-linux-686-dedicated +exec server_allmaps.cfg
top:
2260 nexuiz    15   0 96800  44m 2944 R 23.8  4.4   3:34.38 nexuiz-linux-68


So it seems I was wrong.. The CPU usage is quite hight.. In my head I must have confused it with the memory consumption.

@Crash: btw, I'd like you to read through your posts before actually posting them as they are full of mistakes and I'm sometimes unsure what you are saying because of your syntax and "spelling" which is incorrect for every second word

GreEn`mArine wrote:@Crash: btw, I'd like you to read through your posts before actually posting them as they are full of mistakes and I'm sometimes unsure what you are saying because of your syntax and "spelling" which is incorrect for every second word

I'm dyslexic so dat makes it harder for me to write correctly and special in English dutch gets hard sometimes so i will try to prevent the error buy reading and using OpenOffice to spell check my postes.

Did anyone bougther to check the sysload when the server(s) are in use? idle load is rather uninteresting.. (sorry if its been posted above, im dead tired and should proly be sleeping insted of posting) All my past (but rather limited) experiances with nexuiz servers was that the load was ridiculously low, but i havent realy checked up on it lately. And do make sure that your sys_ticrate isent set to something crazed like 0.0001


and.. sorry for nagging abt this but

it realy simpel i notice the developers talking around it but i don't FUCKING CARE how the system works, i just want it to work right.

plz dont start that kinda crap talking. this is a very good comunity when it comes to (among other things) keeping the talk above 4'th gtade level.

Crashzor wrote:

GreEn`mArine wrote:@Crash: btw, I'd like you to read through your posts before actually posting them as they are full of mistakes and I'm sometimes unsure what you are saying because of your syntax and "spelling" which is incorrect for every second word

I'm dyslexic so dat makes it harder for me to write correctly and special in English dutch gets hard sometimes so i will try to prevent the error buy reading and using OpenOffice to spell check my postes.

That sounds tedious. Better idea: use Konqueror (included with KDE) as web browser. It can check the contents of web form fields automatically. In this post, it highlighted Crashzor, GreEn`mArine, btw, dyslexic (why that, looks right to me), OpenOffice and postes.

Is there perhaps a Firefox extension which does the same? Would be nice to have...

divVerent wrote:Is there perhaps a Firefox extension which does the same? Would be nice to have...

the Google toolbar has a spell-checker

Something is really wrong and I think it has someting to do with architectures because when I run dedicated server on my PPC 7400 466MHZ (CVS darkplace from today compiled with -mcpu=7400 -O3 -pipe -maltivec -mabi=altivec -fno-strict-aliasing -ffast-math -funroll-loops) without any players it starts out high (Well duh) but soon sink and disapear from top. With ps it sinks slowly.

Also if I run it with 10 bots ps says the load is about 80% after 30 minutes. And when I run it on my 3GHZ P4 the load is about 20% on empty server...

Rizado wrote:Something is really wrong and I think it has someting to do with architectures

Yeah this might be possible. We got Crashzor to compile the engine himself and after that it was at a stable 15% full and empty.. (no bots just players) and i tested it on my little server it was around 3% when empty and at 14% with 5 players (using the binary that came with Nexuiz). So maybe this is more a library problem then architecture..
Its very strange, guess i have to test the svn/cvs version of Nexuiz/DarkPlaces those days to see if they behave better. But as i did not have problems with the 1.5 version maybe others should try too. Pretty please

the way the server handles CPU usage is about to be changed by LordHavoc and div0. my v1.5 Nexuiz server is running at 30% (idling), however, CPU usage does usually not get any higher (have not really monitored that)

however, div0 had a test build which has variables to change some handling-times (how long to wait, and so stuff, didn't understand much of it) and with his help I was able to get a server with fluent movement, 0% load when idling and 10% load with 7 bots and me playing on the server, with full action all the time (and please notice that the AI of the bots also needs CPU time), so there is a great improvement possible

Just wait, it will be in v1.6 for sure I guess

Yeah i noticed the DP changelogs..

rewrote timing code, now a much better and very different sleeping method, no longer wastes any cpu time
changed timing code to have independent sleeping for client and server frames in a listen server
changed console execution to occur in sync with server frames if a server is running, this fixes frikbot loading of .way files
eliminated several host_* variables, replaced host_realframetime with cl.realframetime (which is how long since the last client frame)
removed the clamping of sys_ticrate to >= 0.1, so now sys_ticrate 0 is allowed (run as fast as possible), as well as silly values like 0.001
removed serverprofile cvar as it was not easy to preserve it in the rewritten timing
merged Host_FilterTime, Host_ServerFrame, Host_Frame, _Host_Frame into Host_Main (which also calls Host_Init), this eliminates some duplicate code in all the sys_ modules

So another reason why 1.6 will get great