Jump to content
Eternal Lands Official Forums
Sign in to follow this  
raila

update nvidia graphic 169.07

Recommended Posts

hello,

i updated my suse linux 10.3 with the new nvidia graphic - driver 169.07.

while trying to start the client, el-linux.bin starts, stops, and using full power of the processor.

not even a failure in error_log.txt posted. :devlish:

system was working fine before the update - no further changings of the configuration done.

sombody with similar problems?

thanks

raila

Share this post


Link to post
Share on other sites

Are these the beta drivers? If so, these don't seem to work at all on EL. I tried them and get the same problem. /var/log/Xorg.0.log and /var/log/messages also show no error messages

 

I am running the most up to date non-beta drivers (NVIDIA-Linux-x86-100.14.19-pkg1.run) and everything is running great.

Share this post


Link to post
Share on other sites

Are these the beta drivers? If so, these don't seem to work at all on EL. I tried them and get the same problem. /var/log/Xorg.0.log and /var/log/messages also show no error messages

 

I am running the most up to date non-beta drivers (NVIDIA-Linux-x86-100.14.19-pkg1.run) and everything is running great.

 

 

 

it's an official release!

http://www.nvidia.com/object/linux_display_ia32_169.07.html

intersting is, there is no failure on the rest of my opengl applications, all are running nice.

even glxinfo is reporting no failure

just the EL-client still won't start ...

Share this post


Link to post
Share on other sites

You are correct. It also lists that same release under their beta drivers, that's where I got confused.

 

Maybe someone here with a good OGL programming background here can point us to a way to enable debugging in these drivers and we can maybe figure out what's going on here and find a workaround.

 

http://www.nvidia.com/object/linux_display..._100.14.19.html These are the lastest driver for Nvidia under linux that I have found to work with EL. Good luck :D

Share this post


Link to post
Share on other sites

I think something is gone a little crazy at nvidia.

 

I currently have forceware 163.75 installed.

On their site the main page list 163.75 as the current. If you do a search for your

card you get 169.21. Also if you look for Beta drivers you get 169.21 and forceware 163.75

is NOT listed as an archived driver.

 

I tried the fubar'd 169.21 and el loads but the bmp background and then the following green text, is

garbled in a mosiac like pattern. I rollback to the 163.75 and works fine again.

 

So I wonder did nvidia accidently release beta drivers as official?

Share this post


Link to post
Share on other sites

I am running 100.14.19 drivers, and am having the same problems i have referred to in other threads, ie the freeze on window focus change etc. This is the latest "stable" offering from nvidia.

 

I tend to agree with robotbob, something is up at nvidia, someone has borked something.

 

S.

Share this post


Link to post
Share on other sites

Same problem with 169.07, i've been debugging the cvs sources and discovered the issue in gl_init.c, the call for check_fbo_formats() near the end of init_gl_extensions function seems triggering an endless loop within library code. I'm not sure it's due to e-l code (which seems quite fine) or a bug in nvidia code.

Commenting out this call will helped me running the game.

 

I hope this will help.

Marco

Edited by lama

Share this post


Link to post
Share on other sites

I contacted Nvidia about this issue, hope we'll be able to hear from them.

Meanwhile, is there anyone with this driver that doesn't have problems?

 

Win2k SP4, 1 Gig RAM, Nvidia 6600LE PCI-Express 256mb

 

Current Driver Version: 93.71 (came with card,install cd)

 

I had tried to beta drivers, forceware, with some improvement in performance.

I was doing some testing/diagnostics, to see what the best perfromance I can get

from the game.

 

This seemed to be most stable, not with just this game, but others as well.

I dont overclock anything, as I choose reliability over speed, but, I tried a few things.

 

The discovery in which I found (for this game):

 

In choosing from the pre-configured profiles, the best one I have found sofar, is GLStress.

With doing this, the game performance has significantly increased. I dont generally see

any gfx anomalies or distortions, but there was one. Strange vert/hori white lines, displayed

within the general viewing area. I encountered this, with GLStress and reflections turned on.

That being, the only thing I see, with that profile selected. This was a one time occurance, so

the update may have fixed it.

 

I have kept the setting on GLStress with no problems, in which, with deafult settings, the game

doesnt perform quite as well.

 

Well famed CoolBits, works for most situations, where performance vs. stable is an issue. The

option to see the GL tweaks/settings is enabled. (you have to check with your unique card)

I am not suggesting overclock, infact, I am suggest NOT to overclock. CoolBits just makes the

options available.

 

Linux users, not sure this will work on your machine, so you are on your own with that.

 

WARNING: dont play with overclock settings if you dont know what you are doing, as it can

damage your video card, beyond repair.

Share this post


Link to post
Share on other sites

Same problem with 169.07, i've been debugging the cvs sources and discovered the issue in gl_init.c, the call for check_fbo_formats() near the end of init_gl_extensions function seems triggering an endless loop within library code. I'm not sure it's due to e-l code (which seems quite fine) or a bug in nvidia code.

Commenting out this call will helped me running the game.

 

I hope this will help.

Marco

 

Well, even if it is a fault in our game, the driver should not have infinite loops..

 

Anyway, thanks for the info, we'll check it out!

Share this post


Link to post
Share on other sites

Well, even if it is a fault in our game, the driver should not have infinite loops..

 

Anyway, thanks for the info, we'll check it out!

 

I agree, but you know that those things can be easily triggered in C by wrong call sequences which can lead to small buffer overflows and so on. Anyway I'm still thinking it's more likely a driver issue even if other 3d programs seems to be unaffected (i.e. OpenArena based on Quake3 engine). I'll try some extra checks in those days. :D

Share this post


Link to post
Share on other sites

After a little debug i found that in framebuffer.c the testing code in check_fbo_formats for color_format 14 and 15 (GL_RGB16F_ARB and GL_RGBA16F_ARB) is triggering the problem. By skipping those modes everything starts.

Share this post


Link to post
Share on other sites

Can I make a client that disables those modes? I have the issue on my other PC but my solution was to uninstall the driver and install the old one.

Share this post


Link to post
Share on other sites

After a little debug i found that in framebuffer.c the testing code in check_fbo_formats for color_format 14 and 15 (GL_RGB16F_ARB and GL_RGBA16F_ARB) is triggering the problem. By skipping those modes everything starts.

 

Hi, my distribution is updating our nvidia drivers in the repo to 169.07 today. Can you provide a patch (file.diff) for your workaround so I can rebuild the el client?

 

Edit:.

Editing gl_init.c as lama suggested fixed the bug.

Edited by Fonzi

Share this post


Link to post
Share on other sites

Ok, the problem happens only with the beta drivers (released today). With the 'stable' driver, it works fine after the patch.

 

There are no new drivers for linux, as of today. I had same problem and "fixed" it by downgrading them.

 

Taking a look at it, so to see if I can spot anything else.

 

Linux IA32
Latest Version: 169.07
Latest Legacy GPU version (1.0-71xx series): 71.86.01
Latest Legacy GPU version (1.0-96xx series): 96.43.01
Archive

Share this post


Link to post
Share on other sites

Yes, the patch fixes it. I'm using 24 bit for display, not sure if that will break those that use plain 16.

 

I was trying to look deeper. There's already some nVidia hack code there. Maybe we're missing some more "default" values.

Share this post


Link to post
Share on other sites

still the problem since 3 weeks now ...

i was trying to edit gl_init.c, with no results...

maybe i was not perfect in doing it ...

can somebody post me how and where to edit exactly...

unless there is no patch available ....

greets

Share this post


Link to post
Share on other sites
Guest
You are commenting as a guest. If you have an account, please sign in.
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoticons maximum are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×