Go Back   nV News Forums > Linux Support Forums > NVIDIA Linux

Newegg Daily Deals

Reply
 
Thread Tools
Old 07-03-04, 01:46 PM   #13
arokh
Registered User
 
arokh's Avatar
 
Join Date: Jun 2004
Posts: 287
Default Re: Slow performance with 6106

Quote:
Originally Posted by Doogman
Yes, but ut2004, et, q3 and every other 3D game I have works fine. I just tried Wolfenstein in a window instead of fullscreen and it still hits my framerate cap (90 fps). Everything seems to work well except this goofy glxgears.
You haven't by any chance turned vsync on?
__________________
Asus A8N32-SLI : AMD64 3700+ @ 2.85GHz : 2GB PC4000 OCZ Platinum EB : 2x 7800GT SLI
arokh is offline   Reply With Quote
Old 07-03-04, 02:30 PM   #14
Kamel
Hardware Mass Murderer
 
Kamel's Avatar
 
Join Date: Oct 2003
Location: Jacksonville, FL
Posts: 1,035
Send a message via ICQ to Kamel Send a message via AIM to Kamel Send a message via MSN to Kamel Send a message via Yahoo to Kamel Send a message via Skype™ to Kamel
Default Re: Slow performance with 6106

Quote:
Originally Posted by Doogman
Yes, but ut2004, et, q3 and every other 3D game I have works fine. I just tried Wolfenstein in a window instead of fullscreen and it still hits my framerate cap (90 fps). Everything seems to work well except this goofy glxgears.
so... what's wrong? who cares if glxgears don't work but everything else does.. go play some games and be happy, lol.

i suppose i should add a disclaimer to what i mentioned previously though... there's thousands of things that could cause glxgears to run slowly, and they might not always be affecting other opengl applications.
__________________
hardware murdered so far: geforce 5900u, deceised: sat on it. geforce 6600gt deceised: improperly mounted heatsink. 3 asus a7n8x-e deluxe motherboards, deceised: unsure. mobile amd xp2600 processor, deceised: chipped core due to compusa heatsink. dfi lanparty ultra b, deceised: unsure, third memory bank went out. samsung cd burner 48x40x48, deceised: unsure, will only read pressed cd's very slowly. samsung TH552C dvd burner (dual layer), deceised: same as other samsung . 10gb seagate 5200rpm drive, deceised: tried fixing a pin that was pushed into the drive by inserting an ide cable upside down. 40gb maxtor drive, deceised: maxtor manufacturer. 80gb western digital, deceised: unnatural causes. swan xt10, deceised: brutal pre-meditated murder, even video taped (it wasn't y2k compliant). logitech mx510, deceised: death by electricution. ati remote wonder, deceised: death by electricution.
Kamel is offline   Reply With Quote
Old 07-03-04, 02:31 PM   #15
Kamel
Hardware Mass Murderer
 
Kamel's Avatar
 
Join Date: Oct 2003
Location: Jacksonville, FL
Posts: 1,035
Send a message via ICQ to Kamel Send a message via AIM to Kamel Send a message via MSN to Kamel Send a message via Yahoo to Kamel Send a message via Skype™ to Kamel
Default Re: Slow performance with 6106

Quote:
Originally Posted by gagent
Below's the glxinfo output. I don't see anything about MESA. What confuses me it that GeForce2 ti performes so much better with the same setup. So I guess the setup is ok.
glxinfo looks good... hmm, i really can't think of many other things atm that would cause this.
__________________
hardware murdered so far: geforce 5900u, deceised: sat on it. geforce 6600gt deceised: improperly mounted heatsink. 3 asus a7n8x-e deluxe motherboards, deceised: unsure. mobile amd xp2600 processor, deceised: chipped core due to compusa heatsink. dfi lanparty ultra b, deceised: unsure, third memory bank went out. samsung cd burner 48x40x48, deceised: unsure, will only read pressed cd's very slowly. samsung TH552C dvd burner (dual layer), deceised: same as other samsung . 10gb seagate 5200rpm drive, deceised: tried fixing a pin that was pushed into the drive by inserting an ide cable upside down. 40gb maxtor drive, deceised: maxtor manufacturer. 80gb western digital, deceised: unnatural causes. swan xt10, deceised: brutal pre-meditated murder, even video taped (it wasn't y2k compliant). logitech mx510, deceised: death by electricution. ati remote wonder, deceised: death by electricution.
Kamel is offline   Reply With Quote
Old 07-04-04, 03:12 PM   #16
Doogman
Registered User
 
Join Date: Jul 2004
Posts: 21
Default Re: Slow performance with 6106

Well, I fixed my "problem" with glxgears. It wasn't the graphics driver at all, it's actually the new 2.6 scheduler and the dc project folding at home (f@h) client. The f@h client is "niced" to 19, which usually means it gets very low priority and mostly idle CPU cycles. With the new scheduler, it's going to get cycles regardless, which seems to play havoc with the goofy glxgears benchmark. I killed 1 of the 2 f@h processes (I have a SMP box), and glxgears went from ~170 fps to ~3400 fps which is quite the difference.

With all my old games (et, q3) I had plenty of CPU power to spare and it's only since I've started playing ut2004 which can handle everything this box has that I've noticed how much the new scheduler gives to even highly niced apps. Anyway, the solution to this is to try another scheduler which idle batch, but gotta make sure what I'm running now is stable.
Doogman is offline   Reply With Quote
Old 07-04-04, 08:47 PM   #17
how
Registered User
 
Join Date: Dec 2002
Location: Australia
Posts: 142
Default Re: Slow performance with 6106

what scheduler is it? if not the kernel anticipatory one.
how is offline   Reply With Quote
Old 07-05-04, 08:55 AM   #18
Doogman
Registered User
 
Join Date: Jul 2004
Posts: 21
Default Re: Slow performance with 6106

http://members.optusnet.com.au/ckolivas/kernel/

That is the web page with the kernel patch. I'm interested in schedbatch. By the way, I'm talking about the CPU scheduler, not the disk I/O scheduler or "elevator."
Doogman is offline   Reply With Quote
Old 07-05-04, 09:04 AM   #19
how
Registered User
 
Join Date: Dec 2002
Location: Australia
Posts: 142
Default Re: Slow performance with 6106

I already tried the ck4 patch for kernel 2.6.7. performance is the same no increase at all
how is offline   Reply With Quote
Old 07-05-04, 08:07 PM   #20
Doogman
Registered User
 
Join Date: Jul 2004
Posts: 21
Default Re: Slow performance with 6106

Performance is the same for _what_?

Using schedtool to mark the F@H tasks scheduled_batch certainly helps out a great deal. Glxgears works the same with both F@H processes running as with none. It helped Ut2004 a little; perhaps I will have to kill F@H now but I need to test on a few more levels.
Doogman is offline   Reply With Quote

Old 07-15-04, 12:24 PM   #21
Nubi
Registered User
 
Join Date: Jul 2004
Posts: 2
Default Re: Slow performance with 6106

Doogman could you please explain in newbie terms what you are talking about, and step by step how you implemented your solution? That would help me out a lot because Im having the same isssue, every game runsbut glxgears performance sucks.
Nubi is offline   Reply With Quote
Old 07-17-04, 09:02 AM   #22
Doogman
Registered User
 
Join Date: Jul 2004
Posts: 21
Default Re: Slow performance with 6106

First off, if all your games are working fine with expected FPS don't worry about glxgears, it's a silly benchmark and isn't very accurate.

Anyway, on my computer I donate my spare CPU cycles to the Folding at Home Project (f@H) Link, so both my processors (I have a 2 CPU box) are utilized at 100%. Now, the f@h processes are both "niced" to the maximum value of 19, which means they will only take a 19th of the processor usage of a normal (foreground). Basically, this means that your normal foreground apps and such seem to run at normal speed. BUT, on this glxgears thingy, it seems that even highly "niced" programs like f@h will greatly interfere with the glxgear's FPS score.

One solution: stop all other background processes. Not interested. Another is to Con Kolivas's kernel patch here and the program schedtool to set the f@h processes to sched_batch; this means that they will ONLY use idle cycles, not the 19th percent of a foreground app.

Help any? Well, normally:

[doug@ghidorah doug]$ glxgears
622 frames in 5.0 seconds = 124.400 FPS
778 frames in 5.0 seconds = 155.600 FPS
781 frames in 5.0 seconds = 156.200 FPS
771 frames in 5.0 seconds = 154.200 FPS

And now I set both f@h processes to sched_batch

[doug@ghidorah doug]$ glxgears
14344 frames in 5.0 seconds = 2868.800 FPS
17142 frames in 5.0 seconds = 3428.400 FPS
17312 frames in 5.0 seconds = 3462.400 FPS
17306 frames in 5.0 seconds = 3461.200 FPS
17260 frames in 5.0 seconds = 3452.000 FPS

Mind you this is all processor usage!

So, if you are really worried about getting your glxgears FPS up, understand the this benchmark is definately influenced by CPU availability and latency.
Doogman is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


Similar Threads
Thread Thread Starter Forum Replies Last Post
Boost Your Performance Goals 10x ' This Week on inside* Publications News Archived News Items 0 05-13-12 05:20 PM
NVIDIA Unleashes the GeForce GTX 670 Graphics Card ' Performance Perfected (WCCFTECH) News GeForce GTX 670 Reviews 0 05-10-12 08:40 AM
Very slow X startup Jeremy NVIDIA Linux 96 05-23-03 10:11 AM
My UT2003 Tweak Guide DXnfiniteFX Gaming Central 48 10-30-02 11:59 PM
GeForce4 Ti 4600 performance in UT2003? Turnbolt NVIDIA GeForce 7, 8, And 9 Series 9 08-17-02 08:50 PM

All times are GMT -5. The time now is 03:12 PM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.