You are not logged in.
Pages: 1
Enemy Territory: Quake Wars performs a lot worse on arch than it does on ubuntu. Even with lower quality settings frequent graphics lags occur on arch though I never even noticed one of these kind of lags on ubuntu.
I'm using the most recent catalyst (7.12) on both OSes.
Shouldn't arch actually perform better than ubuntu due to its i686 optimization (and a lot less background-tasks)?
I didn't find anything concerning that issue in the wiki, so I'd like to know if there's a way to improve my gaming performance on arch.
Thanks for your help in advance.
Offline
Have you setup Xorg to use your actual graphics card and not just using the generic vesa module?
Offline
Have you setup Xorg to use your actual graphics card and not just using the generic vesa module?
It wouldn't be running at all without hardware acceleration.
Offline
Try the Ubuntu kernel in Arch.
There's lots of other potential tweaks - this is just one of them.
Offline
xorg is setup correctly.
i'd like to test the ubuntu kernel, but my ubuntu install is as 64bit one, so i just copying it is out of the question.
i'm gonna try compiling the ubuntu kernel but this can take a while as my knowledge about kernel compilation is borderline zero and i do not intend to spend too much time with my pc during the holidays.
however i don't think using the ubuntu kernel on arch is a proper way of experiencing arch's superior performance.
Offline
See my howto regarding compiling the kernel.
This doesn't work by magic, ya know. Code that fixes bugs is good code - it doesn't matter whether it comes from the vanilla kernel, Arch's patches, Ubuntu's patches, or wherever.
Offline
i'm gonna try compiling the ubuntu kernel but this can take a while as my knowledge about kernel compilation is borderline zero and i do not intend to spend too much time with my pc during the holidays.
Once you compile the first kernel that actually works, you just forget how difficult it was and you can compile it with your eyes closed.
Offline
im just on creating a .config from scratch.
got a question: do you know any tool that helps you create a made-to-measure kernel-config (i.e. automatically deselecting drivers that are not required by one's system and optimizing the config for a high-performance desktop system) or a proper example config?
Offline
I've been reading a lot of posts in the forum complaining about poor performance in games. Some suggested that there was a problem with the cfs scheduler ? in the last few kernels ? can anyone confirm that ?
Offline
I've been reading a lot of posts in the forum complaining about poor performance in games. Some suggested that there was a problem with the cfs scheduler ? in the last few kernels ? can anyone confirm that ?
I run a custom kernel in both ubuntu and arch. ETQW runs great in ubuntu, but in arch I can't even get the graphics to load. Maybe that's because I am using the new nvidia beta drivers. That's more of a headache then I wanna go through, u can't have nvidia beta drivers installed for one kernel and just nvidia for the next. Tells me I got a conflict because of nvida-beta-utils. can I safetly avoid this conflict and work around it?
Offline
i just spent probably two hours configuring that son-of-a-bitch-kernel, just to fail with a bloody kernel-panic on boot up (sth about it not being able to find my root-partition) ... i'm done for now ... gonna give it a try after christmas.
Offline
ok, i got past that nasty sata-controller-driver-missing-issue, which caused the last kernel panic.
however i got a new problem which looks like i won't be able to fix it without help (made a quick search via google, but all i found was unanswered posts or non-helping replies).
That's all I'm being given to figure out what's causing the problem, but i don't even have a clue about how to solve it:
run-init: nuking initramfs contents: Directory not empty
Kernel panic - not syncing: Attempted to kill init!
Thanks for your help, again!
Offline
i got to admit, that discussing two topics in one thread is stupid (so just forget my kernel-compilation issues, as i made some further investigations concerning the cfs scheduler-thingy).
according to kerneltrap.org the cfs-scheduler was as effective as the old sd scheduler or even better in some situations. as i read through that mail-log and the following comments i stumbled upon a post, stating how wonderful it would be when schedulers would be no longer required as hyperthreading- and multicore-technology is in rapid development. this was kinda thought-provoking.
as i got a dualcore processor i tried setting all processes' (except for etqw) cpu affinity to the 1st core and only etqw to the 2nd core. the result was just amazing. while sharing a core with the other processes* (and the scheduler hence distributing its ressources) etqw was, even when niced to -20, lagging. using it's own core it went as fluently is it does on ubuntu.
this kinda viciates the statements made on kerneltrap, but the only logic explanation for that behaviour would be that the scheduler is causing the trouble.
any objections to my conclusion?
*all i had running at that time was system daemons, xorg, openbox and etqw.
Offline
Do you have 64 bit arch OS version, or are u using 32 bit?
Also, can u give me a list of the libraries that are being used when u play ETQW?
I am having a crashing problem and I just can't figure out why in the hell its happening, I get absolutely no errors, just a seg fault and that's it. Doesn't even tell me why its seg faulted.
Offline
i tried running etqw on arch64, but it siply didnt want to (but i cant tell you why, any more ... forgot it) ... thats why i switched to arch32.
if you'd tell me how to find out which libs are used, i'll post you the list asap.
EDIT:
I figured it out myself. Thats what etqw uses on my 64bit ubuntu;
COMMAND PID USER FD TYPE DEVICE SIZE NODE NAME
etqw.x86 14074 user mem REG 8,1 4254420 7225447 /home/user/.etqwcl/base/compiledscriptx86.so
etqw.x86 14074 user mem REG 8,1 862500 7225451 /home/user/.etqwcl/pb/pbcl.so
etqw.x86 14074 user mem REG 8,1 2876036 7225432 /home/user/.etqwcl/pb/pbsv.so
etqw.x86 14074 user mem REG 8,1 6764 4210746 /lib32/libnss_mdns4_minimal.so.2
etqw.x86 14074 user mem REG 8,1 9620444 7225428 /home/user/.etqwcl/base/gamex86.so
etqw.x86 14074 user mem REG 8,1 30436 4210704 /lib32/libnss_compat-2.6.1.so
etqw.x86 14074 user mem REG 8,1 29264 7225434 /home/user/.etqwcl/pb/pbag.so
etqw.x86 14074 user mem REG 8,1 67408 4210717 /lib32/libresolv-2.6.1.so
etqw.x86 14074 user mem REG 8,1 38420 4210706 /lib32/libnss_files-2.6.1.so
etqw.x86 14074 user mem REG 8,1 34352 4210708 /lib32/libnss_nis-2.6.1.so
etqw.x86 14074 user mem REG 8,1 83712 4210703 /lib32/libnsl-2.6.1.so
etqw.x86 14074 user mem REG 8,1 2176700 7225404 /home/user/Spiele/ETQW/libCgx86.so
etqw.x86 14074 user mem REG 8,1 17462688 10485775 /usr/lib32/dri/fglrx_dri.so
etqw.x86 14074 user mem REG 8,1 485888 10371482 /usr/lib32/libGL.so.1.2
etqw.x86 14074 user mem REG 8,1 32800 10372671 /usr/lib32/libXcursor.so.1.0.2
etqw.x86 14074 user mem REG 8,1 22136 10372684 /usr/lib32/libXrandr.so.2.1.0
etqw.x86 14074 user mem REG 8,1 17884 4210705 /lib32/libnss_dns-2.6.1.so
etqw.x86 14074 user mem REG 8,1 30624 4210723 /lib32/librt-2.6.1.so
etqw.x86 14074 user mem REG 8,1 5384 10372967 /usr/lib32/gconv/ISO8859-1.so
etqw.x86 14074 user mem REG 8,1 16616 10372673 /usr/lib32/libXdmcp.so.6.0.0
etqw.x86 14074 user mem REG 8,1 6988 10372667 /usr/lib32/libXau.so.6.0.0
etqw.x86 14074 user mem REG 8,1 59832 10372590 /usr/lib32/libdirect-0.9.so.25.0.0
etqw.x86 14074 user mem REG 8,1 20388 10372592 /usr/lib32/libfusion-0.9.so.25.0.0
etqw.x86 14074 user mem REG 8,1 354504 10372591 /usr/lib32/libdirectfb-0.9.so.25.0.0
etqw.x86 14074 user mem REG 8,1 807584 10371834 /usr/lib32/libasound.so.2.0.0
etqw.x86 14074 user mem REG 8,1 9684 4210700 /lib32/libdl-2.6.1.so
etqw.x86 14074 user mem REG 8,1 1339816 4210697 /lib32/libc-2.6.1.so
etqw.x86 14074 user mem REG 8,1 224402 7225405 /home/user/Spiele/ETQW/libgcc_s.so.1
etqw.x86 14074 user mem REG 8,1 149332 4210701 /lib32/libm-2.6.1.so
etqw.x86 14074 user mem REG 8,1 4387755 7225407 /home/user/Spiele/ETQW/libstdc++.so.6
etqw.x86 14074 user mem REG 8,1 56156 10372674 /usr/lib32/libXext.so.6.4.0
etqw.x86 14074 user mem REG 8,1 986540 10372666 /usr/lib32/libX11.so.6.2.0
etqw.x86 14074 user mem REG 8,1 140058 7225406 /home/user/Spiele/ETQW/libjpeg.so.62
etqw.x86 14074 user mem REG 8,1 80472 10373408 /usr/lib32/libz.so.1.2.3.3
etqw.x86 14074 user mem REG 8,1 111699 4210711 /lib32/libpthread-2.6.1.so
etqw.x86 14074 user mem REG 8,1 420760 10372655 /usr/lib32/libSDL-1.2.so.0.11.0
etqw.x86 14074 user mem REG 8,1 14128 10372675 /usr/lib32/libXfixes.so.3.1.0
etqw.x86 14074 user mem REG 8,1 28744 10372685 /usr/lib32/libXrender.so.1.3.0
etqw.x86 14074 user mem REG 8,1 117340 4210694 /lib32/ld-2.6.1.so
Last edited by Actium (2007-12-26 23:35:16)
Offline
ty actium, that is exactly what I was looking for. Cheers, even though it doesn't look like it will help me.
I just got a note from the programmer of of the linux ETQW client and he seems to think my problem is arch related. He thinks it is a broken package, but as we all know the dev's won't help get 32 bit apps running on 64 bit arch so I guess I am screwed.
Its a shame really, the whole hardware/software bs. Force people to buy 64 bit hardware and then no one wants to ever support old outdated software. They just want you to run 32 bit OS irregardless of the hardware u have. That in my opinion is the wrong one, but what do I know? I am only the end user who is getting disgruntled over the fact that everyone wants to force to use this and that. That's why I stopped using microsoft and joined linux, maybe that was a mistake too?
Offline
About low fps, the cpu affinity could be the solution.
FPS
People with dual core cpu's might need to turn off affinity on 1 core (through the windows taskmanager or
automated using this). Nvidia users may need to change a setting in the nvidia control panel to disable
threaded optimization and/or setting Multi-display/Mixed GPU acceleration to "Single display performance mode".
Also, make sure you have the newest drivers from either nvidia.com or amd.com (ATI).
To get better fps on slow machines, try disabling anti aliasing and anisotropic filter in your video card control
panel. Also in the Urban Terror system menu, try a lower resolution, and low texture quality. Turning on
'Compress textures' helped some people too.
That's from Urban Terror faq. (http://www.urbanterror.net/faq.html)
I don't know if it's true because I don't have dual core.
Offline
Its a shame really, the whole hardware/software bs. Force people to buy 64 bit hardware and then no one wants to ever support old outdated software. They just want you to run 32 bit OS irregardless of the hardware u have. That in my opinion is the wrong one, but what do I know? I am only the end user who is getting disgruntled over the fact that everyone wants to force to use this and that. That's why I stopped using microsoft and joined linux, maybe that was a mistake too?
they don't want to force you to do anything. it is simply much easier to maintain just a 32 bit client instead of a 32 and 64bit one. and as 64bit hardware is backwards compatible and 32bit hardware will never be forwards compatible (would be kinda ridiculous, anyway) the choice, which one to chose, is easy.
as soon as 64bit hardware is widespread enough and its advantages prevail against the disadvantages of the application not being able to run on old hardware, the commercial software development studios will probably switch quickly.
Offline
Do you have 64 bit arch OS version, or are u using 32 bit?
Also, can u give me a list of the libraries that are being used when u play ETQW?
I've been running it on arch64, here's the list...
[etqw]$ ldd32 ./etqw.x86
linux-gate.so.1 => (0xffffe000)
libSDL-1.2.so.0 => /opt/lib32/usr/lib/libSDL-1.2.so.0 (0xf7eeb000)
libpthread.so.0 => /opt/lib32/lib/libpthread.so.0 (0xf7ed3000)
libz.so.1 => /opt/lib32/lib/libz.so.1 (0xf7ec1000)
libjpeg.so.62 (0xf7ea3000)
libX11.so.6 => /opt/lib32/usr/lib/libX11.so.6 (0xf7dba000)
libXext.so.6 => /opt/lib32/usr/lib/libXext.so.6 (0xf7dab000)
libstdc++.so.6 (0xf7cc9000)
libm.so.6 => /opt/lib32/lib/libm.so.6 (0xf7ca4000)
libgcc_s.so.1 (0xf7c99000)
libc.so.6 => /opt/lib32/lib/libc.so.6 (0xf7b63000)
libdl.so.2 => /opt/lib32/lib/libdl.so.2 (0xf7b5e000)
/lib/ld-linux.so.2 (0xf7fa3000)
libxcb-xlib.so.0 => /opt/lib32/usr/lib/libxcb-xlib.so.0 (0xf7b5c000)
libxcb.so.1 => /opt/lib32/usr/lib/libxcb.so.1 (0xf7b44000)
libXau.so.6 => /opt/lib32/usr/lib/libXau.so.6 (0xf7b41000)
libXdmcp.so.6 => /opt/lib32/usr/lib/libXdmcp.so.6 (0xf7b3c000)
Hope it helps.
Offline
libc.so.6 => /opt/lib32/lib/libc.so.6
I had to downgrade to 2.6 from 2.7 to fix my problem. Posted here in another thread. If u are using arch64 and libc-2.7 32 bit compatibility library then maybe we can research this problem and little more.
Do you have a 8k series nvidia graphics card? Are u running a single or dual core processor? Which version of the nvidia drivers are you using?
Last edited by jacko (2007-12-29 16:32:10)
Offline
If u are using arch64 and libc-2.7 32 bit compatibility library then maybe we can research this problem and little more.
Do you have a 8k series nvidia graphics card? Are u running a single or dual core processor? Which version of the nvidia drivers are you using?
Have lib32-glibc 2.7-7 installed,
Don't have 8k card,
Using a single core cpu,
Am running nvidia 100.14.19-6 drivers.
Offline
then the difference is between the dual core and the 8k series card. Everything else seems the same.
I will pass this info on to the programmer of ETQW. Maybe it will help them track down the specific bug that is causing dual core 8k series nvidia GPU setups to freeze while using libc-2.7 32 bit compatibility library on 64 bit OS's only.
Last edited by jacko (2007-12-30 04:15:32)
Offline
Pages: 1