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

update nvidia graphic 169.07

Recommended Posts

i will try - its kind of a puzzle to me ...

 

Fonzi:

Editing gl_init.c as lama suggested fixed the bug.

 

lama:

.....testing code in check_fbo_formats for color_format 14 and 15 (GL_RGB16F_ARB and GL_RGBA16F_ARB) is triggering the problem.....

--------------------------------------------------------------------------------------------------------

gl_init.c:

if (have_extension(ext_framebuffer_object))

{

check_fbo_formats();

}

Share this post


Link to post
Share on other sites

How fucking incapable the pople from Nvidia have to be to:

1. Create such a bug that kills the whole machine.

2. Not even respond to my e-mail that detailed the probelm.

3. Release a windows Beta driver that fucks up the game even more...

 

I am not going to buy an Nvidia card next time.

Share this post


Link to post
Share on other sites

How fucking incapable the pople from Nvidia have to be to:

1. Create such a bug that kills the whole machine.

2. Not even respond to my e-mail that detailed the probelm.

3. Release a windows Beta driver that fucks up the game even more...

 

I am not going to buy an Nvidia card next time.

I, for one, am going to write to nvidia about this now, too. Maybe if they hear this from enough people they'll get sick of hearing it and actually do something about it...

 

For all the years I heard ATI sucks, get Nvidia, is the opposite now becoming true? (right after I finally invest in an Nvidia card... figures...) Seems to me that ever since AMD took over Nvidia, it's gone down hill. I'm not sure if there's cause and effect there, but I know I'm going to suggest this to them when I write. They've had ample time to publish their new drivers and fix the vista problems and other crazy issues with their new cards/drivers.

Share this post


Link to post
Share on other sites

Umm, actually AMD got ATI, and their drivers and support has been improving since then.

I had some issues with ATI, and they did respond to my e-mails.

Share this post


Link to post
Share on other sites

Umm, actually AMD got ATI, and their drivers and support has been improving since then.

I had some issues with ATI, and they did respond to my e-mails.

LOL, well now I feel like an ass. Maybe you could find some appropriate picture....

 

Either way, I did write to Nvidia about this (even though I don't use linux or this driver version...) I think this lack of support warrented a rant to them. They'll probably ignore it anways. :devlish:

Share this post


Link to post
Share on other sites

For me problem formats were:

GL_RGB32F_ARB

GL_RGBA32F_ARB

GL_RGB16F_ARB

GL_RGBA16F_ARB

 

Just comment them out ;)

 

Are you on Linux or Windows ?

 

If you're on Linux, could you post the output of "xdpyinfo" and "glxinfo" ?

 

And if you're on windows, can you tell me your resolution and color depth ?

 

I'm curious about why RGB?32F does not work for you.

 

Álvaro

Share this post


Link to post
Share on other sites

Does it work for you, with the latest Nvidia drivers?

Yes, but for the RGB?16 formats.

 

But the 32 ones work ok, at least they do not crash the client.

 

$ glxinfo | grep "version string"|tail -n 1
OpenGL version string: 2.1.2 NVIDIA 169.07

Share this post


Link to post
Share on other sites

Well, a bit of Framebuffer magic (i guess I need to advance to level 100 to handle it, though :P

 

After playing with Framebuffers and Renderbuffers I've found that only

GL_DEPTH_COMPONENT16 from depth format list was causing this hangup.

 

Finally I've changed PROBE_SIZE to 512 which seems to work with all formats,

depths and stencils correctly.

 

However, I've also made small test programs (glX and SDL_GL) which work

fine in this situation (buffer crate/attach/detach/cleanup code) even with SIZE 16.

So, nothing is clear yet.

 

P.S.

Also try to use GL_TEXTURE_RECTANGLE_ARB instead of GL_TEXTURE_2D, it might help

(for me it helps with GL_RGB32F_ARB and GL_RGBA32F_ARB (with PROBE_SIZE 16)).

Share this post


Link to post
Share on other sites

got burned by this one today when I took openSuse 10.3 to the 2.6.22 kernel. poking around the download.nvidia.com site, I found the following which may be useful if those reporting to nVidia have been doing so via a different method:

 

The driver RPMs hosted in this location are entirely built, maintained and supported by Novell/SUSE.

NVIDIA hosts them as a courtesy to Novell, however all problems and support requests related to

these RPMs should be reported to Novell via their bug tracking system:

http://bugzilla.novell.com

 

If the issue can be reproduced using the official NVIDIA driver package from this location:

http://www.nvidia.com/object/unix.html

 

please check and/or post to the linux discussion forum

( http://www.nvnews.net/vbulletin/forumdispl...&forumid=14 ). If you don't find an answer to

your question there, you can send email (in English) to linux-bugs@nvidia.com. When emailing

linux-bugs@nvidia.com, please attach an nvidia-bug-report.log, which is generated by running

"nvidia-bug-report.sh" as the root user.

 

Thank you.

Edited by nocnoc

Share this post


Link to post
Share on other sites

This is not a Linux bug, it is driver bug, and it happens on all the OSes.

I wrote them twice at developers@nvidia.com (their official e-mail for developers contact) and they didn't give a fuck about me, or about our users.

Share this post


Link to post
Share on other sites

If at all possible, avoid the 169.07 drivers. There is something majorly wrong with them. If at all possible, use the 100.14.19 drivers for 32 bit linux.

 

The lads and lasses at nvidia do not have a precompiled kernel module for the latest ubuntu kernel, which I believe has the largest proportion of linux users. I think nvidia may have taken their hands off the wheel when it comes to keeping the linux driver ball rolling. It's a shame, as they had developed a reasonable reputation for keeping the drivers working well.

 

S.

Share this post


Link to post
Share on other sites

Unfortunately, I'm using OpenSuse for the machine I've been running EL on. At least if it was Debian, I would have an easy path to backout of the nVidia update which came with the newest kernel.

 

I'm going to try bringing down the cvs copy and commenting out the modes sending nVidia into fits 'cause actually having to boot into windows for the first time in 3 months so I could play made me kinda twitchy. However, it is interesting to see the differences in the linux & win client.

Share this post


Link to post
Share on other sites

the only issue i had while running the 169.07 driver on my fedora 7 machine was that when changing between full screen and windowed modes, the screen would just go black. (game wouldn't crash though. but i understand that's a client bug which cvs version fixes, (although i seem to have other issues there) they have released a new driver though which is 169.09.

 

personally i've rarely had issues w/ NVidia's linux drivers, and when i did, they usually were fixed quickly. (i do remember one bad driver from a few years ago that was quite annoying)

 

system

athlon 2600+

epox 8rda3+ pro motherboard (nforce2 chipset)

geforce 5200 w/ 256 MB ram

 

fedora 7 w/ latest kernel from atrpms

169.09 official nvidia drivers

Share this post


Link to post
Share on other sites
Well, a bit of Framebuffer magic (i guess I need to advance to level 100 to handle it, though :P

 

After playing with Framebuffers and Renderbuffers I've found that only

GL_DEPTH_COMPONENT16 from depth format list was causing this hangup.

 

Finally I've changed PROBE_SIZE to 512 which seems to work with all formats,

depths and stencils correctly.

 

However, I've also made small test programs (glX and SDL_GL) which work

fine in this situation (buffer crate/attach/detach/cleanup code) even with SIZE 16.

So, nothing is clear yet.

 

P.S.

Also try to use GL_TEXTURE_RECTANGLE_ARB instead of GL_TEXTURE_2D, it might help

(for me it helps with GL_RGB32F_ARB and GL_RGBA32F_ARB (with PROBE_SIZE 16)).

 

PROBE_SIZE was the easy winner for me. The one thing I noticed in code was the "problem" modes were only defined if compiling for OSX.

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  

×