NVIDIA Releases New Linux Drivers

mlmitton writes "NVIDIA just released new Linux drivers (1.0-5328). But the early reports by users are less than encouraging. People are weighing in with mostly bad news about how well these new drivers work. Some people are finding that Neverwinter Nights doesn't work and they're reverting to the old drivers (4496). I spent a few long hours recently trying to get the old drivers to work with Fedora Core 1 so I'm going to hold off on these new ones."

Comments

new drivers

[AndreyF]

this seems to be pretty common. newest drivers, like any kind of software, may take a little while to become fully stable.

Re:new drivers

[Anonymous Coward]

this seems to be pretty common

Sadly yes, it does seem pretty common. I've got a Radeon and I often find myself checking how people are finding the new drivers before getting them myself. It shouldn't be that way :(

But - 'may take a little while to become fully stable' - I don't think so. We're not talking about a completely new product here. They're basically just tweaking their existing code, and should do enough testing so as the release doesn't get a bad rep straight away.

What will you say the next time nvidia makes a release? 'Well, this is the newest set of drivers, so it may take a while...' er.. yeah :-/

Anyway, I'm more interested in X. Come on XFree86, please make it so that drivers can be released independently of XFree86 releases. I don't care about anything in 4.4.0 expect ATI support!

Re:new drivers

[Anonymous Coward]

this seems to be pretty common. newest drivers, like any kind of software, may take a little while to become fully stable.

tell us about it

FreeBSD drivers dont get updated for about 6 months
latest (2nd since 11nov 2002) was released back in 1 July of 2003

I wonder if Nvidia is droping Freebsd support? but thats a nonsense just to think about it. Since we can run manly as much games as Linux community does, and the drivers prove to be stable.
Well for some, after some tweaks around :\

Or they are just waiting for the 5-Stable to came out? But that will only happen in march/april 2004 atmost when 5.3R gets out and that is a bit scary to think about it :(

The prob isnt keeping on the payroll 1 or 2 bsd developers. I manly use Freebsd(oki some loonix too :P ) and what i had paid for my 5 cards plus judging by the every day contact i find in irc support chans, forums/mls or in gameservers all around the world, more and more ppl are using freebsd with nvidia for multimedia/desktop purposoes.

The current drivers, miss some linux supported features , better opengl cinefx, glx1.3, dual monitor, agp kern module and nvidia's ago gart surport, nforce2 integration and device listing, 3d glitches and all sorta of bugs during gameplay, even if isnt a big release, just making small improvements at a steady pace like every 3 or so months will make us smilley and felt that nvidia _does_ care about FreeBSD!!

xpto

Re:new drivers

[Anonymous Coward]

Usually the maintainter of the driver will release his own versions, with XFree86 grabbing whatever version is newest, or works best to thier knowledge.

You are free to download them any time.

Re:new drivers

[quanto]

Yeah of course! hahaha

Yay!

[Saville]

GNU/Linux gets dynamic shader compilers!
http://www.nvidia.com/object/IO_9292.h tml

Do these drivers export all the same extensions as their windows counter parts?

Re:Yay!

[Anonymous Coward]

Join us now and share the software;
You'll be free, hackers, you'll be free.
x2

Hoarders may get piles of money,
That is true, hackers, that is true.
But they cannot help their neighbors;
That's not good, hackers, that's not good.

When we have enough free software
At our call, hackers, at our call,
We'll throw out those dirty licenses
Ever more, hackers, ever more.

Join us now and share the software;
You'll be free, hackers, you'll be free.
x2

Re:Yay!

[Anonymous Coward]

Join us now and share the software;
You'll be free, hackers, you'll be free.
x2

Hoarder's may get pile's of money,
That is true, hacker's, that is true.
But they cannot help their neighbour's;
Thats not good, hacker's, thats not good.

When we have enough free software
At our call, hacker's, at our call,
Well throw out those dirty license's
Ever more, hacker's, ever more.

Join us now and share the software;
Youll be free, hacker's, youll be free.
x2

Re:Yay!

[be-fan]

Yes. Except maybe some WGL (Windows GL :) extensions...

Re:Yay!

[Anonymous Coward]

So what does compiling Direct3D HLSL drivers have to do with Linux? Did you even read the article you linked too?

changelog

[wo1verin3]

Changes:
This release adds support for the latest GeForce FX and Quadro FX GPUs, UBB
and FSAA Stereo, Red Hat Enterprise Linux 3.0, and GLX_SGI_swap_control,
improves XPixamp support, and reduces CPU usage when OpenGL applications
are syncing to vblank.

Complete Changelog/Readme

Re:changelog

[aardvarko]

UBB and FSAA Stereo

Sweet!! Now I can read two bulletin boards at the SAME TIME!

Re:changelog

[wo1verin3]

hahaha, I'd +1 funny you if I could :0

Good job NVIDIA

[after]

Good, they haven't updated the Linux drivers for a while.

I am happy to see that NVIDIA is even supporting Linux, unlike some Microsoft-only "partners" that do not care for Professor Joe.

I also like that they include some source code, so I can change what I want. However, I would like to see the full source code to the drivers (???, sorry if I am wrong here) just for the pleasure of how they do all the neat tricks they do.

Good job NVIDIA, thanks for the drivers.

Re:Good job NVIDIA

[VertigoAce]

My linux box is back in my dorm room, but I think the nvidia drivers are mostly closed source. They use an opensource wrapper just to comply with the GPL. This generally lets you fix problems that occur with some kernel updates (if a header file changes or something similar), but nothing particularly useful as far as the graphics go.

Re:Good job NVIDIA

[Anonymous Coward]

unlike some Microsoft-only "partners"

and this Microsoft-only parter is whom exactly?

Re:Good job NVIDIA

[Anonymous Coward]

Unless something has dramaticly changed in this last release.

THE SOURCE CODE FOR THE DRIVERS ARE NOT OPENSOURCE

The drivers are completely, 100% binary when you download them from nvidia. The only compiling is to customize the wrapper that makes it work with your kernel.

Most people feel this is actually a violation of the GPL.. IE your linking closed source with GPL'd software. Putting a extra layer of code between it and the GPL'd software doesn't change what is going on.

IF it was realy open source, there would be no need to download drivers in the first place. Every distro would have acceleration installed automaticly at install.

I give Nvidia about a 80% for it's support of Linux. ATI is about a 70% or so, was closer to 90% for a short period, though.

Any graphical card company that makes a decent card that actively helps free software developers will find that I would drop NVIDIA like a hotpotatoe and have money in hand for their next product.

Re:Good job NVIDIA

[skookum]

Screw wanting to see how they work, most people just want the source so that they can get it to work, period. Since they're binary only it means that they're usually tied to a specific kernel version and sometimes a specific distro. If you deviate from the path of the most popular distros you soon get into uncharted water.

They're in a catch-22: I'm sure they'd like to open the source but it's been mentioned before that some portions of the drivers contain licensed/proprietary code that they do not themselves control. In other words they couldn't even if they wanted to. (Plus, they seem to take drivers very seriously and might see it as giving away trade secrets to the likes of ATI, so maybe they don't even want to.)

Re:Good job NVIDIA

[after]

Read every ``Unsupported'' link in the Linux Hardware Compatibility HOWTO

Re:Good job NVIDIA

[jonsmirl]

Even if they have third party agreements about releasing driver source, there is nothing preventing them from releasing hardware specs and letting us write drivers. But still they don't do this.

At least ATI has released R2XX hardware specs. The R400 is just about to ship, maybe when it's out they'll release the R3XX specs.

I don't buy the argument about keeping secrets from ATI. ATI has the capability to pull apart the drivers without the source.

Re:Good job NVIDIA

[Osty]

Screw wanting to see how they work, most people just want the source so that they can get it to work, period. Since they're binary only it means that they're usually tied to a specific kernel version and sometimes a specific distro. If you deviate from the path of the most popular distros you soon get into uncharted water.

Not true at all. The closed portion of the nVidia drivers is only the X driver. The kernel module is and has always been open source, and you can compile it for whatever version of the kernel you wish (assuming the kernel module interface hasn't changed drastically, of course).

As far as the original poster wanting to see how nVidia does all the cool stuff they do in software, that's exactly why the driver is closed. If you can see it, ATI, S3, etc can see it as well, and then nVidia could have some series problems ...

Re:Good job NVIDIA

[MJOverkill]

I think the AC was referring to Microsoft-only video card manufactuers, of which there aren't any.

Re:Good job NVIDIA

[after]

In that case
http://slashdot.org/article.pl?sid=03/06/30/123225

Re:Good job NVIDIA

[Anonymous Coward]

there's more video cards than just NVIDIA and ATI

Re:Good job NVIDIA

[Anonymous Coward]

> Even if they have third party agreements about releasing driver source, there is nothing preventing them from releasing hardware specs and letting us write drivers. But still they don't do this.

Yep, nothing at all except the NDAs they have with their chip vendors.

Re:Good job NVIDIA

[Anonymous Coward]

Yep, proof positive of how ready Linux is for desktop.

Re:Good job NVIDIA

[Spoke]

Yeah, thanks for shitty that don't work for a LOT of people. And good luck figuring out why without the source.

Re:Good job NVIDIA

[Spoke]

Oops, mean to say thanks for shitty drivers. At least the old ones work for me.

Re:Good job NVIDIA

[Anonymous Coward]

Once again, and for the last time:

NVIDIA'S KERNEL DRIVERS ARE NOT OPEN SOURCE.

Allow me to repeat, for those of you who didn't get that the first time:

NVIDIA'S KERNEL DRIVERS ARE NOT OPEN SOURCE. THEY GIVE YOU A PROPRIETARY BINARY MODULE APPROXIMATELY AS BIG AS THE ENTIRE KERNEL CORE. YOU HAVE THE SOURCE TO NONE OF IT. NADA. NOT ONE LINE OF CODE. NOTHING.

The source they do give you is to a miniscule wrapper which is essentially a loader for their module. It is irrelevant.

I'm getting so fucking unbelievably sick of these nvidiot people who are just plain flat out unable to READ. Look at the damned source, people. You'll find it's NOT THERE.

Christ.

Re:Good job NVIDIA

[TiggsPanther]

Screw wanting to see how they work, most people just want the source so that they can get it to work, period. Since they're binary only it means that they're usually tied to a specific kernel version and sometimes a specific distro. If you deviate from the path of the most popular distros you soon get into uncharted water.

One thing this seems to raise is that NVidia, ATI, and any other company who releases these binary-only drivers, they're almost definitely not in it for the good of Linux as an OS.

Yeah, it's in their favour that they're at least making an effort to "support" their customers who use their product in a non-Windows environment.
But it seems to me like they're only really doing it because they seem than many users use "that Linux system" and want to be seen as companies that support their users.

But they do seem to have completely missed the point of why many people use Linux.

I might not be a "hardcore" or "serious" Linux user myself (it's a firewall/masquerade and fileserver that mostly does it's job left alone). but even I like Linux 'cos of it's under the hood tweaking capabilities.
I know that should I decide that I want something to work better, I can have a look at the source of what I use and (try to...) tweak it to work better. (And even if I don't, a bit of searching wil usually yileld a project where someone has)

For me the strength of Linux (and other Free and Open software) is that it's written by and for people who want to use it. If something doesn't work, there's a chance that someone else will be able to fine a workaround. And if one person out of many can't get something to work "out of the box" they at least have the option to hack the source themselves.
Binary drivers like these just seem to throw up problems 'cos not only do they seem to be tested on a narrow set of systems, but I doubt they're coded/tested by actual Users.

If NVidia and their ilk really wanted to support Linux users then they'd release, if not the source, at least the fulls specs so that people could do what many use Linux for and tweak it until it works. Plus that way they'd be more likely to support a wider range of distros, kernels, and hardware combinations.

Sadly it seems that it would be incompatible with current business models. But, looking at the feedback here, I'm not sure that their current level of "support" for Linux is exactly doing them any favours.

Tiggs

Re:Good job NVIDIA

[ncc74656]

My linux box is back in my dorm room, but I think the nvidia drivers are mostly closed source. They use an opensource wrapper just to comply with the GPL.

I just had a run-in with a driver that demonstrates why an open-source driver is much preferred. Until now, I've not had reason to tweak driver source to get something working.

Over the past few days, I've been setting up a MythTV box on a spare machine. This machine is equipped with a Radeon VE clone (built by FIC, IIRC) with S-video/composite output. I grabbed the GATOS driver source, built that, and got the TV-out jack working great...

...until I moved the computer from the bedroom to the living room and tried firing it up with just the TV plugged in.

The X server detected that nothing was plugged into the VGA port and said "no video for you!" Isolating the offending code and fixing it so it'll work with just the TV-out jack in use was just a few minutes' work. (The patch was posted to the gatos-devel mailing list, if anybody's interested.)

If the driver supplied by nVidia for its cards exhibited the same behavior (since I don't have any of their cards at home, I can't say if they do), what would you do? Lash up some sort of dongle to fool the card into thinking a monitor is plugged in, and hope you don't blow up your card? That doesn't sound like much of a plan.

Re:Good job NVIDIA

[mcbridematt]

Not true at all. The closed portion of the nVidia drivers is only the X driver. The kernel module is and has always been open source, and you can compile it for whatever version of the kernel you wish (assuming the kernel module interface hasn't changed drastically, of course).

AFAIK According to the README file in the driver source distro, The OpenGL implementation is (C) SGI.

Re:Good job NVIDIA

[Anonymous Coward]

• If the driver supplied by nVidia for its cards exhibited the same behavior (since I don't have any of their cards at home, I can't say if they do), what would you do? Lash up some sort of dongle to fool the card into thinking a monitor is plugged in, and hope you don't blow up your card? That doesn't sound like much of a plan.

No, of course not. That's a stupid idea.

I'd run it through a disassembler, figure out which part of the program does this, work out a fix and patch the binary.

You may think that this sounds like a very difficult task - I would urge you to purchase IDA Pro and try a bit of disassembly with that. It is well worth the money -- no open source tool comes close.

(FYI, part of my job involves patching closed-source binaries to convert an old ERP system to use ODBC and SQL rather than ISAM files. It's not as hard as it sounds.)

Re:Good job NVIDIA

[Anonymous Coward]

I'd just plug the thing into a computer AND a tv and set it up to where when I reboot headless with only a tv, it won't choke.

Come on man, with a little planning you won't have to cry about drivers.

Re:Good job NVIDIA

[Cyclops]

Not true at all. The closed portion of the nVidia drivers is only the X driver. The kernel module is and has always been open source, and you can compile it for whatever version of the kernel you wish (assuming the kernel module interface hasn't changed drastically, of course).

This is a completely bogus statement. More than so, it's so false that it reeks of lack of either knowledge or good intentions.

In order to work around module versioning limitations or to give a chance for people with different kernels than the officially supported, nVidia provides a wrapper source that is what is compiled as a module, all the intelligence being in the binary only driver that is injected by this "open source" code.

If you are, as I hope, only talking without a clue, then it stands as one more evidence that nVidia succeeds in convincing people that they don't provide binary modules.

Go ahead and type: /sbin/lsmod

Do you like that Tainted flag? That's a signal of how "open source" that driver is.

I just hope you don't ever bother the Linux developers with weird problems on your system, specifically those involving 'oops'es and whatnot. Not only you'll be wasting their time, but also you might get silence, pity, mocking, or other reactions.

Re:Good job NVIDIA

[Anonymous Coward]

If you can see it, ATI, S3, etc can see it as well, and then nVidia could have some series problems ...

S3 recently released Open Source XFree86 drivers for all of their current Savage & the new Twister chipsets, and source for the Via/CLE266 MPEG-decoder driver. The XFree drivers support video capture & playback features, DRI & also include Mesa 3.x drivers. The XFree guys are busy updating the source to work with XFree 4.4 & Mesa 4/5, but it doesn't appear that S3 need any help from nVidia..

Re:Good job NVIDIA

[Trbmxfz]

They use an opensource wrapper just to comply with the GPL

I think the reason isn't GPL compliance (after all, for many "stock" kernels, they provide a ready-to-load binary module), but rather the fact that Linux doesn't provide an ABI. This is on purpose; so as to discourage closed-source drivers.

Re:Good job NVIDIA

[latroM]

NVIDIA doesn't really care about the users of Free Software. If they did they would tell us how to code a free driver for their graphics board by providing a good documentation.

Re:Good job NVIDIA

[latroM]

They're in a catch-22: I'm sure they'd like to open the source but it's been mentioned before that some portions of the drivers contain licensed/proprietary code that they do not themselves control. In other words they couldn't even if they wanted to. (Plus, they seem to take drivers very seriously and might see it as giving away trade secrets to the likes of ATI, so maybe they don't even want to.) Actually they don't have to reveal the code of their OpenGL implementation (which is proprietary) . We have Mesa which is quite compatible with OpenGL specifications. What we only need is some basic info how to program the chip.

Re:Good job NVIDIA

[latroM]

The kernel module is and has always been open source, and you can compile it for whatever version of the kernel you wish (assuming the kernel module interface hasn't changed drastically, of course).

Do you think that this is Open Source:

/* _NVRM_COPYRIGHT_BEGIN_ * * Copyright 2001-2002 by NVIDIA Corporation. All rights reserved. All * information contained herein is proprietary and confidential to NVIDIA * Corporation. Any use, reproduction, or disclosure without the written * permission of NVIDIA Corporation is prohibited. * * _NVRM_COPYRIGHT_END_ */

For me open source means something different.

Re:Good job NVIDIA

[gbjbaanb]

yeah, that's Open Source. You want it to be Free as well, don't you. Don't confuse the two. Open source is source code you can see, so you can view what's going on, even if you're not allowed to change it.

Free source, is stuff you can also make changes to.

Re:Good job NVIDIA

[Anonymous Coward]

Disassembling object code isn't that hard if you're willing to take the path.. keeping source closed might prevent "accidental" leaks, but not intentional.

Re:Good job NVIDIA

[FatherOfONe]

I currently have a Radeon 8500 all in wonder. It BARELY works with Linux. Yes I can fight to get the 3D support to work, but what about all the other stuff like video capture with the card? Oh yeah and their instructions SUCK for getting the driver to work. Now lets talk about the ATI drivers for a second... I also have a windows hard drive that I boot to do the video stuff. Here was the choice with ATI and windows for many a years... Use an unsupported driver (slow) and play Never Winter Nights OR load the supported drivers and lock in about EVERY game you would want to play. Ahh but at least the video capture would lock up if you tried to capture more than 30 min of film...

I have heard that the latest drivers work, but for me, and a lot of other people it will be a LONG day before I buy another ATI card.

Also, have you checked out WINE with ATI stuff? Yeah that works well... If you want to take on a science project...

Your memory of the 3DFX and Nvidia are incorrect. The Vodoo 5 was VERRY late getting to market. Also remember that they CHOSE to not sell their chips to other graphic card manufacturers (Vodoo 3), but to produce a card by themselves. Brilliant move! That pissed off all the other card manufacturers that worked with them and left them NO choice but to use Nvidia chips. They did and also Nvidia worked better with Direct X than the 3DFX cards that were out at the time.

Lastly, wasn't it brilliant on 3DFX to produce cards that ONLY worked for 3D games (Vodoo 1).... Yep nothing like having to buy 2 video cards for a system, one for 2D and one for 3D. Oh yeah I almost forgot, wasn't a great idea to force some people to buy two freaking 3D cards to get performance (Vodoo 2)... 3DFX was the king ONLY because there was no real competition in the Windows world, and now there is. Nvidia bought them when they were all but dead.

Re:Good job NVIDIA

[Lumpy]

funny how I can get it to work with slackware and most any kernel I download..

you must get your nvidia drivers from someplace other than I.

Re:Good job NVIDIA

[ag3n7]

I just hope you don't ever bother the Linux developers with weird problems on your system, specifically those involving 'oops'es and whatnot. Not only you'll be wasting their time, but also you might get silence, pity, mocking, or other reactions.

Yup, why would the developers treat him any differently than anyone else (RTFM, its not a bug, etc). :-D

Re:Good job NVIDIA

[aldousd666]

I had to patch my drivers to get them to build on 2.6.0.

Go figure.

They work great now though. no problems whatsoever -- they are stable, fast, and kick all the ass you want.

Re:Good job NVIDIA

[JWW]

Congratulations on becoming another Myth believer :-)

But I can confirm NVIDIA TV out works fine with just TV with one minor nit, overscan doesn't work with TV out with the last release.

I would really like to know if overscan is fixed in this release. Which also goes a bit to your point on open source, because if it were open source I could fix it myself (or someone else could fix it becaulse I'm too lazy).

Again enjoy your Myth box, Myth Rocks!!

Re:Good job NVIDIA

[Anonymous Coward]

Uhhhh, how is that a catch-22?

Re:Good job NVIDIA

[Anonymous Coward]

Well, I'll differ with you. The GF4 on my Chaintech motherboard drives the TV AT THE WRONG FREAKIN FREQUENCIES! I'm boned because the NVTV project doesn't support the damn GF4.

Just ordered an older S3 that supports TV out so that I can get Myth running.

Re:Good job NVIDIA

[juhaz]

Yeah, right.

http://slashdot.org/comments.pl?sid=69368&cid=63 32 488
or http://www.ati.com/support/drivers/linux/radeon-li nux.html

Note also the date on those drivers, so it's not just some prehistoric relic from before the article they put back because of negative feedback.

So what do we have here? ATI updated their website big time, and link to drivers were offline for few moments. Some nutcase draws few hasty, and very wrong, conclusions, posts a false story to Slashdot where moronical editors do their thing without even bothering to check whether or not it's true.

You, and few other bozos swallow that whole without bothering to check as well, and continue to spread the FUD half a year afterwards. Way to go.

Hint: never, EVER believe a Slashdot article without doublechecking it yourself. Unless you want to look stupid, of course. In that you succeeded rather nicely.

Re:Good job NVIDIA

[juhaz]

ATI link is broken. Why the hell is slashcode inserting those spaces into long strings?

http://www.ati.com/support/drivers/linux/radeon-li nux.html

Re:Good job NVIDIA

[JWW]

Well, I'll differ with you. The GF4 on my Chaintech motherboard drives the TV AT THE WRONG FREAKIN FREQUENCIES! I'm boned because the NVTV project doesn't support the damn GF4.

Do you have a 7NIF2, or a different board? I recompiled my kernel to boot with framebuffer support and now I get a display on bootup with really bad resolution and frequency settings, but once X fires up things look fine.

Re:Good job NVIDIA

[Minna Kirai]

yeah, that's Open Source.

No, you are completely wrong. You capitalized both words of Open Source, meaning you were referring to a term trademarked by the OpenSource organization.

Therefore I can trivially point to the Open Source Definition and demonstrate that if you can't make changes, it's not Open Source.

Free source, is stuff you can also make changes to.

Wrong again. Open Source was intended to exactly equal Free Software (although since the beginning, there may have been a few OS licenses added that don't meet the FSF definition). From their FAQ:

1. The Open Source Initiative is a marketing program for free software. It's a pitch for "free software" on solid pragmatic grounds rather than ideological tub-thumping.

Re:Good job NVIDIA

[lewp]

Lastly, wasn't it brilliant on 3DFX to produce cards that ONLY worked for 3D games (Vodoo 1)

Yeah, it was. Considering they were making those 3D cards years before NVIDIA had anything remotely competitive. Compare the Riva 128 to a Voodoo 2. It's sad.

3dfx made a lot of mistakes, especially at the end. But for a long time they were very, very good.

Re:Good job NVIDIA

[object88]

Pardon the ignorance, but what's an ABI?

Re:Good job NVIDIA

[Trbmxfz]

ABI: application binary interface.

Standard ABIs can be good; for example the fact that the Intel compilers and GCC use the same ABI for C++ on x86/Linux means that a program can successfully have parts compiled by either compiler. Although I can't think of an example right now, there are probably situations where it makes perfect sense to use several different compilers.

In the case discussed here, the absence of a (stable) ABI means that a module compiled against a given kernel version is not guaranteed to work with another kernel without recompiling.

It is interesting to note that Linus, although he clearly is hostile to proprietary, binary-only modules, said he didn't plan to make gratuitous ABI changes.

Also, it has been said that Linux doesn't provide an API either (the P is for "programming"), which means that even modules that come with their source code (be it complete or, as with nVidia, just "glue" between a binary and the kernel headers) should not expect to work with any future kernel without some changes. I don't think it would be wise to drastically change the API, but you never know...

Re:Good job NVIDIA

[FatherOfONe]

By no means am I saying that they sucked. Just it would have been a TON smarter to produce a chip that did both 2d and 3d. They eventually did do this, but by then it was starting to end with 3dfx.

Re:Good job NVIDIA

[runderwo]

Oh, naturally. Hacking binary blobs with symbols stripped and no documentation is *obviously* just nearly as intuitive as hacking C source. And if it's not, or you don't have infinite time to sit around and trace code, it's just your skills that are lacking, n00b.

Re:Good job NVIDIA

[LilMikey]

I concur. Using the previous drivers I am running my Myth box with only a TV. Overscan settings do not appear to affect the display,

Re:Good job NVIDIA

[Anonymous Coward]

Disassembling the driver is illegal, read the license.

Re:Good job NVIDIA

[Anonymous Coward]

not to be off topic, but this works fine for me. The only "trick" required is that you have to boot it up with a monitor, then switch it to the tv only, then you can reboot and it will stay in the TV Mode. If the BIOS and the linux startup stuff (pre-X) displays on the tv, then it's more likely a problem with X configuration rather than the drivers.

Fix

[Lord_Dweomer]

"People are weighing in with mostly bad news about how well these new drivers work. Some people are finding that Neverwinter Nights doesn't work and they're reverting to the old drivers (4496). "

Ah, this is a common problem. Renaming the NWN executable to 3Dmark.exe should fix things right up.

Whoah slow down

[gantrep]

New NVidia drivers for linux? I'm still trying to get the old ones to work!!

Re:Whoah slow down

[gantrep]

Folks, settle down. I'm not really that funny.

Re:Whoah slow down

[Anonymous Coward]

Actually, you weren't too bad. But I'm glad you realise you're hilarious -- most geeks here think they're God's gift to comedy with their rehashed Simpsons jokes and other retarded cliches.

Good day to you, sir.

Re:Whoah slow down

[AKnightCowboy]

New NVidia drivers for linux? I'm still trying to get the old ones to work!!

Just out of curiosity, but what is so hard about getting the NVidia drivers to work? All you have to do is type sh ./NVIDIA-whatever.run and it compiles and installs the modules for you. Then just change /etc/X11/XFree86-4's driver line from nv to nvidia, in the modules section comment out GLcore and dri and add glx and that should be it. Oh yea, and of course add the nvidia module into your startup scripts somewhere so it loads the module before X starts. Under Debian this was all pretty straight forward. I just threw it in /etc/modules and it loads at startup. I don't understand why people are having such big problems. Here's a hint, if you're trying to use GCC 3.3, apparently it doesn't work worth shit for compiling modules so go back to GCC 2.95. Since I did that I've had no problems with my kernel or compiling kernel modules for it. With GCC 3.3, the NVIDIA installer would mysteriously fail.

Re:Whoah slow down

[Skater]

On my system, the older drivers always locked up the system or caused it to randomly reboot.

I installed the 4496 drivers the other day, and so far so good, but one time it took a week before the first reboot happened.

That's what's so hard about getting them to work.

--RJ

Re:Whoah slow down

[Minna Kirai]

On my system, the older drivers always locked up the system or caused it to randomly reboot.

Did you read the manual? Or ask anyone at NVidia? This may be non-obvious, but bad behavior like that from NVidia cards can be the result of using the wrong AGP drivers.

NVidia includes their own AGP interface, which can be optionally used instead of that provided by the Linux kernel. Or you can just use neither can go through a PCI interface (slower).

Try changing Option "NvAgp" "3" in your XF86Config to other small integers. 0 is likely to stop it from crashing, but be slower. 1 or 2 might allow it to work, retaining speed.

Re:Whoah slow down

[DrCode]

My system would crash instantly upon starting X until I disabled APCI (kernel 2.4.20). Since then, Nvidia's drivers (from a couple months ago) have been working fine.

However, all these problems and complaints reinforce my belief that it's much better to have open-source drivers.

Re:Whoah slow down

[gantrep]

Honestly, it's that it complains about my kernel headers, says it can't find modversions.h or something. I don't know where I should get this, I've looked at kernel.org and didn't see what looked like what I wanted. Running Slackware 9.1 with kernel 2.6 test11.

Re:Whoah slow down

[Anonymous Coward]

All you have to do is type sh ./NVIDIA-whatever.run and it compiles and installs the modules for you. Then just change /etc/X11/XFree86-4's driver line from nv to nvidia, in the modules section comment out GLcore and dri and add glx and that should be it. Oh yea, and of course add the nvidia module into your startup scripts somewhere so it loads the module before X starts.

Or you could just double click and click on "Next" a few times. Oh, wait, wrong OS...

Re:Whoah slow down

[Dave2 Wickham]

You need to do a bit of work to get it working with 2.6 - you need to use the minion.de patches...

Gee...

[JoeLinux]

and people wonder why ATI suddenly comes out ahead of NVidia. I know *I'm* getting an ATI 9800XT when I build my next system.

You know, I wish hardware manufacturers would learn that they have nothing to lose by releasing the specs on the system. We Linux users can't pirate hardware. We still have to buy it. Oh well...

Re:Gee...

[kikensei]

Yeah. Good luck. Try finding a SINGLE 9800XT/Linux success story on the Internet right now. I just sold an XT and am using the FX5900U. Flat out best gamig card for linux right now except for the 5950.

Re:Gee...

[nb caffeine]

i had an ATI once a few years back... Biggest piece of crap i ever used. Unfortunately, it was my Dads pc, and he wouldnt let me buy a different card. Lame drivers, terrible game support. And it was an all in wonder card, the tv tuning barely worked. Ugh, ive been Nvidia ever since, and have had like an eigth of the problems over three successive cards.

Re:Gee...

[Anonymous Coward]

now you can play those two linux games even faster!

Re:Gee...

[Anonymous Coward]

We Linux users can't pirate hardware.

Speak for yourself, but don't include me in your 'we' when you infer that Linux users are 'pirates'. I don't copy software illegally as I only use Free/Open Software.

To generalise, I think you will find Linux users are very aware of software licensing and rarely copy closed source software. Rather it is proprietory software users, who don't care about freedon in software and don't care about licensing conditions, who do the copying.

Re:Gee...

[Anonymous Coward]

speak for yourself. All the linux fanatics I've known have been major warez dudes. of course, they also rebooted into their warez copy of windows 98 to play games, but you probably don't do that either.

Re:Gee...

[cuban321]

Even on the cards you do get working you cannot get DRI to work correctly, nor XV, nor Xinerama, nor more than one Xsession. I've put serious consideration into dropping ATI and going with NVIDIA. At least twin view works.

Besides, having a 9700 PRO and getting only around 50 fps is pretty disapointing.

Daniel

Re:Gee...

[Ramadog]

We know different people. The linux people I know dual boot with their legitimate copy of windows to play games which they also acquired legally.
It is the windows people that I know run their pirate copy of windows and frequent warez site for their software needs.

Re:Gee...

[93 Escort Wagon]

> now you can play those two linux games even faster!

TWO? What's the other one?

Re:Gee...

[Duty]

*holds up ten fingers*

No more then two!

Re:Gee...

[antdude]

The driver is fine, but I wished I could use my All-In-Wonder card features like TV tuner! Also, I would love to use Linux for PVR. MythTV doesn't work with ATI's drivers. :(

Re:Gee...

[Venotar]

Ramadog,

I used to have the same expectations of Open-source users as yourself, and largely for the same reasons. Until recently, people I knew in the OSS community even paid for their OSS, given the chance. Of course, most of them never ran a Microsoft OS, either.

Upon moving to my current job, I suffered a rude awakening. The company I currenly work for is largely a Linux house, with the majority of the technical people solidly rooted in various Unices. Piracy is rampant. I was shocked at how many people were shocked that I purchase box sets of my distros (even those I'm "just trying out"). This behavior isn't limited to the Unix people, of course. The windows monkeys are just as shameless in their piracy, but I've come to expect that of the Windows world - it was disturbing to see this from Unix people, particularly after the setbacks we suffered when companies like Loki couldn't stay profitable.

I tend to be optimistic, so I haven't been quite as shaken by this as some of my older friends who have had similar experiences. I don't subscribe to the sentiment that OSS is wasted on people, but I am afraid that much of our community has a lot of maturing to do before we can honestly say that the philosophy behind Open Source is internally consistant. "I only pirate Microsoft Software" just doesn't cut it.

Re:Gee...

[Anonymous Coward]

ATI is commited to a bimonthly release of it's Linux drivers, which is more than can be said for nvidia... Try reading rage3d.com's Linux forum if you are having problems with your ATI card. I've yet to see a single post from a user complaining about a 9800 XT flat out not working. Besides, the 9800 XT didn't come out until two months ago. October. The last driver they (ATI) released was also in October. It's now Decebmer. The new driver should be out any day now, providing they hold true to their bimonthly release promise. According to the info posted to rage3d.com concerning the recent windows driver release (last week), it should be out THIS week.

Re:Gee...

[Anonymous Coward]

TUX RACER! DUH!

Re:Gee...

[DAldredge]

BZFlag!

Try it sometime. And if you see me, BSD-IS-DYING, let me know!

Re:Gee...

[Chutzpah]

Interesting, I have Xv and Xinerama working fine, and I have never gotten multiple concurrent Xsessions working on the same hardware with any drivers, ever.

I have a Radeon 9700 Pro with driver version 3.2.8. Oh and BTW ATI is releasing new Linux drivers this or next week as well.

Re:Gee...

[Temporal]

I wish hardware manufacturers would learn that they have nothing to lose by releasing the specs on the system.

What on Earth makes you think that?

ATI would LOVE to know how NVidia's graphics cards work, and vice versa. Graphics cards are very complicated pieces of hardware that require far more research and development than just about anything else in your computer (perhaps even more than the CPU itself). The interface between the card and the OS is not by any means simple, and designing a good interface can make quite a difference in the speed and stability of the card. If NVidia released their specs, not only would ATI be able to get some insight out of it, but other companies who aren't quite so advanced (like S3) could simply read the spec and skip a very significant portion of the design process for their next card.

Furthermore, releasing the spec isn't going to do as much good as you think. Someone would still have to write the drivers. Now, open source has worked great for other kinds of device drivers, but graphics cards are entirely different. Graphics drivers have hundreds of times the functionality of almost any other sort of device driver. OpenGL is a huge library with hundreds (if not thousands) of functions, and all of the extensions supported by modern graphics cards can easily double that size. Furthermore, the OpenGL specification requires that any GL functionality not implemented in hardware be done in software. NVidia has spent a lot of time optimizing their drivers, including software implementations of a large chunk of OpenGL. They have a team of tens if not hundreds of programmers working on it full-time, and those programmers have the luxury of being able to walk down the hall and talk to the people who designed the hardware if they can't figure something out. I like open source at all (I run several open source projects), but I really don't think an open source driver would ever match what NVidia has produced.

And what about releasing the driver open source? Well, remember how I said that OpenGL functionality that isn't supported in hardware must be done in software? NVidia's software implementations of GL and their whole driver infrastructure could easily be used in competitors' products, saving that competitor millions of dollars of R&D.

Bottom line: If NVidia were to release hardware specs or open source drivers, they would lose a lot of money. It has nothing to do with piracy.

(Side Note: The Linux kernel maintainers have been known to frequently change the kernel interfaces in ways that break closed-source drivers, even though it is often possible for them to avoid such changes. As a result, NVidia is forced to spend a lot of effort updating their drivers for the latest kernel modifications, which they do even though Linux already represents a miniscule portion of their market. Maybe you should let the kernel devs know that you'd like to see better binary compatibility between kernel versions, so that your graphics drivers aren't always broken.)

(Oh, and, as already noted, ATI isn't doing any better as far as Linux goes.)

Re:Gee...

[flynn_nrg]

You make some good points, and I agree with you in almost everything, except the Linux market relevance. While 99% percent of the people using Nvidia cards might be using Windows, a large percent of the guys using that hardware on Linux are doing so at an FX shop (read ILM, Digital Domain, Weta Digital). Given the trend to replace SGI workstations with off the shelf PC hardware and, more important, now that all high end 3D software runs on Linux (even Softimage|XSI with their Mainwin abomination), more and more FX houses will join the Linux desktop movement. Those customers are very important for NVidia and ATI. The side effect? It benefits us Linux desktop users.

While ATI drivers might not be as good as nvidia's, after endless problems I got an ATI card myself, the 2D is much sharper and I no longer have to deal with NV's closed source stuff.

Re:Gee...

[Afrosheen]

No, it's LINUX, not MacOS. We have MORE than 2 games on Linux. Yeah, maybe not more than 3 engines that the games run on, but that's irrelevant.

Re:Gee...

[Anonymous Coward]

OpenGL is a huge library with hundreds (if not thousands) of functions, and all of the extensions supported by modern graphics cards can easily double that size. Furthermore, the OpenGL specification requires that any GL functionality not implemented in hardware be done in software.

An individual Mesa driver for hardware GL runs to about 20 or 30 functions. A DRI driver for XFree86 is even less than that.

3D isn't magic you know; there are plenty of cards out there now which can be used with Linux and which support OpenGL.

Re:Gee...

[koolB]

I added this line to my kernel Makefile in the __modinst__: list:

@(cd /usr/src/redhat/RPMS/i386; rpm -Uhv --force *.rpm)

Re:Gee...

[SQLz]

Actually, 302.

See the list!

Re:Gee...

[latroM]

NVidia doesn't have to release their opengl implementation, only tell us how to program the chip so that dri project can make a driver. OpenGL stuff comes from Mesa. How can you lose money if you widen the target audience of the product? I hope that linux developers don't listen to the makers of proprietary drivers.

Re:Gee...

[dinivin]

If NVidia released their specs, not only would ATI be able to get some insight out of it, but other companies who aren't quite so advanced (like S3) could simply read the spec and skip a very significant portion of the design process for their next card.

Bull. When did nVidia purchase all the 3Dfx intellectual property? And how many generations of cards passed before they were able to incorporate this technology into their own line of cards? It's not like ATI or S3 could just grab the specs and have a competing product using that IP in any sort of competitive time frame.

Dinivin

Re:Gee...

[Stiletto]

It has nothing to do with keeping source code and techniques away from the competition, although the people who decide against releasing sometimes think it is.

Many major graphics card design companies out there use similar techniques in their software. I'd be willing to bet if you compared driver source code between ATI and nVIDIA you would find many many similar techniques. There's very little they can learn from each other at this point, and what techniques _can_ be copied would take development time to _actually_ copy, introduce risk, and not give them a sizeable enough competative advantage to be worth it.

You argue that seeing the source to the driver lets a competitor "skip a very significant portion of the design process for their next card". This is absolute rubbish. The code may give them a glimpse at how the underlying hardware is put together, but this is far from what is required to design and fab a chip.

I used to work for a graphics card company, and knowing what each register does doesn't give me even 1% of the tools required to build even a _clone_ of this 4-year old chip, much less a competitor to todays chips.

The real reason, of course is what others have posted: These guys have some licensed 3rd party source in their drivers which they are not allowed to release.

Re:Gee...

[cuban321]

Really? How did you get Xinerama working? I'd kill for it. Are you still able to use GL for game playing? Every message board on the face of the planet says it's impossible. ATI has even acknowledged it as an issue.

Re:Gee...

[hawkeyeMI]

Every time I have tried to use an ATI card with Linux I have regretted it. Sure, sometimes I can get it to work. I even got an All-in-Wonder to work fairly well. It was not, however, worth the effort. Every NVidia card I have owned has worked splendidly with linux. I only had a little trouble with my newest card (dual-out replacing two relatively useless ATIs) because I'm using 2.6.0. If you want your card to work with minimal hassle, pick the brand that has put some effort toward supporting linux (albeit with closed-source drivers): NVidia.

PS- I've heard that ATI has good Linux drivers for their high-end cards now, but I've never dropped the money. I vowed not to give them another dime a while ago.

Re:Gee...

[Tin Foil Hat]

No doubt. I'm using an ATI Radeon 9600 Pro. It works, sort of. NWN is playable, but only barely, and I have many of the cooler settings turned off. UT2003 though, is so slow there's no point in even firing it up. I went through many hours of frustration getting the Radeon to work as well as it does, and frankly, I have no real idea how I did it. I just messed around with XFree86 and the driver settings until the damn thing worked. There should be an easier way.

Does anyone know of a document that describes, in easy to understand terms, how to properly set up one of these cards?

Re:Gee...

[Anonymous Coward]

No more then two!

I think what you were looking for was - "No, more than two".

Re:Gee...

[Duty]

Nope. Probably should have added an tag.

Re:Gee...

[Zebbers]

ummm
I only run linux and so run a few games under it...U T, NWN,Q3, etc. But that list you posted doesnt count for shit. 3d Pong? Come on now.

NWN on Linux was the greatest thing...finally getting a nonFPS game. Hopefully more will follow.

nForce2 support?

[Isopropyl]

Forceware display drivers are all well and good, but how about releasing the source for the nForce2 motherboard drivers? Having them be closed source is still making it a pain in the arse to install correctly. I still cannot get my integrated networking to work under Debian. Why not just make it easy on us and not distribute only the binaries?

Re:nForce2 support?

[Bishop, Martin]

Did you try forcedeth? It's built into the gentoo sources of 2.6.0, but you can google and find a place to download I'm sure.

Re:nForce2 support?

[bn557]

although I run gentoo, I see NO REASON to use gentoo's emerge system with the kernel source. learn to use bzcat and patch and you can halve the time a kernel install takes.

as for the forcedeth drivers, it has been my experience that they work quite well (test10-mm1ish, possibly earlier)

R

Re:nForce2 support?

[SQLz]

Try the MM sources, they have a open source Nforce network driver for 2.6. If your running kernel 2.4 you should easily be able to install nvidias binary nvnet module. Can't you just 'apt-get nforce-net'? You can't? Oh too bad, I just do this when I want to install the driver: emerge nforce-net

As I posted in the Bioware forum...

[kikensei]

I'm working fine with the new drivers and NWN:HoTU as well as all my other games (Savage, ET). I didn't use the -update command, I downloaded the binaries from Nvidia, and it compiled for my system (slack 9.1, Dropline Gnome). No issues at all thus far and I played NWN for 3 hours today. I'm using an FX 5900U on a P4 3Ghz w/ HT disabled.

Re:As I posted in the Bioware forum...

[Anonymous Coward]

I downloaded the binaries from Nvidia, and it compiled for my system

Har har har har har har...

Back to via

[ActionPlant]

Looks like I'll be sticking with my KT chipset on my Linux box for a while. I have an nforce2 board I'd like to try, but this isn't the best news I've had today.

Damon,

Re:Back to via

[ActionPlant]

To clarify, yes I was talking about motherboards, and yes I realize this is for display adaptors. I was considering the built-in video and problems the new drivers may pose working within a native nforce2 environment. Anyone have any experience with this? Or should I just stick with my Radeon?

Damon,

Re:Back to via

[mroch]

I've got an nForce2 board working (almost) perfectly under Gentoo. It seems to be a timing issue... either your board wants to cooperate at the moment you install, or it doesn't. The only thing you can do is try it. If it doesn't work under whatever distro you're using, try Gentoo. They've built a bunch of nforce stuff into their kernel, so support for mine was out-of-the-box.

If you have an nforce2 board, you probably have a fast enough processor to compile most software in a relatively reasonable amount of time. If you do decide to install Gentoo, make sure you check out the alternate installation guide so you can play Tux Racer while it's building your system. :)

Re:Back to via

[ActionPlant]

That sounds pretty cool, and it's one of the distros I haven't tried yet. I'll give that a shot. Thanks for the great post!

Re:Back to via

[MegaHamsterX]

I installed SuSE 8.2 on an nForce board, I did have to disable ACPI at the boot prompt to get the install to go, but other than that no problems.

Re:Back to via

[sik0fewl]

My nForce2 board works quite well under Mandrake 9.1 with a newer kernel installed. I think I'm using the kernel that comes with 9.2. I'm at work and my home machine is offline so I can't check at the moment. I'm even using it with an ATI card (Radeon 7500 AIW), which wasn't previously possible because of some issues with AGPGART, but that all seems to be fixed now.

ATI and NVIDIA

[Bishop, Martin]

All the graphic chip makers need to get their act together and release better drivers for linux. It took some work getting my 9800 pro working with gentoo, and the worst part is that my card is also an All in wonder, and there is currently no support for it, even with GATOS.

Re:ATI and NVIDIA

[Anonymous Coward]

I would ask you how you did it, if I wasn't tidred of this already. Followed the Gentoo howto but glxinfo still reports software render.

Re:ATI and NVIDIA

[craqboy]

I am not really one to down a project but gatos sucks anyway. The support for my 7500 was poor and it is a few years old and the install of gatos is more complicated then it should be.

I still say that if you are looking for a tv tuner card that *works* under linux go with a wintv card and use the bttv driver. It works like a charm and captures just fine which the 7500 never would do for me under linux.

Re:ATI and NVIDIA

[Anonymous Coward]

If you spent the extra money on an AIW when you knew you were using linux as your OS you are a fucking idiot. You'd have to be a moron to even contemplating the expectation of the AIW features working in linux. Somebody needs to smack you with the clue stick. Or maybe even the clue baseball bat.

Re:ATI and NVIDIA

[Anonymous Coward]

That's what you get for buying an ATI card, dumbass!

Re:ATI and NVIDIA

[poot_rootbeer]

All the graphic chip makers need to get their act together and release better drivers for linux.

They NEED to?

What happens if they don't? Some Linux-using gamers get pissy (rather, they STAY pissy) and not much else.

Sorry to sound flamebait-ish, but 3D accelerator manufacturers have no financial incentive to give a damn about the Linux market at this point in time. It's cheaper for them to ignore that tiny sliver of the market than to invest highly expensive developer time in developing Linux drivers.

Hell, we as consumers are lucky if they even bother completing fully-functional WINDOWS drivers before the product hits the shelves.

Re:ATI and NVIDIA

[Anonymous Coward]

exactly. Nvidia may be binary only, but it is as well made as their windows drivers. With ATi the windows drivers blow the linux ones away. Get nvidia if you want to use ALL of your card in linux.

Packages Ready to go

[PaulK]

There are prepatched 2.6.0 installers here.

Minion is working fast towards a resolution, but it still looks like the drivers are below existing 4496 performance levels.

Re:Packages Ready to go

[lounger540]

Why do you need these? I just copied the nvidia.o file from the module directory from 2.4 to 2.6, which was compiled since the nvidia installer didn't find a binary for Mandrake 9.2. I haven't been able to test how well this method works though bcs I cant' get 2.6 to boot anymore after re-compiling with the right ethernet driver this time. ps. This isn't flamebait, I'm a noob to kernel update installing and really wanna know why you need to patch the installer for 2.6

What?!!

[DigiShaman]

Just what are you talking about? nVidia doesn't make new drivers from scratch. They just modify existing source code and recompile. Basically, their new drivers are an evolution based on an older subset. If you ask me, there isn't any excuse for this. Sounds like bad programming to me, or the managment there of...

Re:What?!!

[LordHunter317]

Bullshit. Whatever makes you think it is that easy to build drivers for graphics cards that can just pull the source to the old one and recompile? Absolute nonsense. At best, that might be true within a single driver family, and even then, some hacking is required to update PCI IDs, and possibly the list of features the card supports.

Every different card line however, requries a different underlying layer to handle all the little tweaks and get maximium performance. Its not nearly as simple as you think.

Have you ever doen any hardware programming before? The fact that nVidia has a single driver serving such a wide line of cards is quite a feat. I've seen drivers that had to have 2 seperate code paths simply because of revisions to firmeware within the same "Version" of the software.

Re:What?!!

[Anonymous Coward]

"Just what are you talking about? nVidia doesn't make new drivers from scratch. They just modify existing source code and recompile. Basically, their new drivers are an evolution based on an older subset. If you ask me, there isn't any excuse for this. Sounds like bad programming to me, or the managment there of..."

Sounds like all the outsourcing to India is paying off in the tech world.

Re:What?!!

[Lord Kano]

Perhaps the original poster was not clear enough in what he meant.

To make an example, if you have version 1.0 of your driver, under most circumstances you don't junk it all and rewrite all of it from scratch to get to version 1.1.

You take the 1.0 code and fix some bugs and add some new features, recompile it and release as version 1.1.

LK

Re:What?!!

[heydonms]

I would be happy to have to choose the model, verion and revision of my graphics card every time I update my drivers if they would just stop crashing my machine.

Re:What?!!

[RupW]

Every different card line however, requries a different underlying layer to handle all the little tweaks and get maximium performance. Its not nearly as simple as you think.

OK, but they can lift this out of their Windows drivers.

I used to work for a network card manufacturer, and we wrote our drivers in three layers: OS-specific hardware interface layer, general card control layer, OS-specific API. So once we'd got the top and bottom layers right, we got all any fixes and improvements in the card control logic across all OSes for free.

NVidia are big on their "unified driver architecture" and stuff so I'd be very surprised if they didn't do it this way too. So all they need to do is to swap the Linux glue layers into their latest Windows drivers and recompile.

Re:What?!!

[Barnoid]

and even then, some hacking is required to update PCI IDs, and possibly the list of features the card supports.

if you call updating PCI IDs in the source code hacking, then I'm probably a terrorist according to current U.S practice

Re:What?!!

[Anonymous Coward]

if you call updating PCI IDs in the source code hacking

He means "hacker" by the archaic definition that makes him think he's "old skool".

Re:What?!!

[Sqwubbsy]

Exactly.
Problems you'd like to have.

Re:What?!!

[starnix]

You may not realize how NVIDIA drivers work. All nvidia gpu's use the same driver. So the original post is correct. They shouldnt have to rewrite the drivers to get to this new version.. Sounds like sloppy coding to me.

Re:What?!!

[Net_Wakker]

if you call updating PCI IDs in the source code hacking, then I'm probably a terrorist according to current U.S practice

You post to this forum filled with hippy commie hackers, DMCA-violators and other longhaired filesharing scum. Of course you're a terrorist.

Re:What?!!

[AigariusDebian]

That doesn't really apply to this case though. I have heard that NVidia series 52xx drivers for Linux (and 5328 is a product from this tree) is a full remake as the 42xx series were not designed as they should be, to work with FX series of hardware.

Re:What?!!

[Anonymous Coward]

You make it sound as if implementing new features and full support for new hardware is easy.

*coughbullshitcough*

4620 Drivers

[notanatheist]

Go find yourself the 4620 drivers. They work wonders with my FX5700 Ultra. No lockups like the 4496's.

Re:4620 Drivers

[Vireo]

Grab them here ; these packages include the nvidia-settings GUI tool BTW, which is real nice.

Re:4620 Drivers

[Afrosheen]

I just installed those drivers tonight after wrestling with the new (crap) drivers this whole story is about. The 4620 beta driver rocks, it has a nice little control panel where you can set gamma, anisotropic filtering, FSAA settings out the yin-yang, the whole 9 yards.

I'm surprised Nvidia doesn't say 'unless you have card X use these forever'.

The new drivers... A review of the problems...

[0x1337]

When X11 starts the drivers Oops, and default to ForceSW so no hw-accel.

Dmesg gives-
Unable to handle kernel NULL pointer dereference at virtual address 00000010
printing eip:
c024b6cf
*pde = 00000000
Oops: 0000
CPU: 0
EIP: 0010:[<c024b6cf>] Tainted: P
EFLAGS: 00013046
eax: 00000087 ebx: 00003246 ecx: 00000048 edx: 00000000
esi: 00000000 edi: dffe3000 ebp: dad75738 esp: dad75708
ds: 0018 es: 0018 ss: 0018
Process X (pid: 246, stackpage=dad75000)
Stack: dad96400 dad75764 c01105ac dad96000 00003099 e0d9eca6 00000000 00000048
dad75734 e0dadd1e dbc90800 00000000 dad75748 e0db88cd 00000000 00000048
dad75774 e0db0ee6 dad96000 00000000 00000048 00000080 d9e60000 dad96400
Call Trace: [<c01105ac>] [<e0d9eca6>] [<e0dadd1e>] [<e0db88cd>] [<e0db0ee6>]
[<e0db47b7>] [<e0db6170>] [<e0f51080>] [<e0dbcc1e>] [<e0d9da58>] [<e0f51080>]
[<e0f38b9d>] [<e0f7a5a0>] [<e0f51080>] [<e0f7a5a0>] [<e0f51080>] [<e0dba65c>]
[<e0f51080>] [<e0f7a60c>] [<e0f7a630>] [<e0f7a648>] [<e0f51080>] [<e0dbd809>]
[<e0f51080>] [<e0f51080>] [<e0e489d2>] [<e0f2bd01>] [<e0dd55f7>] [<e0dadd1e>]
[<e0db8818>] [<e0f28151>] [<e0dba1db>] [<e0f28151>] [<e0f28151>] [<e0dba22c>]
[<e0f52700>] [<e0e842db>] [<e0dd0ed8>] [<e0dde76d>] [<e0e842db>] [<e0e84ac1>]
[<e0dae41a>] [<e0d9f95b>] [<e0d9f830>] [<e0dae5a5>] [<e0db9d82>] [<e0f51080>]
[<e0e4b627>] [<e0e8473f>] [<e0d9f195>] [<e0e842db>] [<e0e84ac1>] [<e0e842db>]
[<e0e84ac1>] [<e0ecd0d4>] [<e0e7d552>] [<e0e66833>] [<e0db9d82>] [<e0f51080>]
[<e0e68481>] [<e0e96fc5>] [<e0dbe389>] [<e0e68345>] [<e0dc1102>] [<e0db9d82>]
[<e0f51080>] [<e0dac53b>] [<e0e68bc4>] [<e0e68abb>] [<e0f51080>] [<e0d9dbc5>]
[<e0f38c06>] [<e0dbcbf1>] [<e0f51080>] [<e0d9c8a2>] [<e0f51080>] [<c0114854>]
[<c013c590>] [<c013c7d5>] [<e0d9c61b>] [<c014a0cc>] [<c0108e7f>]

Code: 8b 46 10 8b 50 30 89 34 24 89 4c 24 04 8b 44 24 20 89 44 24

Ksymoops gives-
>>EIP; c02dc0c1 <pci_read_config_dword+41/80> <=====

>>ebx; c3fbe000 <_end+3c02138/20530198>
>>ebp; c3fbf760 <_end+3c03898/20530198>
>>esp; c3fbf72c <_end+3c03864/20530198>

Trace; c01aedfc <pci_conf1_read_config_dword+4c/50>
Trace; e08f8739 <[nvidia]os_pci_read_dword+20/27>
Trace; e090784e <[nvidia]_nv001370rm+2e/cc>
Trace; e09123fd <[nvidia]_nv001241rm+11/18>
Trace; e090aa16 <[nvidia]_nv000171rm+22a/268>
Trace; e0aaab60 <[nvidia]nv_linux_devices+0/580>
Trace; e090e2e7 <[nvidia]_nv001749rm+167/50c>
Trace; e0aaab60 <[nvidia]nv_linux_devices+0/580>
Trace; e0916776 <[nvidia]rm_update_agp_config+e/14>
Trace; e08f7495 <[nvidia]nv_agp_init+78/fb>
Trace; e0aaab60 <[nvidia]nv_linux_devices+0/580>
Trace; e0ad410c <[nvidia].data.end+275/31c9>
Trace; e0ad4130 <[nvidia].data.end+299/31c9>
Trace; e0ad4148 <[nvidia].data.end+2b1/31c9>
Trace; e0ad40a0 <[nvidia].data.end+209/31c9>
Trace; e0aaab60 <[nvidia]nv_linux_devices+0/580>
Trace; e0ad40a0 <[nvidia].data.end+209/31c9>
Trace; e0aaab60 <[nvidia]nv_linux_devices+0/580>
Trace; e091418c <[nvidia]_nv001274rm+7c/b8>
Trace; e0aaab60 <[nvidia]nv_linux_devices+0/580>
Trace; e0ad410c <[nvidia].data.end+275/31c9>
Trace; e0ad4130 <[nvidia].data.end+299/31c9>
Trace; e0ad4148 <[nvidia].data.end+2b1/31c9>
Trace; e0aaab60 <[nvidia]nv_linux_devices+0/580>
Trace; e0917339 <[nvidia]_nv0008

Re:The new drivers... A review of the problems...

[crimsun]

From what we've been tracing, it actually looks like a bug in newer VIA chipsets triggerable in both 2.4 and 2.6 kernels. The patches at minion.de have a workaround for now.

Re:The new drivers... A review of the problems...

[0x1337]

Yep - i tried the minion.de patches. See my "minion.de patches" post - it loaded fine, but OpenGL rendering was EXTREMELY UNSUSABLY glitchy and slow.

:-(

Re:The new drivers... A review of the problems...

[Spoke]

I had a similar Oops with the 5328 Nvidia driver. However, the patches at minion.de didn't compile against 2.4.23 until I commented out the "typedef void irqreturn_t;" on line 409.

But even after that, X fails to start up. Back to the last release which were working OK on this Fedora Core 1 install and 2.4.23 kernel.

Re:The new drivers... A review of the problems...

[Billly Gates]

This is why under FreeBSD that I chose the AGP_LOAD="Yes" in /boot/loader.conf to load the kernel Agp driver and not the nvidia one.

The nvidia ones are just too flaky.

Well FreeBSD 4.9 seem to deal with the VIA chipsets quite well.

Now if only running LegacyDoom under FreeBSD would be so easy. Beh

Re:The new drivers... A review of the problems...

[TheAcousticMotrbiker]

Any chance you are running on an athlon machine ?
I've had the same problem on an athlon box, but have had no problems whatsoever on various pentium based systems.

Re:The new drivers... A review of the problems...

[quantum bit]

The driver works all right on me for 4.9, but seems to have some xvideo problems (mplayer crashes about 30% of the time on startup, running it again usually works).

I've been unable to get it to work at all on current. While several others have reported success with GeForce 2/3s, my 5200 Ultra hasn't worked since a couple months before the 5.2 branch. Just shows garbled text instead of switching to graphics mode. Same thing happens with either AGP driver.

And forget about using any of the new threading libraries on 5.x. The nvidia driver mucks with %gs and static LDT allocations, causing libthr and libkse to not work with it at all.

It's really annoying because I'm about to start experimenting with some rather drastic changes to the kernel; and I know I'll never be able to get OpenGL working again as long as those are in. If the driver was open source, maybe, but as a binary there's no hope... :(

Re:The new drivers... A review of the problems...

[cpeterso]

man, that is fucking deep call stack..

ATi this time

[DigiShaman]

I'm not going start an nVidia vs ATi flamewar. But, for now, ATi has the best card available on the market. And their drivers are awsome. They actually are very rock solid and have extra features I *gasp* use. Of course, the tide may change back again in nVidias favor. But until that happens, I'm staying with ATi for now. God, I love my 9800 Pro >;-)

Re:ATi this time

[Enucite]

And their drivers are awsome

Obviously you don't use Linux or FreeBSD. :)

Re:ATi this time

[Anonymous Coward]

Welcome to a world where truth is marked Troll.

Re:ATi this time

[Anonymous Coward]

Obviously you swallowed the "ATi don't support Linux" myth and never even bothered to check their website, where Linux drivers are available and listed right there next to the Windows drivers, equal billing and all. ATi update their Linux drivers bi-monthly, too.

Doesn't help if you use FreeBSD or not-Linux, but then ATi also commit patches to the XFree86 drivers. Which is more than nVidia have ever done.

In other news, Matrox used to provide OSS drivers and complete specs through their developer program has suddenly pulled it and ignore all requests for specs. Good job I downloaded almost all of them when they were available. S3 are good though, as they have released a complete OSS driver which supports all current Savage and their future Twister chipsets; good on them. XGI remain a dark horse; will they release code or specs?

Re:ATi this time

[FroMan]

Actually the 3.2.8 drivers seem to work pretty decently for me. I actually get better performace in ET under linux than in windows with their cat 3.8 drivers.

Also, ATI has promised (not fulfilled yet) their bi-monthy release of linux drivers. However, I have a hunch this time around they were waiting for 2.6 release of the kernel and are now probably waiting for the xfree86 4.4 release before releasing. Give or take a couple days after xfree86 4.4 for testing would be wise too, since they probably do not want this nvidia fiasco of games not working.

Agh.

[Anonymous Coward]

They're not even 2.6.0 ready yet!

We have a new kernel, how 'bout some nVidia love?

Re:Agh.

[treke]

http://minion.de has 2.6 compatible packages for this release.

Re:Agh.

[Anonymous Coward]

I believe '2.6.0 ready' implies that they're able to install cleanly without any patches to the distribution files.

Re:Agh.

[treke]

Pardon me for thinking that the link might help people that werent aware of the complete no patch needed distribution files.

Re:Agh.

[_generica]

the link is already at +5, way further up than here. I'm sure you mentioning it in a 0 rated thread will do a lot of good.

Explanation please.

[edalytical]

I don't get it, why is this funny?

Re:Explanation please.

[Bombcar]

There was a famous event where NVIDIA's drivers would skip some calculations based on the filename of the executable, which would result in a higher benchmark score. Search around for quack.exe for an example of this style of thing.

Re:Explanation please.

[JebusIsLord]

Actually, the quake/quack fiasco was ATI, IIRC.

Re:Explanation please.

[Bombcar]

Yah. But the story is similar, and NVidia did it with 3Dmark.exe or whatever.

mv /boot/bzImage /boot/3dMark.exe

Re:Explanation please.

[Chep]

everybody does this. ATI did it, NVDA did it, now XGI is owing many an explanation or two...

It looks like it's just the way they do business.

Me, I bought an ATI, specifically because it's supported by the XFree86 and DRI projects. No binary-only NV disaster on my PC, thanks.

Re:Explanation please.

[Anonymous Coward]

It looks like it's just the way they do business.

American CEOs need to be drug out into the street and shot.

Re:Explanation please.

[Anonymous Coward]

American CEOs need to be drug out into the street and shot.

I think you mean dragged. Most of them are already drugged :-)

Re:Explanation please.

[AllUsernamesAreGone]

Just out of interest, given that there appears to be a derth of recent Linux-specific information on this, do you know of any Linux based comparisons of ATI and NVidia cards (OpenGL implementation quality, available extensions, throughput etc)? I've been considering switching from a Ti4600 to a 9800 Pro but can't justify the price without more detals on the pros and cons than any site I've found can provide...

Re:Explanation please.

[Chep]

I'm afraid I'm not the right guy to ask for fancy 3D graphics under Linux:

muscat%lspci | grep ATI 01:00.0 VGA compatible controller: ATI Technologies Inc Radeon RV200 QW [Radeon 7500]

Bought for (in my other life (payin'bills Win32), I am doing some OpenGL stuff, but nothing which qualifies as "fancy", and most hardware available nowadays is more than adequate. Sorry!

That said, I'd stay away from any solution with binary-only drivers on Linux. Just out of principle (and, that was #1 factor in choosing the Rad7500 over similarly-priced Nvidias).

Re:Nvidia's Detonators are designed to force upgra

[Anonymous Coward]

You are a fool sir.

Their driver architechture is unified.

Do you complain about the linux kernel being 130 megs unpacked, containing code for architectures you dont care about?

Same with nvidia's drivers. Its not the same drivers, its the same install package.

Re:Nvidia's Detonators are designed to force upgra

[MoronGames]

"Doesn't anyone find it odd that they reccomend the SAME drivers for an 8mb TNT card and a 256mb quadra or FX? The latest windows drivers are larger (8.5 Mb) than the amount of memory on the former!!"

Sure, but it's not like the entire driver file is being loaded into to the video card's memory. That's just to control the video card, you know? I'm fairly certain, as well, that the driver that is loaded is different for the newer FX's and the older TNT cards. I'm fairly certain, actually, that each generation has its own driver set inside these big releases.

They sell MB chipsets, too....

[veg_all]

...and they are way late in releasing compilable source (never mind a binary) for Mandrake 9.2 for their nForce2 chipsets. While they were busy worsening their video drivers, I had to go out and buy a new NIC to replace the onboard ethernet! Is this complaint worth tossing out my moderation rights on this topic? Oh, who gives a fuck. Someone else can modify the damn GNA posts.

Re:They sell MB chipsets, too....

[Anonymous Coward]

I guess you forgot to upgrade your kernel to get the nforce2 platform drivers to work correctly.

We all know that average Joe user can upgrade the kernel and instantly have "effortless" compatibility. Why would NVIDIA want to release the source to their drivers when kernel recompiling is sooooo easy?

Re:They sell MB chipsets, too....

[Anonymous Coward]

I don't know what you're talking about man... both the drivers on nvidia's site AND the nvidia nforce packages from mandrake worked great on my 9.2 system. I've also used my nforce2 board (audio, video and ethernet nvidia drivers) on my mythtv box running knopmyth. No hiccups, nothing. It just worked.

Re:They sell MB chipsets, too....

[Foddrick]

I noticed forcedeth on Kernel Traffic a few weeks ago. It's an OSS driver for the onboard nvidia ethernet card. It's alpha but it couldn't be any worse than the closed source driver could it ?

Re:They sell MB chipsets, too....

[Foddrick]

It would help if I included a link

Re:They sell MB chipsets, too....

[Tin Foil Hat]

Is this complaint worth tossing out my moderation rights on this topic? Oh, who gives a fuck. Someone else can modify the damn GNA posts.

Thanks, that was nice of you. We really don't need a new GNA version anyway.

Re:Nvidia's Detonators are designed to force upgra

[DeathPenguin]

>>Doesn't anyone find it odd that they reccomend the SAME drivers for an 8mb TNT card and a 256mb quadra or FX?

Not at all--Their unified driver architecture helps to make sure that no matter which card a user has, he only needs to download one driver package.

Believe it or not, many people don't even know which video card they own, much less could they tell the difference between a GeForce MX 200, GeForce MX 400, GeForce MX 440, GeForce FX 5600, GeForce 5900, GeForce 5950 Ultra, one of many varieties of "Quadro" based products, etc. It's easier and a lot less error prone to tell people just to download the 'latest nVidia graphics driver' than to go into the device manager or lspci or whatever and figure out exactly which piece of hardware they own.

Re:Nvidia's Detonators are designed to force upgra

[Anonymous Coward]

*chuckle* I've ran Linux since about '92. I've watched the size of just about every executable grow due to "new features/bloat" as you call it. Here are some sample processes from the OS I'm running now (RH 9):

X - 108 megs (Sure, I know - AGP Aperature is in here too)
kdeinit - 25megs
kdeinit: dcopserv - 27 megs
kdeinit: klaunche. -28 megs
kdeinit: kded - 39 megs
kdeinit: knotify - 38 megs
kdeinit: ksmserve - 29 megs
kdeinit: kwin - 31 megs
kdeinit : kdesktop - 31 megs
kdeinit: kio_file - 26 megs
kdeinit: kwrited - 30 megs
pam-panel - 11 megs
gconfd - 10 megs
kdesktop - 28megs
gnome-volume-control - 17 megs (WTF?! a mixer?)
bonobo-moniker-xml - 11 megs
evolution - 28 megs
kdeinit: konsole - 31 megs (a terminal program?!)
kdeinit: kicker - 32 megs
galeon - 120 megs!!!

Notice, I just happen to be running KDE at the time. I bet it would be the same had I chosen Gnome as my desktop - as I frequently do.

So, lets stop talking about Bloat and the "MS Office ever-increasing footprint", people... It's here too, come to terms with it. Red Hat 9 will *not* run reasonably well on a 486. Heck, I wouldn't really enjoy running it on a 500mhz machine.

Who needs a desktop environment anyway? Time to switch back to plain and simple window-manager.

NWN works

[DeathPenguin]

Funny, I can play Neverwinter Nights just fine with my Ti4600 and 5328 drivers. Heck, I'm even using kernel 2.6 with the Minion.de driver patch.

Yes, I've seen a lot of complaints about the drivers on the nVnews.net forums, but I really wish Slashdot editors would refrain from making blanket statements.

Re:NWN works

[Natalie's Hot Grits]

its not a blanket statement. Nvidia drivers for both nForce motherboard chipset and GeForce Graphics chipsets has never _EVER_ been of production quality. nVidia under linux is asking for trouble.

It's too bad too because combined with their drivers for windows, they have the best motherboard platform even when compared to intel chipsets... its a real shame.

Re:NWN works

[SQLz]

ts not a blanket statement. Nvidia drivers for both nForce motherboard chipset and GeForce Graphics chipsets has never _EVER_ been of production quality. nVidia under linux is asking for trouble.

Well, how do you explain the pretty much community wide notion that if you want to game or code GL apps in Linux, nVidia is the way to go?

Nvidia drivers cause problems for n00bs and the unitiated because they have no clue what they are doing.

For example, to get the nvidia drivers working, you may have to: (same things go for ATI) edit your XF86Config recompile your kernel untar a tar.gz file! edit /etc/modules.conf type: make install

Some of these things may pose a huge problem for people, especially the kernel and if not done right the drivers will never work correctly. Who knows what other stuff they have done to their system as well.

Take it from me, I help people get nVidia cards working in X via IRC/Forums all the time and the problem is rarely the drivers, its always the configuration.

Also, there really is no Linux Nforce chipset drivers. The sound driver is just the normal OSS Intel i810 driver. Their is an open source version of the Nforce-Net driver now as well so for the Nforce platform you should be using that for network and either the ALSA i810 driver or the commerical OSS i810 driver if you need digital out/hardware mixing.

Re:NWN works

[Asdex]

> Nvidia drivers for both nForce motherboard
> chipset and GeForce Graphics chipsets has never
> _EVER_ been of production quality.

Then, which other 3D-capable driver has ever been of production quality?

You are writing best of FUD.

- Nvidia is one of the few companies that releases fast and feature rich x86,AMD64,IA64,...drivers
- They even provide a own AGP Gart driver if the standard Kernel driver does not work well.
- Their first drivers have had really some problems but over the years(!) they got the best working and most feature rich drivers for XFree.

It's good to see a hardware company that actually employes several Linux-driver-developers.

Re:NWN works

[Zebbers]

Ive never had a problem, in fact Ive had more problems with windows graphics drivers than I have ever had with linux ones.

Re:Nvidia's Detonators are designed to force upgra

[Anonymous Coward]

'the ever-increasing memory footprint of explorer in windows?'

This is due to new features, like really nice thumbnail browsing. Many people, like me, like these new features and have no prob with memory requirements. Ditto nautilus, konqueror etc.

'they convince you that last year's $400 AGP wondercard needs to be replaced yet again'

Maybe on Windows, where there are apps (games) to stress such features. But on Linux (the target for the drivers in this story), where there isn't much common 3D stuff going on, no way. My GF4 would last me a heck of a long time in Linux.

'They have grown exponentially in bells, whistles, background processes and systray apps for the last 3 or 4 years and will (I'm sure) continue to do so for the forseeable future'

So what? We're talking a system tray applet and a set of control panel tabs (again, on Windows). Who gives a tooble flooble? Has this impacted upon your computing AT ALL?!

'Doesn't anyone find it odd' - this last bit is just stupid. I suppose you have the same beef with the linux-2.6.0.tar.bz2 and other stuff.

5328 is awesome!

[Anonymous Coward]

The newest drivers are great. I don't think people should complain about their northbrige being unsupported by nvagp, or about their own lack of skillz, that's just my 2 lemurs though.

The 4496 drivers totally sucked. They had display glitches in 3D programs (at 640x480 visual tearing in the middle of the screen even though I enabled vsync, and even worse at 800x600 there is distorted 'garbage' at the lower right corner of the screen, no glitches at higher resolutions though).

So the 4496 drivers are unusable to me, however the 5328 drivers rock! The performance is faster for me, no more strange artifacts or tearing, and yes with vsync enabled the fps on ut2003 has DOUBLED!

5328 is faster on linux kernel 2.4.x than on 2.6.x, but really I am [YOU ARE] lucky to have the very latest kernel supported so quickly :p

BTW for those who never RTFM, you have to set __GL_SYNC_DISPLAY_DEVICE to your device if you want it to actually sync to refresh. Also, pageflipping is now on by default and the Option name is changed, so look at your XF86Config people...

Thanks Nvidia and thanks Zander too :)

Re:Linux really needs to make inroads in this area

[Anonymous Coward]

Thankyou for your tremendous insight, telling the Linux community something they have never heard before. Are you interested in doing any professional speaking engagements? I'm sure you would be in high demand.

Anyway, I love these kinda comments. Here I am just thinking about these drivers, and some guy pipes up and starts talking about 'adoption of Linux'. Always with the adoption of Linux. News flash - I've already adopted Linux, and don't give a shit whether half-assed companies like Redhat and Novell can make it commerically viable or not!

Rambling Post:

[big_groo]

From the article, it sounds as if it's a certain card that seems to be having problems - if it doesn't work for you, well, you should have backed up your kernel before trying. I have no sympathy for people who are *going* to say 'I can't even get X to start!'.

Another poster mentioned that someone has already built 2.6.0 .run files - that's cool, but I have no reason to change my *stable* system. Maybe next week when I'm bored.

Reading though this thread I've seen people extole the virtues of ATI and slam Nvidia. One particular poster said (s)he loves h(er/is) 9800. The first Google I get on this card shows a price of $299.00 US. I don't know about anyone else, but I think this is a *total* waste of money. I upgraded my last video card from a TNT2 (32MB) when I couldn't install Unreal Tournament 2003. Time to upgrade. Picked up a GeForce4 MX440. $99.00 CAN. I bought this card for one reason: Nvidia had drivers for Linux - and as a recent Linux convert, let me tell you, this is good news. Cudos to Nvidia - they'll get my $$ when it's time to upgrade again, and I'll get a card that's equivalent to the 299US card for 99CAN.

Re:Rambling Post:

[Anonymous Coward]

nvidia won't release card specs to the dri project

i bet you advocate slavery too

Re:Rambling Post:

[mcbridematt]

AFAIK Utah GLX is working on an Riva 128/TNT driver. Good luck to them. I was using my TNT with the NV binary drivers until the start of this year. Still put out playable fps at 1280x1024 in Quake 3(Linux)

Re:Rambling Post:

[BiggerIsBetter]

Funnily enough, I did exactly the same thing. This despite being a Linux convert of about 6 years now. As much as I think open sourced drivers would be great I've had minimal problems with them, and even if I was using an ATI I'd probably run the XiG server anyway...

Re:Rambling Post:

[Anonymous Coward]

GeForce 3, GeForce 4, GeForce FX. Yup, one card. Enjoy the $99 GeForce 4 MX, you could have bought an old GeForce 3 for less and gotten better performance. I'll keep in mind that I should take my advice from you in the future, seeinng as how you're stupid.

And how is this a good thing ?

[anti-NAT]

I'm of the opinion that you should only dumb down tasks to a certain level. After that point, the task passes its efficiency threshold, and become less efficient..

Not at all--Their unified driver architecture helps to make sure that no matter which card a user has, he only needs to download one driver package.

It's easier and a lot less error prone to tell people just to download the 'latest nVidia graphics driver' than to go into the device manager or lspci or whatever and figure out exactly which piece of hardware they own.

Honestly, I can't see how this philosophy saves time. I believe it would take no more than 10 seconds to perform (1) open a terminal window (2) run lspci, and look for the "nvidia" item. Bang, you then download the appropriate file.

Assuming it takes 30 seconds, I'd suggest downloading the much larger single file would take a lot longer than the time saved.

If people can't cope with simple tasks such as this, they probably shouldn't be performing any maintenance on their PC at all, in particular installing new video drivers.

Re:And how is this a good thing ?

[Anonymous Coward]

Since I have well over 100 computers to maintain with almost as many different video chipsets, it is nice to know that as long as it is an nVidia chipset, I just have to use that one driver. You have no idea how nice that is, apparently.

Re:And how is this a good thing ?

[bovinewasteproduct]

I belive that they have ONE driver source tree and just do builds outside the tree based on platform. If they had to recreate their entire build system/layout just Linux, they may not have even supported it...

BWP

Re:And how is this a good thing ?

[DeathPenguin]

>>If people can't cope with simple tasks such as this, they probably shouldn't be performing any maintenance on their PC at all, in particular installing new video drivers.

How wonderfully arrogant.

Why not simplify maintenance tasks so that more users can perform them on their own? I guess some would rather keep this sort of elitist attitude that makes the geek community look bad, particularly the GNU/Linux users.

Besides, maybe it's not just novices that could use a little help. Perhaps some people have multiple machines with different nVidia graphics cards. Should they be expected to log in to each machine, lspci, and download + install the proper driver for each different machine? Or just download one driver for which they can hack up a simple script to distribute to all machines at once and install / update automatically?

Re:And how is this a good thing ?

[anti-NAT]

How wonderfully arrogant.

I don't think so. If you can't cope with opening an xterm, and running lspci, how can you cope with performing any other tasks under linux, possibly even using it, or managing to find and use the power switch ?

User friendliness is fine up to a point. After that, it just gets in the way of people who don't need it, and gives a false sense of competence to those who do "need" the user friendliness.

Re:Nvidia's Detonators are designed to force upgra

[xlogicalxendx]

i love linux but all this stuff is just so darn scary.

what about sleep?

[ejaw5]

Judging from the Drivers page and README, it seems they haven't yet addressed the problem of the computer not able to go into sleep/suspend while the driver is loaded. A bit of a nuicance for notebook users...

Re:what about sleep?

[mbertini]

I've installed the drivers yesterday on my DELL Inspiron 8200 (RH 9). APM suspend works, it's standby that is not supported.

Re:Wonderful!

[Anonymous Coward]

The "source code" consists of some headers and a couple large binary object files, so you can relink the kernel driver against whatever kernel you have to be running. The X11 driver is pure closed source.

Nobody should praise NVIDIA as a perfect Linux citizen for providing these drivers. They do work, and it is a better solution than some hardware companies that provide no support at all. But they have consistently refused to provide documentation to write a fully free driver, so there is no 3d support at all and usually the XFree86 driver does not support the latest generation or so of chipsets at all.

Ignorance is not bliss, it's annoying.

[Anonymous Coward]

Way to be a brain-dead l0ser boy there Digital.

I'd define bloat differently.

[anti-NAT]

I'd consider bloat to be when there is very little increase in functionality, yet the cost in consumed disk space or RAM is significant. I would consider MS Clippy would be a canonical example.

Does KDE offer increased, and more importantly, useful additional functionality to you ? If it does, then you have decided to accept the extra CPU, RAM and disk space it requires. It could also be argued that the visual "beauty" of the environment makes your computer more pleasurable to use, which will increase your efficiency.

OTOH, if you consider KDE to be pure bloat, then you are right, you probably should go back to using twm, resurrect your old 486.

From that site (as mangled by google)

[DAldredge]

Following the news concerning the bizarreries in drivers XGI, we carried out some tests by modifying the file of drivers XGI so that the applications is not recognized any more (modification of the character string 3 D m has R K 0 3. E X E in 3 D m U R K 0 3. E X E for example) since the simple renaming of file did not make it possible to circumvent detection. The results are rather impressive, since for example one passes from 20.6 fps to 8.8 fps in the scene Mother Nature of 3DMark03 (1024*768, the total score follows the same tendency), or 88.2 with... 18.7 fps in the bench BotMatch integrated into Unreal Tournament 2003 (1600*1200). It will be noticed that the fall is however less in our own demonstration, more realistic, since the score passes "only" from 28.6 to 14.6 fps... Does D?ou come this deceleration? D?optimisations specific to these applications of course. These last are can be multiple, but the checking for example of plane l?integration of clip for the demonstrations (like NVIDIA in 3DMar03 at one time) n?est not verifiable without assistance of the developer. However, we noticed that when the driver detected one of the listed applications, it decontaminated trilinear filtering to make only the simple bilinear one, i.e. without any transition between different the level from mip map. Of course that reduces the workload by two (interpolation from 4 texels instead of 8). In fact being given the importance of the fall one can even wonder whether the second GPU is well activated "by defect"... Here to illustrate our remarks a screenshot under Unreal Tournament 2003 with the drivers d?origines, and another screenshot once that these detections are decontaminated: We used the order firstcoloredmip here so that the various levels of details is coloured in order to highlight well the difference at the level of filtering. Worse, under Halation with detection d?origine, graphic quality is really deplorable, has such point qu?on would be believed in 640*480 when one is in 1024*768, so that l?on thinks qu?il s?agit not only d?un trilinear problem of filtering. This problem is solved when l?on decontaminates optimizations, but blow one attends slide show. It should be noted that the function screenshot does not function correctly when Halo is detected by the driver, which prevents us from showing to you the difference in graphic quality. Here are which thus explains mainly why in the benchs running Volari s?en leaves relatively well, and why in less current plays for the benchs but inevitably less not played c?est the catastrophe. In our protocol of usual test using in major part of the not detected plays, Volari Duo V8 Ultra arrived indeed at 55% of the performances d?une GeForce4 Ti 4600, with in the best of the cases 77% of the performances of the latter under UT2003 and specific optimizations which go with (40% without) and 15% of better under Quake 3, since in OpenGL by defect and contrary to what occurs in Direct3D XGI decontaminates office trilinear filtering and also lowers by defect the level of detail of textures as you can note it by comparing filterings of ATI and XGI: With final, one can wonder which is l?interet for XGI d?integrer this kind of things in these drivers. Indeed, if XGI thinks that its chips are not enough fast to carry out a trilinear filtering, of decontaminating it completely as much drivers rather than to decontaminate it that in certain app

Works for me,

[anti-NAT]

but that is because I refuse to run binary kernel modules, and only buy hardware with open programming specs.

Re:Works for me,

[jejones]

Could you tell me what graphics card you use, then? I'd like to go buy one of them, and send the vendor email telling them why I'm doing it as positive reinforcement.

Re:Works for me,

[anti-NAT]

The last video card I bought was a Matrox G550. I don't play OpenGL based games, where high FPS mostly seems to matter, so this card is fine for "normal" 3d stuff, and provides great visual quality, which Matrox have a reputation for.

However, since that time, they've released the Parhelia, with binary only drivers (at least that is what I last found - I'm not looking for a video card at the moment, so I'm not following this sort of information all the time.). Although not ideal, at least these binary drivers are only XFree86 ones, and don't have to be installed inside the kernel, risking system stablity.

Have a browse around Direct Rendering Open Source Project as well as XFree86 for open source X Windows support

Re:REPORT YOUR EXPERIENCES BELOW

[Clockwurk]

WORKS FINE WITH COUNTER-STRIKE AND QUAKE III,

I almost bit on this, but when you posted the bit about Counter-Strike running under Linux ( it only runs in Windows), I knew you were trolling.

HTH. HAND.

Re:face it nVidia

[Fnkmaster]

Um, no. How about the fact that the Windows driver APIs remain constant within any major release of Windows, whereas Linux driver APIs (or really ABIs here) change slightly even between minor releases of Linux kernels? This is of course intentional - it's supposed to be hard to make closed source drivers. If Linux developers wanted to make it easy, there would be a simple, clear, well documented set of driver development and interface ABIs that only changed with major version releases (2.2, 2.4, 2.6, etc.). Then it wouldn't be necessary to do a partial/stub recompile for every version and have dozens of different packagings of binary and partial-source versions of the NVidia drivers, and they would probably end up much more stable and functional.

So yes, I agree it's quite hard to produce Linux drivers that are stable and functional across a broad range of Linux kernel versions and XFree versions, and I am sure it is in part because there are more users and thus more developers working on the Windows drivers, in large part it's because of the inherent features of the Linux platform. Which of course may be desireable for many people who want to encourage companies to release specs or truly Open Source drivers.

Re:Linux really needs to make inroads in this area

[Anonymous Coward]

And that, my friend, is why Linux on the desktop will never be successful. The "community" if full of arrogant, selfish halfwits determined to keep the code as archane, so they can pretend to the world that they are untouchably elite and highly skilled.

The minion.de patches...

[0x1337]

The patches on www.minion.de allow the latest NVIDIA driver to compile/run under 2.6 - as well as fix the crash with VIA-chipsets.

The patches against the NVIDIA driver at www.minion.de allowed the driver to successfully load on kernel 2.4.22 on my VIA kt266a-based motherboard with FX5200-based video. (That is - without the Ooops crashes descirbed in my previous post)

X11 starts normally. 2D graphics seems same.

However...

Glxgears has horribly mutilated graphics. The shapes aren't filled in, its all blinking/flashing and there is a lot of extraneous junk/shapes randomly appearing and disappearing really fast. Its slow too at that - 400fps. I tried ut2k3 - same story - horribly slow and mutilated graphics. it says that AGP was working. Checked dmesg again - no errors.

Sigh - just not a very good driver, I'spose.

Re:face it nVidia

[jrockway]

> And Linux nVidia Driver will be always described with one word: crappy.

Let's give nVidia a break. Yeah, one bad version of drivers. They work fine for me (although not with AGP enabled). Anyway, let's see if they fix this. They probably will, and we'll all be happy again.

And BTW many games get higher fps under linux than windows. Crappy indeed.

Re:face it nVidia

[Jacek Poplawski]

And BTW many games get higher fps under linux than windows. Crappy indeed.

FPS is not everything. In Linux I could expect stability. Do you have framebuffer enabled?

How Soon We Forget...

[Anonymous Coward]

NVidia are the same fuckers that stole every 3D technology from 3Dfx and then claimed it was theirs in the first place. They are also the ones who did a hostile takeover of 3Dfx and killed off ALL LINUX development and support for truly great products. At the very least, they could have kept the 3Dfx VooDoo cards from VooDoo3-5 supported after they KILLED 3Dfx. But instead, the day they took over, they eliminated all support and left all of us VooDoo users high and dry. I WILL NEVER buy a card from a company like NVidia. They are the Microsoft of the 3D accelerator world. They didn't have a good technology of their own, so they used corporate espionage to steal it from 3Dfx and then claimed it was theirs. NVidia are criminals and they should be exposed for the evil that they are. I suggest that NO ONE buy any NVidia cards as they don't truly support or understand Linux and the Open Source/Free Software world. For now they are paying lip service to it, but that is only to lure the suckers in and then play bait and switch and force them to move to Windows. Watch for the Linux support to vanish within the next year or two. Instead stick by a company with a long track record of Linux support: ATI. Check out the GATOS project at SourceForge and you will see what a company who truly groks Open Source/Free really acts like. The rest of you NVidia fan boys can continue on to the cliff like the lemmings that you are. I can't wait to see your jaws drop when you find out that NVidia ownz your asses and will force you to use Windows very soon. Fucking moron NVidia fanbois.

Re:How Soon We Forget...

[Anonymous Coward]

I worked with the GATOS project while doing level 2 support at IBM, and the ATI was extremely less then helpful in anything we asked them to do, or in regards to what we could release.

Re:How Soon We Forget...

[Anonymous Coward]

I hate you fucking idiot moderators. I had a pretty good thing going there for a few hours with a moderation of +1 Insightful as it should be. But now you asses have bombed it down to Flaimbate even though that's completey incorrect. It's amazing how you monkey sheep fanbois will come do the defense of anything that is held to be most holy in geek terms. I'm sorry but you guys are totally off base on this one. 3Dfx was THE premiere 3D acceleration company until the assclowns at NVidia stole everything and tried to kill off Linux support. It's not a wonder that the drivers suck so bad. They don't care about Linux, they just care about getting ATI out of the way. I hated ATI before, but at least they aren't NVidia. Yeah... free market my ass. The market has shown that the company with the most money will bury all the other ones by hook or by crook regardless of the quality of their product. That's why America and it's capitalist system will always produce inferior and mediocre products rather than great works that will last for ages. Fuck you to all the moderators. I hope you all get modbombed to fucking hell you jerkoff bastards!!!

fedora + 4496 nvidia drivers

[painehope]

the 4496 drivers would lock my workstation at work ( Ti4200 card )every few days using the default Fedora kernel. I think there's a newer kernel out, might try it again.
My solution was to revert to the latest RH9 kernel, as I don't have time to chase down a bug in a closed-source driver, that locks X, so I can't see any console messages, and prints nothing to syslog.

Re:Linux really needs to make inroads in this area

[Digital Dharma]

I guess if the shoe fits, wear it. I didn't intend for this to come off like a troll, just commenting on my observations regarding the community. I guess if I don't support Linux as a zealot than I deserve the rank of troll in your eyes, is that it? Or maybe the zealots can't stand to have reality presented to them in a factual way. If that is what Slashdot has come to, I'm out. Delete my account please, CN. Your followers have abandoned reason for madness...

So, Intel are going to fail to sell 10GBit NICs ?

[anti-NAT]

because, I've already got their GPL driver, that they wrote themselves ...

-- The full GNU General Public License is included in this distribution in the file called LICENSE. Contact Information: Linux NICS Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497 Makefile for the Intel(R) PRO/10GbE driver --

/usr/src/sys/kernel/linux-2.6.0/drivers/net/ixgb

As I understand it, hardware companies are in the business of making money from selling hardware. Drivers for that hardware would usually be considered a cost. Giving the community the hardware specs allows the community to absorb the driver development costs, allowing the hardware company to both sell more hardware, as they are community friendly, and the community supports its friends, and make more profit on the hardware, because they reduce some of their costs, as well as have the possiblity of selling hardware to people who may not have been customers in the first place. (for example, I remember a post to LKML with a patch for running an e1000 aka Intel 1Gbit card under PPC. If Intel developed closed source drivers, do you think they ever would have developed PPC drivers ? I doubt it, it would have cost them too much, for possibly a too smaller customer base. By open sourcing, they have sold at least one extra card that they wouldn't have sold before, under the closed driver development model)

Re:Once again, Linux displays it's flaws.

[Anonymous Coward]

Not true... The most current nVidia drivers for my Gainward Geforce2 GTS installed on Win2k are crap. They completely break it. Seems to me nVidia just isn't doing a good job in supporting its older variety cards.

Notice all the "hey these drivers kick ass" comments are comming from users of GF4s or better?

Note: Here's what I mean by being horribly broken.
http://www.hayenga.com/mitch/mario1.gif

My only complaint about nvidia...

[rsilvergun]

is that they sell chips to whoever buys. True story: I bought a Geforce2 MX440 with 64 megs of DDR ram that _under_ performed my old 32 meg Geforce2 MX SDR. Near as I can tell the trouble is cheap ram, really cheap ram. It's the easiest place to save some money, and it seems the Best Buys and Circuit Cities of the world will buy cards from whoever offers the best profit margine. I mean, why sell an $80 card for $99 when you can sell a $50 card for that.

That said, I tried the ATI equivalent and gave up on it because of lousy driver support (couldn't get xv working, dvd play back sucked even under windows). Right now I'm stuck with my old Geforce2 mx. I'll try again later to track down a better card.

Re:My only complaint about nvidia...

[DeathPenguin]

ATi has recently turned to this practice as well. You can get ATi-based cards ("Powered by ATi") from Asus, Tyan, and a variety of others.

Re:My only complaint about nvidia...

[BenjyD]

The DDR and SDR Geforce2 MXs are virtually identical in performance - the DDR one should be a few percent slower, as it has the same memory bandwidth but a slightly higher latency (IIRC).
That's because the SDR has a 128bit bus, while the DDR has a 64bit bus, at double-speed. Same bandwidth pretty much. It's nothing to do with cheap RAM.

Re:My only complaint about nvidia...

[incom]

My geforce2mx actaully came overclocked from the factory. and is faster than average. only problem is they used a shitty heatsink and I had to do some crazy cooling improvizations for it to run stable. And since it's only worth peanuts now, it's not worth it to buy a fancy new heatsink. This is a pci version too.

Howto with Fedora Core 1

[ags]

I had lots of problems getting NVidia to work with FC1. Things would kind of work, but other things wouldn't. Getting TuxRacer to work is a good litmus test.

Then I found this page of unofficial FC1 FAQ. Yay...!!

Here's what to do - it worked for me:

Use these instructions if there are no RPMs available, or if the available RPMs don't work for you.

Make sure you have the lastest drivers.

Now print this out, or write it down. Then:

1. You must have kernel-source installed for this to work. Check the "Add/Remove Applications" tool in System Settings on the red-hat menu.
2. Shut down X (as root, do telinit 3 in a terminal).
3. Log in as root and go to the directory where you downloaded the nVidia drivers.
4. Type CC="gcc32" sh NVIDIA-Linux-x86-1.0-4496-pkg2.run --add-this-kernel.
5. There will be some whirring and wailing, and then do CC="gcc32" sh NVIDIA-Linux-x86-1.0-4496-pkg2-custom.run That will do a correct install of your nVidia drivers.
6. Follow nVidia's instructions to set up the driver. (Read the "EDITING YOUR XF86CONFIG FILE" section.)
7. You can now restart X by typing telinit 5.

If none of this works, do rpm -e --nodeps XFree86-Mesa-libGL and then restart your computer. The need to do this should soon be eliminated -- watch this FAQ or the fedora-list. Note that if you update XFree86, this package will be reinstalled and you will need to remove it again. This solves the "DRI" problem.

Re:Howto with Fedora Core 1

[Anonymous Coward]

You don't have to have gcc32 to make it work. I seem to be hearing a lot of complaining about how much time it took to make the driver work in FC1 and honestly I don't get it. It was no harder than making it work in any other distro in my opinion, there was just one difference in that you would get a compiler WARNING, you could tell it to ignore it and it would work fine. The installer even tells you how to ignore it. This is by far one of the EASIEST linux problems I have ever had.

Re:Howto with Fedora Core 1

[Anonymous Coward]

I had to set CC="gcc32". I tried to provide whatever argument it was that bypassed the 3.3.2 compiler warning/blocking, and that didn't work (X wouldn't start).

This was with the 4496 driver.

Not just Linux Drivers

[I Want Your $$$$]

I know their windows drivers have a few . . bugs . . to work out as well. Last time I installed their 'new' drivers it left my WinXP box dumping core everytime it tried to load past the boot-up screen.

Re:Not just Linux Drivers

[Anonymous Coward]

Yes, but as a windows user, that type of behavior is considered acceptable, even expected.

Re:REPORT YOUR EXPERIENCES BELOW

[0x1337]

*cough*

Well - I've gotten CS 1.3 and CS 1.4 (note - not CS 1.5, gah) to successfuly run under Wine (which was whatever the latest CVS was two summers ago). I compiled WINE with the OpenGL wrapper and the results weren't bad - it was playable. Not as fast as under windows (50 fps instead of 70 - and yeah, I dual-booted then - mostly cause of CS) - but pretty good. The cheat protection tho saw Wine's opengl.dll as a cheat - so I had to play on unprotected servers.

This entire summer I played Day of Defeat 1.0 on top of HalfLife under Wine (whatever the latest CVS was, lol). HL acted more buggy. I couldn't use the VGUI without a crash - so I had to run HL with the servername on the commandline. Its alright if yu use XQF or some online server browser. The really cool thing was that the Wine folks fixed the whole "wine's opengl.dll cheats" bug - so I could play on real servers.

However, I had no luck at all with Steam under wine. I'll probably go back to it sometime - see if it works - and if it doesn't , attempt to implement the numerous unimplemented functions it needed.

Success story

[BigBuckHunter]

Loaded the minion 2.6 patched driver. Seems to work fine on my FX5600 KT333 using kernel AGPGART. Played some SMB in mupen64, RTCW, and NWN. I don't use any special features (twinview and TV-out) so I have a pretty vanilla config. Using renderaccel and cursorshadow

Gentoo with Gentoo-dev-sources with Accept "~x86".

Thank you for your time,
BBH

Improved performance

[locus_standi]

On my system, the new driver seems to perform much better if frame rate from glxgears and quake3d is any indication. I am seeing as much as 25% improvement in performance over the 4496!(using 2.4.21 kernel). I also tried the new drivers with patches from minion.de on a 2.6 kernel and could see improved performance. However, synaptic touchpad and pcmcia Wi-Fi seem to be broken in 2.6 so had to go back to 2.4 kernel.

Re:Improved performance

[DF5JT]

"However, synaptic touchpad and pcmcia Wi-Fi seem to be broken in 2.6 so had to go back to 2.4 kernel." Huh? PCMCIA WiFi definitely is not borken in the 2.6 kernel, it even supports a wider variety of cards than with the previous versions.

Re:Improved performance

[eGabriel]

Wi-Fi should work fine. Be sure to do whatever is necessary to get your modules to load, using generate-modprobe.conf or whatever tool is appropriate.

As for Synaptic touchpad, you can't just get away with using the PS/2 driver as is. You need to download synaptic's driver and enable it in the X server and have kernel support for it. Then it works better than ever.

Thanks Slashdot ...

[Anonymous Coward]

Yesterday I upgraded my kernel (Red Hat 8) and installed the brand new nvidia drivers at the same time. Everything seemed to work O.K. with the exception of xawtv. As far as I can tell the modules required to run xawtv refuse to load while using the nvidia driver. Because I`m a trusting sort of person I assumed the problems I was having were my fault - I really didn`t expect nvidia to launch buggy drivers ! You`ve just saved me a lot of time and trouble.

And how is this a bad thing ?

[anti-NAT]

I'm pretty sure Linux never needed Nvidia to provide graphics drivers to succeed - it was doing fine by itself. Keep in mind that Linux had been going for something like six years (starting in 1991) before Nvidia came along and released Linux specific binary drivers (around 1998, IIRC).

A lot of people may think Linux is unreliable, because of the instability of binary drivers, such as the Nvidia ones. Linux could do without this undeserved reputation.

Re:And how is this a bad thing ?

[Shinobi]

That's funny.... The GATOS drivers for the Radeon 7500 in my secondary box has caused me more troubles than ATI's binary drivers have, or Nvidia's binary drivers for the Ti4200 in my primary box.

3 days and X is still refusing to run when I use the GATOS drivers

Re:And how is this a bad thing ?

[anti-NAT]

So why aren't you trying to improve the GATOS drivers ? You've got the source code ...

Re:And how is this a bad thing ?

[Shinobi]

Because I'm busy programming other stuff. I simply don't have the time to spend on it.

I'm just annoyed at dubious claims as to how Open Source is always better.....

Re:face it nVidia

[Jacek Poplawski]

I don't get it. Do you mean all nVidia driver problems are related to changing kernel ABI? But that part of drivers is open source.

IMHO problem is somewhere else. There are millions of possible configurations, and few people working on driver just can't solve each problem. In open source driver there are more workers (they don't need money from company), so it's much easier to fix driver for individuals.

That's why they have patents

[anti-NAT]

Patents are intended to protect intellectual property. Nvidia shouldn't be worried about protecting it in their closed source drivers.

Actually, thinking about it, if they really need to protect their intellectual property within closed source drivers, those drivers should be encrypted, preventling disasspembly and decrypted on the fly. Hmm. That's not going to work, cause if you really wanted to find out their secrets, you'd just use a AGP bus analyser or some other similar device.

Other people may disagree, but I belive this "intellectual property, must close source" is just a furphy.

Even more contrary to this argument, companies like Nvidia and ATI want their extensions to be added to the OpenGL specs, which results in both an industry wide endorsement of their techniques, as well as licensing revenue from their competitors.

Re:That's why they have patents

[Osty]

Actually, thinking about it, if they really need to protect their intellectual property within closed source drivers, those drivers should be encrypted, preventling disasspembly and decrypted on the fly. Hmm. That's not going to work, cause if you really wanted to find out their secrets, you'd just use a AGP bus analyser or some other similar device.

You're right, with a certain amount of effort whatever they're hiding in the closed source drivers can still be found and copied. However, the idea is to make it as unappealing as possible to try to steal the information. As it is now, someone at ATI would have to consciously make an effort to get that information. If nVidia opened the source, then any ATI developer could browse it, and inadvertently (or advertently, even) insert nVidia code into ATI drivers.

By your argument, I shouldn't bother with locks on my doors, because if someone really wanted to get in they could easily pick to lock or break a window. However, we lock our doors because it makes us a less inviting target, and thieves will move on to easier pickings. It won't stop them if they're determined to get in, but few of us have to worry about such targetted attacks.

Even more contrary to this argument, companies like Nvidia and ATI want their extensions to be added to the OpenGL specs, which results in both an industry wide endorsement of their techniques, as well as licensing revenue from their competitors.

The GPU-specific portions of the extensions are irrelevent to the OpenGL standards, so why open the source? And I'm sure if ATI wanted to license nVidia's driver techniques, nVidia would be more than happy to do so and give ATI the source. That doesn't mean you or I would see it, though.

Re:That's why they have patents

[Lozzer]

To complete your analogy, you also need to point out that there are only three houses in the world.

not really

[rebelcool]

I rarely upgrade video drivers due to the old adage "if it aint broke, don't fix it" is a good thing to live by. (case in point: upgrading to the newest 50-series drivers causes Win XP to core dump and wrecks certain system files when playing Enemy Territory on my system)

Because of this, I might only upgrade my drivers once a year or less, basically only if I run into a problem. Trying to remember the exact card model I have after months of never thinking about it is rather difficult. "Hm, GF3..Ti200? Ti400? Which one did I buy again..."

they suck on GF3 as well

[rebelcool]

The 50series of drivers for WinXP cause a system crash when playing certain games with Punkbuster. On my system it hosed some system file requiring a reinstall of the OS off its boot disk. I have a MSI GF3 Ti200.

XP makes the process fairly painless (all your settings are saved for everything), but it still requires a couple of hours to reload off the disk and then reacquire service pack 1 and assorted patches to bring it back up to date. And theres the occasional need to reload a driver or two.

I downgraded to a 40series driver thats about a year old, and it works flawlessly...

Just wait until Fedora Core 2...

[relaxed]

"I spent a few long hours recently trying to get the old drivers to work with Fedora Core 1 so I'm going to hold off on these new ones."

1. If one of the Fedora XFree86 developers has his way with FC 2 you won't even be able to start X if any files from XFree86-Mesa-libGL, which currently conflicts with the nvidia driver, are missing or overwritten. And as of right now the 1.0-5328 installer overwrites 4 files so unless something changes nvidia users might get locked out of X in the upcoming release. Scary, eh?

Re:Just wait until Fedora Core 2...

[Anonymous Coward]

Who cares. Mandrake is the way to go.

Fedora is a buggy alpha release that I am disgusted with.

the fact you have ot press the caps lock key repeatedly when installing fedora cord on a newer compaq laptop or as most people are having to do type linux enablecddma to get the installer to even work. let alone the extremely slow overall feeling of fedora core 1.

Mandrake on the SAME hardware had no installer problems and is at least 2 times faster overall.

fedora core = junk. and will be the death of redhat linux.

I've heard a lot of problems with both companies.

[deinol]

and people wonder why ATI suddenly comes out ahead of NVidia.

Last I checked ATI didn't have the best linux support either. I have a friend with the misfortune of having the Nforce chipset motherboard and a Radeon graphics card. Good luck getting the two to work together.

Open source kernel + 1 closed source driver may work just fine. But open source kernel + 2 closed source drivers can mean conflicts and incompatability.

It doesn't help that Nvidia's precompiled agpart driver for the nforce board only supports nvidia graphics cards. After spending an entire day recompiling kernels, I was able to get either 3D acceleration and no onboard network, or no 3D but with the network working.

I'm sure I could have gotten it all working with enough time, but it wasn't my machine. Still, it shouldn't have to be that hard.

I am glad that Nvidia provides a linux driver, but I really wish they would provide open source drivers. If two many closed source drivers are added to the kernel, it will make it unusable.

Re:face it nVidia

[Fnkmaster]

No, I said _a large part_ of the problem is the changing kernel ABI. But that's not the only factor. Another substantial part is that there are more users and thus more developers working on the Windows drivers. You're right, with Open Source drivers the number of man-hours put into updating the drivers wouldn't be limited to the small number of folks NVidia has working on them, and they would probably end up more stable faster, not to mention that the problems with the kernel ABI changing would be pretty much irrelevant as they are with most of the Open Source Linux drivers.

Unfortunately, we all know NVidia can't/won't Open Source their drivers and they claim it's the "can't" due to contracts with other companies that have various pieces of IP in their cards and in the drivers.

Poor Nvidia

[superpulpsicle]

They have been getting such a bad wrap one after another. I'll tell you from owning both ATI Radeon and Nvidia Geforce cards, that Nvidia has always been the one with the better drivers.

Owning an ATI Radeon 9800 now with so many graphical features disabled and tweaks in general, I swear I am almost better off owning a Geforce FX5900.

It's a bad sign when...

[Caeda]

Your reading about driver problems, go to download the new driver and try it, and 2 different web browsers lock up just trying to download the file...

Re:Once again, Linux displays it's flaws.

[Anonymous Coward]

Oh yeah? This has happened with Windows tons of times, it's just that most people don't notice if their screen is running at 640x480 in 16 colour mode.

Works fine for me without any tweaking

[Little Hamster]

I guess I must have lucked out. Reinstall my desktop today with Mandrake 9.2, and since the old one I have required a kernel interface compile (it's way too old), I downloaded the new one (5328), and worked first time I installed it. The only thing is that I can't load the GLcore module. I think sometimes you can't just take these forums as an indication of how wide spread the problem is. People who have problems are also the loudest ones.

General question

[rsax]

OK this isn't directly related to Nvidia but more of a general inquiry. Are there any video card manufacturers who consider XFree86 users a priority, even a small one? For some reason I always had the impression that Matrox released good drivers for XFree86 and that they're video cards are generally well supported. Whenever the "X is slow" argument arises people usually say that X performance is contingent on good video drivers. So what are the best video drivers out there right now? I want to build a new box soon which will be used to dual boot Windows and Linux. I would like to be able to play games on both operating systems but if that isn't possible then atleast be able to have really good drivers for Linux for normal desktop use.

Re:General question

[craqboy]

I think you mentioned a good name when you mentioned matrox. I will give nvidia a high five also for what you are wanting to do.

One thing about matrox is they have been more toward the video special type of applications and multiple monitor support. Its not uncommon to use 16 total, unique displays in a linux box using matrox video cards. Try that with nvidia. Matrox for gaming isn't up to par like nvidia would be.

If you want a awesome and stable 2d setup then go with matrox, as their drivers for linux are great. The 3d side is sufficient if you are wanting to bust out a game of quake3 or something but I wouldn't be looking for much more from what my experience. The gaming side of matrox is still lagging and that is why you don't see that many reviews for new matrox stuff; mainly because ati and nvidia are advertising and producing for the group that drives hardware: gamers.

Re:General question

[ratfynk]

Odly enough SIS has good Linux support for their onboard agp chips! I have an old Mainboard with SIS620, the cdrom that shipped with it came fully linux supported in 1999! Sound, video, ethernet, modem and everything else on the board. Sure they are not 'gamer' products but then again, anyone who single games on linux deserves what they get, a waste of great cpu cycles, that could be doing real work!

This is where Linux is retarded...

[mcrbids]

Linux is retarded in the area of any kind of kernel module.

Yes, the Linux kernel is nice and modular, and you can make modules and do some neat stuff (like dynamic loading) with them, but....

Why do modules have to be custom compiled to each exact kernel version?

A binary API should be developed with standard hooks that allow for things like video cards, sound cards, soft modems, scanners, and other crap to operate via a protected, binary-compatable API that doesn't change in any minor release. (EG: 2.4.x should be cross compatable)

One of the successes of Microsoft's hardware compatability is that I can frequently use a driver from Windows 3.1 on my Windows 98 or ME system.

Linux developers can cry all they want to about "open" drivers, but there are plenty of times where that just isn't feasible. And, why shouldn't there be a single, well-documented API that allows for binary driver distribution?

Why should this "pollute" anything at all?

Spending any more than 10 or 20 minutes loading a driver is retarded, and even though I'm a firm believer in Linux and its future, I'll be the first to say this.

Create a clear, binary-compatable API for drivers and the drivers will appear like magic, especially if it's similar to the API for Windows drivers.

Hardware companies are begging for more sales, and if they can get them by recompiling their windows drivers, or at the very least putting out and supporting a single driver file for "2.4.x kernel" Linux, you'll find that lots of companies would be perfectly happy to "play nice"...

Re:This is where Linux is retarded...

[fruey]

I can frequently use a driver from Windows 3.1 on my Windows 98 or ME system.

You might get a printer to work with older drivers, but are you really serious in saying that 16 bit Win3.1 drivers worked satisfactorily for you in Windows? We are talking about high end graphics card drivers here, not keyboard, printer or network card drivers. The only things which most Linux users can't get to work these days fall into two categories:

• software driven hardware... that is to say, hardware which is basically a simple I/O device and all processing happens in the main CPU. Like some modems, cheap USB scanners, etc.
• high end specific cards with separate processing units on them (GPUs, low latency audio processors, hardware encoding cards, TV combo cards)

Correct me if I'm wrong but I don't think you'd get any of those kinds of hardware running in WinME with even a Win98 driver let alone a driver from Win3.1!

Re:This is where Linux is retarded...

[hughk]

I can frequently use a driver from Windows 3.1 on my Windows 98 or ME system.

Not really. In fact not at all. Win 3.1 at a driver level looks very different from Win95 or Win98 which were sort-of compatible.

Many Win blue screens are the result of driver problems. One such problem is the difficulty of maintaining an interface across kernel and GDI versions.

The thing is that the Linux kernel has evolved and is still evolving in the area of I/O services even for basic things like disks. Graphics drivers are much more complicated. If a change to the kernel breaks a driver it can be fixed and recompiled relatively easily.

Of course, the real solution is more advanced I/O services that would allow a 100% source code driver but a userland library to provide the advanced functions.

Note the build for NVidia drivers (which are part source code and part object library) seems a little broken. It should be possible to make a build process that takes a maximum of a few minutes.

Re:This is where Linux is retarded...

[The One KEA]

What you suggest will never happen, because Linus has made it absolutely clear that the Linux kernel will NEVER have a binary-compatible ABI. He absolutely refuses to support closed-source binary modules and will NOT accept patches to the kernel to make it friendly to such modules.

Why? Because it goes against the spirit of the GPL AND because it's a legal grey area which could result in frivolous lawsuits which could waste the time of the OSS community, like SCO is so successfully doing right now.

Unless Linus changes his mind (unlikely) or hardware manufacturers GPL their code and/or release full programming specs, then the status quo will undoubtedly remain.

Re:This is where Linux is retarded...

[Billly Gates]

Actually Linus mentioned he kind of favors proprietary modules and drm if someone wanted to do so. To say no would enforce his opinions on someone else and he wants an OS for anyone who wants it. I am not saying that he supports drm and proprietary software. I am just saying he believes that anyone should do whatever they want with Linux and he does not want to get involved.

The reason he wont include such drivers is because of legal and political reasons. Debian would fork the kernel in a second! They refused to even stock KDE because the precious QT was not free enough.

Anyway there is that and what if NVidia or any other company pulls a sco? What if Nvidia decides to charge for their drivers? Who owns the code for the drivers anyway? How can Linux be GNU/Linux if not all of it is open?

Linus himself would voilate the gpl.

He does what he has to do. What I would do is imate debian's installer in xconfig. THe debian installer tells you only gpl software is installed by default. Would you like BSD and closed source apps as well?

If someone replies yes, xconfig could then launch wget to download Nvidia or other drivers. This way the FSF fanatics can have their pure GNU system and distro's would be legally free since the user and not them decides. Of course by default they would have to have a pure kernel but the user can just recompile it.

Perfect solution.

Re:This is where Linux is retarded...

[labradort]

Quote:

One of the successes of Microsoft's hardware compatability is that I can frequently use a driver from Windows 3.1 on my Windows 98 or ME system.

One could also say that 12 year olds can frequently drive a car sucessfully. It doesn't mean it is a good idea. You'll see Microsoft decided that what you can do in Win 98 is NOT a good idea. That is why Windows 2000 and up have digitally signed drivers that warn if your driver is not known to work. But they still allow you to over ride if you need to, because you do not compile your own drivers and you could be stuck without the "wrong" one. In Linux, you normally compile your own and it isn't an issue.

It isn't Linux that needs to change. It is doing the right thing. It is the hardware vendors. nVidia is nuts about this issue. No one else that I know of releases binary drivers for the sound and ethernet modules, except nVidia.

The cooperation amongst several big players on the Linux Desktop initiative will change this in the next few years. Soon, if you want to supply Compaq, HP, IBM with hardware for their PCs, it will require Linux drivers just as they now require Windows drivers.

Re:This is where Linux is retarded...

[Anonymous Coward]

Looks like the Linux developers are really retarded. Who cares about the fscking GPL when the hardware does not even work?
Stupid kids.

Re:This is where Linux is retarded...

[Anonymous Coward]

the perfect solution here would be to support the same API as Microsoft uses for their drivers.

The biggest deal is that then, you'd be able to use re-compiled drivers for Linux that the companies that do a currently poor support for provide. The manufacturers might start releaseing their drivers for Linux then, promptly and without all the funnies people see.

People may not like it, but it would make lots of sense in the real world, FSF-fanaticism can go hang.

Re:This is where Linux is retarded...

[Anonymous Coward]

It is Linux that needs to change. It is unacceptable for every major release of the kernel not not only not be binary-compatible, but not even source compatible with most of the drivers that are even included in the tree itself. 2.4->2.6 has already meant several things that used to work, do not. If you think rewriting drivers every couple years for every piece of hardware is a sound engineering practice, please be sure to never submit a resume to me.

Re:This is where Linux is retarded...

[T'lexii]

>Why do modules have to be custom compiled to each exact kernel version?

They don't 'have' to. read up on CONFIG_MODVERSIONS

Re:This is where Linux is retarded...

[Anonymous Coward]

you can eat my resume cunt

Re:This is where Linux is retarded...

[oshy]

There is a solution to driver problems with Linux.

If everyone went for the hardware that has Linux drivers then there wont be any problems. Eg, not all software modems have linux drivers, only buy the ones that do (or the hardware controled ones), then the manufacturers that don't supply Linux drivers will have to do something or lose out on a share of the market.

Re:This is where Linux is retarded...

[Anonymous Coward]

Where's "Score -1 NAIVE" when I need it?

Re:This is where Linux is retarded...

[Qzukk]

No, whats really needed is for the hardware developers to GROW UP and realize this isn't the Ad Lib generation anymore.

Once upon a time back in the bad old days, there was this company we'll call "Creative". They produced a card called the "Soundblaster" and released the specs for it so that companies could produce games that used this "Soundblaster". Back then, drivers were a rare breed, and the "Soundblaster" enjoyed its market position of being the only sound card to work with many games.

For a few months. Then other hardware companies used the specs to produce "Soundblaster-compatible" clones, and ruined Creative's market position.

But then, "Windows" came about. Part of the function of a Real operating system is to provide an abstraction layer between software and hardware, and "Drivers" are one way to do that. Now, games were written to use Windows, and automatically used whatever soundcard was available. Soundcard competition moved from the monopolistic "My card is the only one supported" to a feature and quality based competition.

And things were good.

But some companies are still afraid of releasing their specs. They don't realize that thanks to the wonder of modern operating systems, any card could replace theirs, specifications or not.

Re:This is where Linux is retarded...

[labradort]

Well it is pretty hard to work for Anonymous Coward anyway. Something like Intelligent Gas I would think.

What do you consider sound engineering practise... Staying with 2.0 kernel? Using tube powered computers?

Do you even use the 2.6 Kernel? I'm using 2.6-test11 on an nForce2 based motherboard and I'm getting good results. The only problem so far is that the CD burner software needs to be updated to work with native ATAPI devices rather than SCSI emulation.

If you care to open one of your two shut eyes, you'll notice that in the rest of the computing world, it is quite normal to need new drivers. Do your Windows 98 drivers work in Windows 2000?

Think before you post.

Re:This is where Linux is retarded...

[ajaf]

>> Spending any more than 10 or 20 minutes loading a driver is retarded, and even though I'm a firm believer in Linux and its future, I'll be the first to say this.
Huh? 10 or 20 minutes loading a driver? what machine are you using? a clock?
API? what's that? Microsoft? are you from this planet?

Re:This is where Linux is retarded...

[gr8_phk]

I have a Linksys 802.11b PCI card in an old computer and it works great. I bought another one for my newer (but still old) computer and it caused all kinds of problems. Video repaint issues and crashes. The software indicated the card was working all the way up to the point of connecting to the access point and then it failed. I checked the box, and it says it works with Windows 98 SE, 2000, XP.... I have windows 98 on both machines (not SE). I suspect the older version of the drivers worked on plain 98 but Linksys decided to "upgrade" them at some point. We're talking network card here with a couple extra features (WEP key) that should be configured independant of the networking API. But it seems MS is as you say, retarded too. I thought I bought another one of the same product to work on the same OS, but I guess thet little "SE" makes a big difference in the Windows world. I'll try the old card in the newer machine and then perhaps a new card with old driver in the old machine. Feels just as "retarded" as Linux doesn't it?

Re:This is where Linux is retarded...

[Anonymous Coward]

Did you hate Win9x because crashes too much? Many of these crashes are because closed source drivers MS don't control. Did you really want to go this way?

Re:This is where Linux is retarded...

[be-fan]

Its not so much that, but the fact that maintaining a stable binary ABI is really hard, and limits what the kernel developers can do technically. He doesn't want binary comatibility to hold back kernel development.

Re:This is where Linux is retarded...

[quantum bit]

The WMP11? Be careful with that one. We bought one for work a while back and ended up using it in a FreeBSD router since it has a Prism2 chip. Since it worked out well, we then bought another one (same model number) a few months later. It turns out that they changed the chipset without changing the model number and that the new (and "improved") one has a Broadcom chip instead, for which there is no BSD driver...

Re:This is where Linux is retarded...

[Derek S]

These are two separate issues. I certainly do favor open sourcing drivers wherever possible, but that's no reason to create binary incompatibilities within a "stable" kernel series. It's hard enough keeping the growing pool of drivers up to date without randomly breaking them for no good reason. A semi-stable (obviously you'd have to allow changes within a development series) ABI would make development and deployment of drivers easier, while not hurting the real arguments for open-sourcing those drivers.

I'm reminded of the political situation surrounding transit in the San Francisco area. The bicycle and mass transit lobbies have it in their heads that anything that hurts car drivers must therefore be good for everyone else. So with the support of the city council they've managed to make a bad traffic situation even worse while failing to provide halfway decent public transportation or safe routes for bicyclists.

Re:This is where Linux is retarded...

[Anonymous Coward]

Thats odd.. didn't Linus actally say that Linux would support trusted computing? That goes against the spirit of the GPL..

Re:This is where Linux is retarded...

[Anonymous Coward]

The hardware does work. You just have to be not stupid and buy the right hardware. People who buy Nvidia hardware deserve what they get.

Oops, no drivers for the 2.6 kernel compiled for the Opteron (native, not 386 mode)? You lose. If you buy Nvidia, then you have to be a luddite and either use a 386 processor or old kernel or MS Windows. Wake up and smell the 21st century -- binary drivers are a stupid relic of the past.

Re:This is where Linux is retarded...

[Zork the Almighty]

Linus said that if someone wanted to implement DRM or "trusted computing" using GPL code, then he would not be averse to it. Yes, there is a subtle undertone to his position, but that's why he's in charge ;)

Re:This is where Linux is retarded...

[Lost Race]

You might get a printer to work with older drivers, but are you really serious in saying that 16 bit Win3.1 drivers worked satisfactorily for you in Windows?

I've done that, sort of. MediaVision ProGraphics 1024 video card (very high-performance in its day) had no Win 95 driver, so I just used the Win 3.1 driver in Win 95. Didn't seem like it ought to work, but it did. It was approximately as fast as in 3.1. (At least, not enough slower to be annoying.)

Re:This is where Linux is retarded...

[Anonymous Coward]

Linux developers can cry all they want to about "open" drivers, but there are plenty of times where that just isn't feasible.

The kernel developers don't give a fuck. Why should they? Inevitably if there is enough demand a free implementation will happen - and linux demand is steadily incresing. In the meantime why waste time on parasites?

Re:Linux really needs to make inroads in this area

[Anonymous Coward]

But there you go again? 'why Linux on the desktop will never be successful'. Successful to who? You? Are you using it? Do you care? Who cares? Where is the vested interest in Linux succeeding? For the good of mankind?

I just don't understand this shit attitude. The 'success' of Linux. Linux is where it is because it obviously works plenty well for so many users. If you want to advance it so it can meet your criteria for 'success', then fine, but stop your bitching everytime something doesn't work the way you think it should.

Calling me arrogant? Do you even do anything to make the situation better? Do you even USE linux? Idiot.

Plea for help here...

[The Master Control P]

Right now, I'd do anything just to GET the Nvidia drivers working at all. As it is, all their drivers have done is left me without any working OpenGL at all.

I've repeatedly tried using their drivers, version 4496. The first time I did so and started X, it worked a little like Windows Airlines. Everything looked good, until I tried to open the menu bar and the system locked up completely. Absolutely nothing could back out of X or get control back. Taking a look at the logs, it said "Unable to find NVidia Xfree-86 driver."

So after a reboot, I uninstalled their drivers and put "nv" instead of "nvidia" back into XF86Config-4. Ok, my desktop works again, except I have no OpenGL server: "Failed to add GLX extension (NVIDIA XFree86 driver not found)" WTF, I just unistalled their drivers!

Alright, I tried it one more time and this time told it to compile it's own kernel interface. Ok, it makes it's own kernel interface, the Windows plane takes off, reaches crusing altitude, and then waits 20 seconds before exploding: This time I tried opening an application, and MELTDOWN.

I've tried just generating a new X configuration file with xf86config, and that doesn't help. At the moment, I'm using the kludge: Replace "nvidia" with "nv" and be happy that it works at all. What can I do to either A) Get the Nvidia drivers working or B) Restore my original drivers? I don't do much gaming, but I DO work with 3-d programs and WINE, neither of which can even start without GLX. Please, help this pathetic Mandrake user!

Re:Plea for help here...

[DeathPenguin]

Did you remember to comment out the "Load dri" and "Load GLCore" lines in the modules section of your XF86Config-4 file and add "Load glx"?

Could you please post your XF86Config-4 file somewhere so we can have a look at it?

Re:Plea for help here...

[GuidoDEV]

It sounds to me like you may have the RenderAccel option enabled (Nvidia's docs call it "experimental"). Look in your XF86Config for it and make sure it is turned off as follows:

Option "RenderAccel" "off"

Re:Plea for help here...

[DAldredge]

What I had to do, at least under Fedora Core 1, was to remove the mesa-gl rpm.

run xdpyinfo and look for a line that says direct rendering: yes or no.

If it says yes it should work much faster, if it says no it is something else.

Re:Plea for help here...

[splint3r]

Hi,

First thing you do is look through the nvidia message boards . It's more than likely that you can find an answer there. If that fails then post a message there with proper details about your problem (machine specs, Xfree86 conf file et al.), the above description (and bless you for trying) is not adequate.

If they fail to give you an answer, which is unlikely, then I'll even take a look at things for you (I've had far more than my fair share of problems with an Nvidia card so I've had many hours of white hair enducing troubleshooting experience).

Re:Plea for help here...

[Little Hamster]

With "nv", you can't have load glx. And that's why you are getting "Failed to add GLX extension (NVIDIA XFree86 driver not found)".

With "nvidia", you need load glx, but you have to delete load dri and load glcore. Actually, I played around with it for a while and find that you can actually still load dri, but not glcore.

As for the kernel interface, if you are using a standard kernel from a major distribution, nvidia has precompiled ones. So it shouldn't reached that step unless you really know what you are doing (ie you've compiled your own kernel).

I agree with other poster that you have a weirdly configured XF86Config-4.

Re:Plea for help here...

[splint3r]

Ammendment: Actually the first thing you do is read the documentation. That's where I started troubleshooting (and okay no it didn't help me but you learn a lot of stuff about the drivers and what you can do with them).

Re:Plea for help here...

[gomoX]

Tippy tip:
When you install the nvidia drivers you have to actually *load* the driver (aka kernel module). No X until you do this. Everytime I install i forget to do this, so it might help for you:
# sh NVIDIA bolablabla
# modprobe nvidia
# exit
$ startx

Hope it helps

Re:Plea for help here...

[arkhan_jg]

the 'nv' driver is the opensource, reverse-engineered driver for nvidia cards that comes with XFree86, and has nothing to do with binary driver you got from nvidia. On gentoo, there is a program called /usr/sbin/opengl-update that switches between nvidia's opengl and xfree's, which you need to do in parallel with changing the 2D driver (i.e. nv or nvidia) in XF86Config-4, might be worth checking to see if such exists on mandrake. In order to use the binary driver, you need to use 'nvidia' in XF86Config, and make the other changes that people have mentioned in this thread.

I understand your POV,

[anti-NAT]

and it is the right type of thinking.

The question is though, who are they trying to protect their intellectual property from ?

For the moment, considering your example of door locks, I choose to use them, not because they make my house impenetrable, but they ensure that most, "casual" theives won't bother to break in, because the risk and / or effort is now higher than the reward. Determined thieves won't bother with trying to break the locks, they will just cut a hole in a wall, creating a new doorway. That is if my assets are worth the risk and effort involved in doing so.

So, who are Nvidia trying to protect their intellectual property from ? Who would gain the most from seeing it ? Individual end users, or their determined rivals like ATi, who have much more at stake, and possibly more to gain from discovering Nvidia's IP ? Assuming it is ATi or other competitors, which is what most people suggest when faced with this argument, then the "locks" that Nvidia have put in place are useless, as they will not stop a determined adversary, such as ATi, who may be willing to invest multiple $100K or $1M decoding Nvidia's drivers, using AGP bus analysers etc. The reward for ATi might be high, so the risk and / or effort involved in decoding the drivers may be worth it.

I really can't guess why Nvidia won't open source their drivers. However, I struggle to believe only reason I always hear - "to protect IP".

Re:I understand your POV,

[Osty]

So, who are Nvidia trying to protect their intellectual property from ? Who would gain the most from seeing it ? Individual end users, or their determined rivals like ATi, who have much more at stake, and possibly more to gain from discovering Nvidia's IP ? Assuming it is ATi or other competitors, which is what most people suggest when faced with this argument, then the "locks" that Nvidia have put in place are useless, as they will not stop a determined adversary, such as ATi, who may be willing to invest multiple $100K or $1M decoding Nvidia's drivers, using AGP bus analysers etc. The reward for ATi might be high, so the risk and / or effort involved in decoding the drivers may be worth it.

I can only believe that reverse engineering nVidia's work is a very expensive and risky proposition for ATI, and the time and expense would be better spent developing their own developers' skills. If on the other hand nVidia opened the source of their code, it would be trivially inexpensive for ATI to co-opt that code into their own drivers. As well, opening the code would require more dilligence and expense on nVidia's part making sure that if any competitors did use the code (even inadevertently, which is where you really have to be careful), those said competitors follow the open license properly. Providing the source code under a "look but don't touch" license would be useless, and still lead to the same problem as licensing under an OSS license, but with even more effort required by nVidia because anybody using the code would be doing so covertly.

In short, there doesn't seem to be any benefit for nVidia to open up the source code of their drivers (OSS mantras of "more eyes" aside), and there are what seem to me good reasons for them not to do so.

Re:I understand your POV,

[Minna Kirai]

The reward for ATi might be high, so the risk and / or effort involved in decoding the drivers may be worth it.

The marketplace for 3d videocards is a speed race, with both sides always trying to edge out the other in raw FPS and rendering features.

The major course of this race is the chipset hardware themselves... but driver releases form an important secondary struggle. Both companies try to repeatedly crank out better drivers to get ever-higher FPS, which can sway the high-end gamer (those customers who spend $400 on a graphics card).

And the thing is, better drivers really do make a difference in performance. On the Windows(r) side, where I suppose NVidia releases more frequently, one can really chart a rapidly climbing level of FPS. 15-20% improvements over the original driver. Driver quality makes a big difference. Having the best drivers matters to these companies. To a significant extent, vidcard companies are also software companies- just look at how tremendously better the NVidia linux driver is to the Free Software alternative.

And because drivers are quickly upgradable over the internet, the driver race is much more about speed than the (otherwise more important) chipset race is. A few weeks late on the release of the newest driver, and they'll feel it in the bottom line.

So, assuming ATI was willing to put in some heavy effort to reverse-engineer NVidia's driver's releases. From NVidia's perspective, that's still not as bad as if they were just reading Open Sourcecode, for 3 reasons.

1. By spending $100,000s on it, ATI is losing money
2. It'll take at least 10-20 times longer to understand a driver from binary as it would from source code. And that's for every new release. If that means it takes ATI 2 months to catch up with NVidia in the driver race, instead of 2 weeks, that's all the time they need to have a new and better driver on the way.
3. Legality. Reverse engineering may or may not be illegal, but it's certainly questionable. If ATI has a hard-core reverse-engineering project, they face the risk of a single upset (or principled) employee making a phone call and sticking them in expensive litigation. Whereas if the NVidia driver was Open Source, studying it for your own technical education would be absolutely legal.

Warning!

[Anonymous Coward]

Parent links to goatse. Just thought you'd like to know.

Re:Warning!

[Anonymous Coward]

You're so stupid, you wouldn't know goatse, if you saw it.

If you were and dumber, I'd swear you were a mac owner.

4496 Drivers Just Fine

[estergum]

Greetings,

For what its worth, I've always been happy with the Nvidia drivers.

So the 5328 drive doesn't work for me with ONE app, the fail back was efortless and I'm playing NWN again.

Kudos to the Nvidia team.

Cheers.

Re:4496 Drivers Just Fine

[DeathPenguin]

Me too. I've run nVidia drivers on everything from an Intel 440BX chipset (Abit BH6, BE6), a GeForce DDR on a KT133 chipset and an AMD 760 chipset (Asus A7V133, forgot the name of the 760 mainboard, it was an Asus), AMD 760-MP (Tyan Thunder K7), SiS735 (ECS K7S5A, SiS745 (ECS K7S6A), AMD 8151 (Tyan S2885), and a number of Dells and other systems at work that use various Intel chipsets. I've used almost all models of their cards in Linux since the TNT2, and haven't had any problems outside the learning curve for configuring X to work (Change three lines in the XF86Config-4 config file).

I know nobody's perfect, but I really do think a lot of people get way over-critical of nVidia at times. Slashdotters can bitch about anything.

Re:face it nVidia

[DataPath]

well, except for the API change from 95 to 95 w/USB, 95 to 98, 98 to 98SE, 98SE to ME.

Then there's the NT line.

The driver model changed slightly with each revision in the 9x line, and not necessarily in a core area.

And under linux, although the ABI may change from minor version to minor version, it IS just a simple compile against the different headers (for, I suspect, not all, but the majority of cases).

Really, the ideal thing for nVidia would be for THEM to publish an API (not necessarily full docs for the chips, just hooks into the binary), and for the hacker/enthusiast community to link it into the kernel. At least, that seems ideal to me.

For what it worth..

[Vilou]

We're actually packaging nvidia drivers for Fedora core 1 for the http://rpm.livna.org/ repository.
See : http://bugzilla.livna.org/show_bug.cgi?id=45

Feedback from the Clueful Ones is welcome.
Good day.

Buy a DRI supported video card

[anti-NAT]

Direct Rendering Open Source Project, it will save you a lot of grief.

Oh, go stuff it.

[Baloo Ursidae]

Manufacturers are (quite rightfully) only concerned with how well it works for most people out there. If it works alright in testbed (usually a Debian, sometimes a Red Hat-based-distro box in general, not sure what nVidia uses for theirs), it ships.

Manufacturers can't and shouldn't be held accountable for changes local to a specific distro or host. Deal with it.

Re:Oh, go stuff it.

[Anonymous Coward]

>> Manufacturers can't and shouldn't be held accountable for changes local to a specific distro or host. Deal with it.

Fine. And I will never again purchase an MB that has binary only graphics and network drivers.

That was my fuckup and it will not happen again. Fuck you Nvidia.

I have wasted far too much time dealing with Nvidia BS that would have been handled *much* better by the open source community. Instead of having a 'set it and forget it' graphics card, I have a built-in that is a tedious pain in the ass whenever I rebuild my kernel.

And for the better part of a year the Nvidia drivers would totally blank your screen upon exiting X11. So you'd have to reboot when you exited X. And they released that crap.

No problem with quake3

[Jayanef]

upgrading to new version
PIII/733 with cheap nvidia 8MB, mandrake 9.2, my own kernel compiled with Con Kolivas patchset

forcedeth driver suggestion

[LinuxGeek]

Others have suggested this driver, and it does now work on my system ( 2.6.0 kernel, NF7-S, GeforceFX 5900NU). It does not work on my system as a module, only when compiled into the kernel. Hopefully this will save someone the few hours I burned finding this out.

Great job on reverse engineering the nvidia binary ethernet driver though. It isn't as fast as the nvidia driver, but is getting better.

Re:This is how I solved the problem

[fruey]

Interesting. You're missing the point.

Do you work for Microsoft?

That's why they have patents-Eyes wired open.

[Anonymous Coward]

"If [Linux] opened the source, then any [Windows] developer could browse it, and inadvertently (or advertently, even) insert [Linux] code into [Windows]"

Well I don't know about you, but I rekin' we better close up our code quick before the likes of Microsoft or SCO get a hold of it, and we lose our competitive advantage.

Re:That's why they have patents-Eyes wired open.

[Osty]

Well I don't know about you, but I rekin' we better close up our code quick before the likes of Microsoft or SCO get a hold of it, and we lose our competitive advantage.

What you fail to realize is that nVidia is selling product which is enhanced by the drivers and the IP contained therein. Driver enhancements are how nVidia keeps their older card lines still attractive in the mid- and low-level markets. An original GeForce256 is still quite capable today thanks to enhancements in drivers, and many of those enhancements support even the ancient TNT and TNT2 chipsets. Because of this, you can still buy TNT2 chipsets on the market and actually get a useful card.

On the other hand, Linux is not in the business of selling product. There are companies making money off of Linux, but none are thriving solely from selling Linux itself, and thus there is no benefit in keeping the source closed. You're right that people could very well take code and insert it into Windows or Windows drivers and applications, and I'm sure it's been done before, but the main difference here is that there's something people (nVidia competitors) want in nVidia's driver code, and it's in nVidia's best interests to make sure that they can't get it without cost.

When you're in a market as competitive as the GPU industry, you hang on to every scrap that could possibly give you an edge on your competitors and cause people to buy your product over others. nVidia has this with their driver technology, on which strength alone many nVidia cards have been sold (even when ATI makes technically superior cards, nVidia's driver prowess often wins the day). The same can't be said for drivers in the Linux kernel for 10 year old SCSI cards, for example.

Re:This is how I solved the problem

[Soko]

Actually, I did that - got Nvidia drivers runnning on XP in under 4 hours. That was for work.

I then installed Fedora Core 1 on another partition, got it all running and happy, and then downloaded the 2.6.0 kernel source and went to town. NVidia drivers and all. It was a riot.

Some things in life are needlessly fucking complicated and obscure - and a hell of a lot fun figuring out how to make them work. For everything else there's Windows.

Soko

Damnit...

[torpor]

I've been trying for a week to get the stupid nFORCE drivers working so I can get my new Shuttle PC on the 'net with Linux, but it's just not working.

Why, oh why, oh why can't they just release these drivers in 'normal' format?

Re:Damnit...

[The One KEA]

You don't need the nForce drivers anymore to make the nVidia Ethernet driver work properly; the 2.6.0 kernel, and IIRC the 2.4.23 kernel, include a reverse-engineered driver called forcedeth for the nVidia Ethernet controller on the nForce, nForce2 and nForce3 chipsets.

The latest patches can be found here

Re:Damnit...

[torpor]

Thank you, you rock ... I figured there might be something like this around, but yeah: couldn't get on the 'net in the environment I'm working in, in order to do a proper search... so, ta!

Re:Damnit...

[Thurn und Taxis]

Try turning off ACPI support in the kernel. With it on, my system would lock up every ten minutes, or whenever I generated a lot of IDE traffic (whichever came first). With it off, my system's rock-solid again.

Nvidia AGP + Fedora = Crash

[noselasd]

Just FYI, the nvidia drivers contain their own AGP driver,
it can be used by e.g. setting the XF86config option NvAGP=1.
This reliably oopses the kernel in Fedora with this new driver.
If anyone else wonder why the new driver don't work, make sure it uses
the kernel AGP driver, not the nvidia one.

Re:Nvidia AGP + Fedora = Crash

[Mr Smidge]

You have to compile your kernel *without* agpgart support if you want to use NvAGP=1 in your XF86Config.

I currently have that option enabled in my configuration, using the older drivers (note: not these ones), driving two monitors from one Ti4200 incidentally, and it has been particularly stable.

Re:Nvidia AGP + Fedora = Crash

[noselasd]

I currently havn't compiled the kernel.
Using the stock 2.4.22-1.2129.nptl from Fedora.

Oh, go stuff it.-Stuffed Bears.

[Anonymous Coward]

"Manufacturers can't and shouldn't be held accountable for changes local to a specific distro or host. Deal with it."

I've found that returning it, and getting your money back, then buying the competition is a good way of "Dealing with it".

Re:This is how I solved the problem

[Anonymous Coward]

Well have fun figuring out how to fix those BSOD's when they happen.

Oh ya when you DON'T figure it out because nobody CAN exept MS and they don't give a shit because it gives you a reason to buy their next OS.

You get to REINSTALL!

Backup that data, lose those apps. Then spend another goddamn 4-8 hours of your life installing all that shit over and over again until by chance it works well enough.

Hope you enjoy spending your 310 dollars, because in another year or 2 you get to do it all over! And again and again and again. Not to mention the thousands of dollars you get to spend on crappy software that's inferior to the shit I get for free.

Rambling [failed economics] Post:[ter]

[Anonymous Coward]

"I don't know about anyone else, but I think this is a *total* waste of money. "

And if that person hadn't "totally wasted" their money, you wouldn't have a $99 card to feel superior about.

Re:I've heard a lot of problems with both companie

[manjunaths]

nForce and Radeon, you mean like this ?
dmesg and XFree86 output

agpgart: Maximum main memory to use for agp memory: 203M
agpgart: Detected NVIDIA nForce chipset
agpgart: AGP aperture is 128M @ 0xf0000000

and

Loading /usr/X11R6/lib/modules/drivers/fglrx_drv.o
(II) Module fglrx: vendor="Fire GL - ATI Research GmbH, Germany" ...
(II) fglrx(0): DRI initialization successfull!

I don't know about you, but I have mine working perfectly. I get 1662 FPS on glxgears. You need to read the manual and try again or ask in #ati on irc.freenode.net.
I had never used an ATI until like a 3 months ago when I bought this card. Well I am glad I did and I paid a LOT less for it too than a comparable nvidia card. I run redhat 9, with kernel 2.4.20-8custom, drivers work perfectly for me.
Oh BTW the motherboard is an Asus A7N-266 VM, in case you are interested.

Testing?

[Doc Ruby]

Don't developers spend money on any testing anymore? Every goddamn program I get these days shows that it was obviously released without any testing feedback. I know Netscape ruined the software biz model by releasing "betas" for consumption, marketing the product before it was tested. And Microsoft has never gotten anything right before version 3.0. But with malware liability looming, when will some kind of quality control come back?

Re:Testing?

[Gothmolly]

And MS stopped getting things right just after 3.11

Re:Testing?

[Deusy]

Oh come on. We've been in 'beta' mode for a very long time. Since Windows was first released back in the 80s. Everything (server-stuff excluded) since then has been of beta quality when compared to the equivalents from other industries.

Imagine how unreliable cars would be if they had the same relaxed release criteria as software?

Cool!

[Trejkaz]

So I guess that makes Gentoo one of the most popular distros.

Wake me up...

[Trejkaz]

... when the ONLY game worth playing on Linux works with these new drivers. Kthxbye.

Re:Wake me up...

[Anonymous Coward]

... when the ONLY game worth playing on Linux works with these new drivers.

It's funny, I have had tons of fun playing Heroes of Might & Magic 3, Jagged Alliance 2 and Alpha Centauri on Linux. It's true that they are not as new and flashy as NWN (bought them about 2-3 years ago, anyway), but playability is just fine for me. Haven't tested the new driver, though, so I can't say 100% surely that they work with it. ;)

Re:face it nVidia

[jeremyp]

Really, the ideal thing for nVidia would be for THEM to publish an API (not necessarily full docs for the chips, just hooks into the binary), and for the hacker/enthusiast community to link it into the kernel. At least, that seems ideal to me

They do publish the source code to the bit which talks to the Linux kernel. They've effectively written their own standard ABI and published it.

Re:This is how I solved the problem

[Anonymous Coward]

what a drama queen. as an mcse and xp user, i have to say...your such a little fag.

haven't lived at home since 17.

or perhaps your "projecting"

anyway...i dare you to respond, and give me a way to contact you. then we can arrange a friendly parlay.

cause i'll show you my mcse card, then i'm gonna beat your ass.

Re:face it nVidia

[minasoko]

Let's give nVidia a break.

I'm with you. Recently, as a complete 'nix beginner, I downloaded Fedora and set it up on an FX5600 with 4496 drivers.

They broke Enemy Territory and some other OGL apps, but 15 minutes, google and the Unofficial Fedora FAQ got things working perfectly.

Now I've got sideband addressing and fast writes working too, performance is a more than a match for Windows. I get no lockups at all and I don't consider 15 minutes of my time too much to ask.

I'm just wondering if most of these people with stability problems have looked into all the FAQs and message boards, like this one linked from nVidia's own driver page.

Re:face it nVidia

[RupW]

well, except for the API change from 95 to 95 w/USB, 95 to 98, 98 to 98SE, 98SE to ME.

I don't know the video interface well, but the networking changes are simple, backward compatible evolutions of the old interface.

Around 98, ME and 2K the driver interfaces for the 9x line and NT line converged. Unfortunately, the kernel interfaces for drivers diverged slightly (e.g. paged memory allocation) - you could use ME drivers on 2k but the kernel spits warnings at you when you fire it up in debug mode.

Something to think about

[Sprinkels]

A lot of people may think Linux is unreliable, because of the instability of binary drivers

A lot of people may think Windows is unreliable, because of the instability of binary drivers

Unified drivers is a good thing

[Sprinkels]

You are implying that milions of different incompatible drivers for every type, mark, model, and revision of hardware makes the world a better place.

Honestly, I can't see how this philosophy saves time.

Actually, I believe that hardware should be engineerd in a way that it will work with generic drivers. This process is called standardisation.

NVidia is showing leadership with this development model.

Re:This is how I solved the problem

[Shinobi]

You need to measure the amount of time it takes to get Nvidia drivers to work under Windows in hours?!?

In 4 hours, you should have installed XP, all the necessary drivers, done the Windows Update thing and setup many of your security policies with the group policy editor..... And those are quite interesting to learn about.... Way too few Linux people do that...

Nevermind :-D

[0x1337]

Applied the minion.de patch again, installed, rebooted - all works... +10 fps increase in UT2003 oh yeah

Hahahahahaha

[Rogerborg]

Ahahahahahaaaaaa haaaaaaaaaaaaaa hahahahaha, ah hahahahaaa, ah haaa, ah haa, ah, ah.

OMFG, you guys are funny. You like the pain. Admit it. You live for this stuff.

Re:Hahahahahaha

[Tin Foil Hat]

Of course we do. What's so funny?

Re:I've heard a lot of problems with both companie

[Anonymous Coward]

I get 1662 FPS on glxgears.

Ah, the other Linux game!

Why not just boycott Nvidia?

[Pope Raymond Lama]

Instead of loosing time and money fighting proprietary software that does not even allow one to give proper feedback to the developers?

One with reverse engineered drivers is far better of, since blocker bugs are resolved in a matter of days by any Free Software developer group.

What are they so afraid of that they do not put these drivers under a Free SOftware License? That people will have copies of thge software without paying them? AFAIK everyone that is interested in such a driver is already a card owner and won't pay anyway.

Keeping these as proprietary is "The Wrong Thing To Do"(tm) under any thinkable point of view.

I keep said for those among the comunity that think it is allright that the drivers are proprietary as far as they exist. Actually it is far worse - in fact proprietary drivers are as incompatible with the Linux Kernel as any Windows only driver. The difference is that instead of a technical incompatibility there is a license incompatibility.

Re:Why not just boycott Nvidia?

[Duty]

Please find this mysterious "any Free Software developer group" and ask them to fix, say, the various sprite display bugs in the DRI Project's OSS r200 driver that have made Aleph One unplayable for somewhat longer then a matter of days.

Not to troll pointlessly, but open source is not a panacea.

Re:Why not just boycott Nvidia?

[TeknoHog]

I agree very much, but how many alternatives do we have? For example, what is the Linux driver situation with latest ATI cards?

Re:Why not just boycott Nvidia?

[Pope Raymond Lama]

Well, since you have posted it here, you could as well post add ons to the bug report on the appropriate place, and write the developers for it.

You could also pick up the U$50,00 you would pay for any other proprietary software and pay someone to check the bugs for you, and at least see if it is feasible in the short term, if not fix it.

Understand that when I caln for Free Software Nvidia drivers, I mean that they could keep the same people taht work opn the drivers today, and just make them free. That simply would not hurt anyone but, possibly, the egos of said programers which might be afraid to expose possibly bad written code to the world.

Re:Why not just boycott Nvidia?

[Anonymous Coward]

alternatives

Matrox G400 or G450.

Re:Why not just boycott Nvidia?

[DeathPenguin]

You're kidding, right? Those cards are on part with the GeForce SDR.

Re:Why not just boycott Nvidia?

[DeathPenguin]

>>For example, what is the Linux driver situation with latest ATI cards?

I haven't tried the X driver in a long time, so I won't comment on that. However, ATi has released their own proprietary driver which I've heard many good things about.

robbIE releases gnu secrets to online dating?

[Anonymous Coward]

you can bet your .asp you won't be 'dating' the won-eyed girl in the jump-you ads, as rumour persists that she is not real, &/or NOT a paid up member of robbIE's dating service.

so, the secret, if there is won, may or may not be revealed in the upcoming questions for va lairIE/robbIE interview.

keep it simple. don't ask about the disposition of the truckloads of phonIE monIE. the top questions will be filtered buy phonIE ?pr? ?firm? talknicians, & high dollar shysters, so that the 'answers' reflect almost nothing to do with stuff that matters(tm). it's a pateNTdead funkshun of the SourceForgerIE(tm).

mynuts won, you need a date buddIE?

Re:Nvidia's Detonators are designed to force upgra

[beezly]

You need to read up on the way memory usage gets reported. All those pixmaps that are being displayed on your screen are reported in the memory usage for *each* application that is using them. Same for shared libraries, same for any shared memory that might be being used.

A good example is XMMS. When you run it, it appears to have a number of processes each using a substantial chunk of memory, but the memory is shared between all the processes, and so total memory usage for XMMS is only just over that of one of the processes.

Drivers work great!

[SQLz]

Maybe its just the Fedora people having problems but the drivers work for me no problem. NWN, Quake3 Urban Terror, UT2003 all worked fine.

NForce2/Ti4200 Gentoo 1.4 (-march=athlon-xp -O3 -pipe -fomit-frame-pointer) with Kernel 2.6.0.

no open source portion

[SQLz]

There is no open source portion of the nVidia driver. The kernel module is closed source as well. The stuff you see compiling is just a wrapper so their are not module versioning issues.

Works fine for me

[trollhead]

My system: P3, GeForce FX5600, Fedora1, kernel 2.6.0 (stock), NV-AGP driver

Native Games: NWN, Railroad Tycoon2 work like before

WineX3 Games: Warcraft3, Starcraft work like before, Railroad Tycoon 3 Demo seems to work a little bit better (movies hang less)

Damn slashcode

[Chep]

"Bought for less than EUR 40"

WHAT IS WRONG WITH YOU MODS,THIS IS DISINFORMATION

[latroM]

Check the facts, do the moderating after that.

Won't people ever learn?

[Anonymous Coward]

I find it very sad that so many people are wasting their time with these drivers.

If it's not open source, it's nonsense.

Closed-source drivers are just not going to be satisfactory with GNU/Linux. Get a clue, people. Just don't buy hardware if the manufacturer won't tell anyone how the hardware works.

Works for me

[T'lexii]

Commiserations to all having problems, but mine installed over 4496 perfectly and are as fast or faster in nwn and q3. glxgears shows 7100 fps fwiw.

I see no evidence of 2d slowdown either.

RH9
normal 2.4.23
Athlon 2800
FX5900

Be thankful...

[kuzb]

...that nvidia does anything for linux at all. Seriously, there is no real game market for linux, and these cards are designed for gamers. You don't see people running out to buy the latest and greatest badass video card to run their office applications.

I'm really suprised that they spend any amount of time trying to cater to people who in most cases could just use some grossly outdated 8mb card to do what they'd normally do. Linux is not a gaming OS, it never has been, it may never be.

This is a real chicken and the egg thing; in order for game makers (and subsequently, video card manufacturers who are geared towards gamers) to take linux seriously, it needs the home userbase. However, before linux has a home userbase worth noting, it needs to be usable to the average home user. Since the vast majority of home users want games, they won't use linux, and you're all the way back to the beginning. And please, don't say "but you can use winex!" this process is hit and miss, and is often more aggrevation than it's worth. Only die hard geeks would have the tolerance for that.

Only with linux is it some kind of major event when a worthwhile game gets released for it. If you love games, but love linux too, you should dual boot linux/windows, it's the only sane solution to the problem. Anything else is going to waste your time that you could be using to kill another bad guy in your favorite game :)

Re:Be thankful...

[m1chael]

I know this is a troll, seeking attention perchance. Having a nvidia card or any such brand, does necessarily mean you must play games or die. Keeping all your bases covered is a good strategy, Trollman kuzb.

ATI 9600 success here!

[CrackedButter]

Plus i'm running it on top of Darwin, Version 7.2.0.
I'm running what seems to be the benchmark game, NWN for hours at a time, i even got Halo installed and playing it with max resolutions, my 10/100/1000 network card is working with no hiccups either, no conflicts.
This is all running under Aqua as well. Sometimes i wonder what you linux guys are getting all this grief for.

Re:This is how I solved the problem

[Anonymous Coward]

You can't even do 1/5th of what he can do right now. I hope you enjoy not being able to use your computer for more than talking to losers on IRC.

kernel driver is binary

[koekepeer]

the kernel driver is closed source, as well as the X drivers (GL stuff)

what you compile is basically a wrapper around the kernel driver.

Obvious Troll

[Gothmolly]

Shame on all you who rose to it!

Problem: it not their code to give

[CarrionBird]

Are you suggesting that they violate thier licenses and agreements by open sourceing software that they don't have the legal right to do? (portions of the drivers are licensed from others)

And even if they could legally realease the code, it would not be in thier best interest to do so. PC hardware is a dirty business and any one of thier competitors would outright copy their best stuff in a heartbeat if they can get away with it.

FYI

[SirNAOF]

For those who want a patch...

http://minion.de/nvidia.html was updated a matter of moments ago.

Go grab the latest patch now!

Secrecy is wasted

[A nonymous Coward]

I worked in a similar industry, communications cards. We also had secret sauce. But so did our competitors, and let me assure you that we disassembled our competitors software as soon as we could get our hands on it. We didn't disassemble chips themselves, but no need, we knew what they did in software.

I am sure every graphics company disassembles their competitors software. Are they each so arrogant as to think their competitors don't?

Re:Secrecy is wasted

[Minna Kirai]

and let me assure you that we disassembled our competitors software as soon as we could get our hands on it

So why didn't your company release the source, then? Were you so arrogant as to think your competitors didn't disassemble it?

Or maybe, you knew they did, but didn't want to make it easy.

Are they each so arrogant as to think their competitors don't?

No, they probably just think that to understand a program from disassembly is more than ten times as slow as if you had the source code. The race to have the best video drivers is time sensitive. A week or two of delay can mean a lot.

They worked for me without problems

[wolverine1999]

The new driver worked for me without a problem.

Thanks to NVIDIA.

I even don't use a normal distribution but a custom built linux version..

OT

[sLaSh_N_bUrN_(.Y.)]

Why did you disable HT? Not a troll, I don't know.

Re:OT

[kikensei]

Because the Savage linux client binaries have trouble with SMP systems, and HT is treated as an SMP setup. Just 'cause of that one title. Not related to Nvidia.

The 'reality'

[Junta]

Is that drivers are *still* a huge deal, but for different reasons. Given equal hardware, well-written drivers that work only with that hardware, doing things in the best possible way with the hardware in mind can make worlds of difference. On top of that, companies often enter restrictive license agreements with other companies to get more optimized code into their drivers, so even if the company was willing to compromise the effort they put into the driver for the good of the linux community, the license agreements prevent them.

Of course, all this seems to sidestep your issue, hardware specs, which in theory hold no secrets (interface shouldn't reveal much about implementation), could be released without giving out the critical secrets which make their drivers so good. In this respect, I would say first that theory does not hold, the interface to the hardware functionality can give extremly enlightening insight into implementation details. Seeing how the data is expected to be passed into the hardware gives a good idea as to the hardware designer's strategy in their implementation. Even assuming the interface itself meets these theoretical requirements, there likely exists no interface specification without extensive comments and notes that do reveal secrets, and cleaning up a spec would take time and money.

Finally, even if all the moons were aligned for a spec release to allow the community to develop a driver around it, it still may not be desirable to the company, for it dilutes the image of their hardware. For example, take the current nVidia situation. The open source driver only has 2D support, so no one gives it a second thought in terms of evaluation. Now imagine nVidia releases spec and the open source nVidia driver takes those specs along with the work for the current ATI offering and implements a driver that would be extremely similar to the ATI software wise. Then presume that the nVidia offering with this driver looks much much poorer than the ATI offering. Reviewers would pounce on this driver to show the hardware on a 'level playing field' to compare the *hardware* independent of software. If the magic of a piece of hardware consists mainly in the free software piece rather than the card, it feels less valuable. On top of that, some hardcore OSS purist reviewers would ignore nVidia drivers now that there was an alternative, degrading nVidia benchmark performance. In addition, if there is a large discrepency between the open source and the binary-only driver, there would be a lot of suspicion cast on the 'optimizations' of the binary driver. What shortcuts must they be taking to perform so well? In your terms, we no longer value simple compatibility, we look at features with the expectation that a hardware company implements them all in *hardware*, and if that is not true, it is worth something to keep the public from knowing.

Ultimately, I wish things were as simple as you say, and I favor open source drivers when available. The reality is that between complex business relations, desire to keep funded driver developer secrets from the hands of competitors, and what the public could do with a more acceptable alternative, the reality is that binary drivers are still much more preferable to hardware vendors.

What's the word on Suse?

[pair-a-noyd]

Anyone had any luck getting the new driver installed on Suse 9.0 Pro with either the 2.4.x or the 2.6.0 kernel??

If so, how did you go about it?

I'm using the 4496 drivers with 2.4.21-144-default and having no problems.

I want to upgrade to the 2.6.0 kernel but am waiting awhile to see what plays out. I, being new to Suse (but not to Linux) am not yet wanting to play guinea pig...

Me Too

[DanielJH]

I hate me too post, but in this case NVidia seems to always get the bad end of the stick. I have been running NVidia drivers for close to 2 years across an array of machines. I have only had three issues in all that time that didn't come down to someone not reading the README file, and one of them is AMD's fault for not disclosing a chip bug!

GREAT JOB NVIDIA! Keep up the good work!

No problem with nvidia and FC1 for me...

[kosmosik]

I spent a few long hours recently trying to get the old drivers to work with Fedora Core 1 so I'm going to hold off on these new ones.

Actualy HOW? These drivers and installer are *very* well made and documented. I don't see any potential problems... Only one - FC1 uses different compiler (gcc32) for kernel and modules and you'll need to use gcc32 (export CC=gcc32 before launching the installer). Thats it! Of course you also need to have kernel source package (as stated in Fine Documentation) but trying to install kernel-source rpm will tell you to install gcc32 (and addons) - so what is the problem here? When you try to compile with wrong compiler. You'll get error message stating that the compiler version is wrong... I really don't understand what can be the problem with installing nvidia kernel module... Especialy for Slashdot editor. :-PP

Oh and if you REALLY don't want to compile anything, just grab (if you trust them, works fine for me) RPMS from here:

http://www2.educ.umu.se/~peter/nvidia/

Then why...

[Anonymous Coward]

is my STB TV Pro card just filling a slot? I put the coaster in the cup holder and nothing. Proprietary software for hardware can render the hardware obsolete.

Knoppix could find it.

Terrible realease

[Anonymous Coward]

These drivers are horrible, I am running FC1 right now and I thought I would give it a whirl. Loaded them and started X back, things looked fine, so I ran a OpenGL app to test it out, no matter HOW simple the app was.. I would get like 1 FPS and CPU usage up at 100% that slowed my entire system down. So I was like, well screw that I will just go back to the old version, BUT now the damned nvidia module will NOT unload. I got out of X and all that, but once it was loaded, it would NOT remove.... ever. I had to reboot the system AND change my XFConfig file so that the module wouldn't load at all in the first place. Just to make sure it wasn't a fluke I tried it again, this time compiling the kernel module myself. Worked just as crappy and again the module wouldn't unload. So after much hassle I finally went back to the older ones and my FPS are normal again. These new drivers obviously weren't tested worth a damn if this many people are having problems!

Works here

[be-fan]

The new drivers work just fine on my system (like every other NVIDIA driver I've tried). I'm getting about a 10% performance improvement across the board. My specs:

Debian sid
Kernel 2.4.22
GeForce4Go 440 (NV17)
Pentium 4 2.0
i845 mobile chipset

Stop thinking like a single home user.

[DanielJH]

Unified drivers are a good things for corporations.

1. It means that they only need a single driver for the entire corporation. Infact you could probably write a simple script to upgrade everyone.

2. One file to download, no mess, no fuss, no long list of cards in corporate machines.

3. Quickly replaced advanced graphics card with yesterday's model in case of a problem. This has saved me significant amounts of time and money. If there wasn't a Unified driver I would need to reinstall the drivers, each time or keep high powered ($$) graphics cards laying around.

WARNING Generial rant:
OSS developers are increasingly locked into the one person, one computer mindset. It is really very Microsoft of you (as it is the way that company thinks). Real power users use hetrogenous systems with multiple distribution and propritary UNIX all using the same home directories and dot files. It is great when a system works in this environment.

driver source for old graphics cards

[r5t8i6y3]

another way to approach manifesting open source graphics card drivers would be to request ATI, NVIDIA, VIA., etc. to release the source for the drivers for their old graphics cards. let the companies decide what an "old" graphics card is.

i'm imagining many open source folks would gladly put energy into stabilizing some older non-bleeding edge graphics cards. if this proved popular a new competition might emerge between graphics cards manufacturers to be the first to release the source for hardware that isn't bleeding edge.

Re:driver source for old graphics cards

[Anonymous Coward]

Yes, but that doesn't work with the unified driver model nVidia use. There's one driver that works with pretty much anything back to the TNT family... no separate driver for the old cards.

Problem: it not their code to give-Secret santa.

[Anonymous Coward]

"And even if they could legally realease the code, it would not be in thier best interest to do so. PC hardware is a dirty business and any one of thier competitors would outright copy their best stuff in a heartbeat if they can get away with it."

That's why I advocate SCO. The OS business is a dirty business, and any of their competitors (Linux) could grab their best stuff in a heartbeat. So it's best that SCO keeps mum about the code.

fixed apm suspend to ram for me.

[Anonymous Coward]

well that says it. latitude c810 w/geforce2go, debian unstable, 2.4.23 w/pac1. Funny because I got the laptop a couple days ago, and sat down yesterday to get it going. only thing I had trouble with was s2r. checked nvidia forum, and blammo, a new driver annoucement. installed, s2r works.

waah! waah!

[j_d]

and if they really loved you then they'd buy you a pony!

And you people say Linux is ready for the desktop?

[Overly Critical Guy]

The submitter even mentioned that it took him HOURS to install video drivers. Uh...excuse me?

Forgive me for wanting to spend my time using my computer, not spending hours setting it up to use it.

Re:Nvidia's Detonators are designed to force upgra

[Anonymous Coward]

Yea, but still. 120megs for a browser? I remember thinking Netscape was bloat at 30 megs. A simple window manager is under a couple megs.

Not quite correct

[autechre]

Debian doesn't differentiate between the BSD and GPL licenses, as they are both DFSG-compliant and thus able to be included in the "main" (Free) section.

What you have are main, non-us, and non-free. Packages in main can be used by everyone and will always comply with the DFSG (Debian Free Software Guidelines). They will also not depend on non-free packages.

Packages in non-us might not be OK for people to use in the US, or might have to be developed outside of the US, for various legal reasons. OpenSSH used to be in there.

non-free is for, well, packages that are not Free-as-in-RMS. Modules to allow the GIMP to process GIFs were in there (not sure if they still are). Netscape 4 is in there.

Debian also has a package which will download, compile, and install the nVidia drivers. They have something similar for the MS TTF fonts, PINE, and some of Dan Bernstein's software. RealPlayer too, IIRC.

I think that this is all very good. It provides a clear distinction, and you will never be in trouble (SCO aside) if you only use packages from main. If this is your home desktop and you don't care, simply tell the installer to include all of these APT sources, and you won't have to think about it again.

Still no powerpc drivers it seems...

[An Anonymous Hero]

*Sigh*... still nothing for the GeForce in 12" PowerBooks.

Come on nVidia, smell the coffee!

I don't understand

[autechre]

Which nVidia drivers are the rest of you using? Did I fall out of some parallel universe yesterday?

Maybe it's the fact that I'm using Debian instead of Fedora, but getting the nVidia drivers working didn't even take me half an hour, and that was before their easy installer. Downloaded both the files, untarred, read the directions, and typed make in each directory.

And that's what I've done every time a new driver is released. I have simply not had any problems. Yes, acceleration is working perfectly. So I ask you...what are the rest of you doing?

I'm not some guru. I'm pretty good at Linux, good enough to be a good sysadmin, but I hadn't even heard of it before 1998. Most of my programs are PHP, and SQL (PostgreSQL if I can) or shell scripts. I spend much of my free time doing non-computer things. I don't understand what the problem is here.

Re:This is how I solved the problem

[moncyb]

So you are saying everyone should fuck up their computer with one of the shittiest operating systems ever made?

Total time spent: just under four hours.

Until tomorrow. Then you'll have to spend another four hours trying to get your system working again because 3 lines of M$ code says "you don't want these settings, let me fix them for you." or "this driver wasn't signed by Microsoft, I'll just delete it."

It figures it'd be now

[Chuk]

I just finished getting the old ones working last weekend...

Nvidia Sucks

[Anonymous Coward]

Nvidia Sucks.
- ATI Radeon Worshipper.

Re:face it nVidia

[DataPath]

what I gather from some acquaintances who do driver development, didn't so much converge as they did confuse. The new driver model is a spaghetti mess that allowed them to make minimal changes to the 9x driver model, and minimal changes to the NT driver model so a driver would work on both, but not without creating a pretty big mess, first. But let me repeat - this is what I understand from talking to some developers, this is not personal knowledge.

Now that think about it some more...

[CarrionBird]

It may be in their interest to open what they can of the drivers (may be hard to separate without breaking things though). They would get the benefit of OSS, but even better if someone does copy thier code it could work for their advantage.

NV would probably have some black box people checking every competitors driver revision. As soon as the code turns up elsewhere Nvidia can engage them in a nice big IP case and bolster their standing while putting a strain on the infringers. (insert evil laugh)

All kidding aside, secrecy has been ingrained in that business for so long that it will take some time for OSS to really get worked into the mindset. This wouldn't be such a problem if all the players would act a little more ethically, maybe some of the new entries to the market can set an example.

No

[Enucite]

Don't try to pass off ATi's cobbled-together drivers as a myth. The only myth is ATi's "Linux support". I bought a Radeon 9500 after hearing about ATi's "support". After using it in Linux I gave it to my girlfriend (since she still uses Windows) and went back to my Ti4200. Yeah, it's that bad.

I want Open Source to succeed more than most people; I only buy hardware from companies that have Linux support. Putting out sub-standard drivers that make brand new cards perform worse than the last-generation cards isn't support in my book. Nvidia is releasing drivers that are at least as good as their Windows equivalent. To me, that's "Linux support".

nVidia is the only choice right now for high performance 3d in Linux. There's no way you can convince me it's better to spend over $300 to get an ATi card with 3D performance on the same level as a $50 nVidia card.

ATi has superior hardware, but their drivers are holding them back. Let me know when they have Open Source drivers that are anywhere close to the performance of their Windows drivers and I'll be the first one to switch. Until then I'll stick with the company that actually supports Linux instead of just using it for PR.

Open SOURCE?!

[hubertf]

It would be nice to have the source for this, so that users of other Open Source operating system can make use of their cards too...

- Hubert

Re:And you people say Linux is ready for the deskt

[Anonymous Coward]

The submitter even mentioned that it took him HOURS to install video drivers. Uh...excuse me?

Forgive me for wanting to spend my time using my computer, not spending hours setting it up to use it.

YMMV.

Re:Linux really needs to make inroads in this area

[Anonymous Coward]

You're a trolling piece of shit. That is all.

New Nvidia Drivers Works 100% for me

[Anonymous Coward]

I just installed the driver and I have no problems. It would seem that it's very dependant on the chipset on people's motherboards.

Thanks Nvidia....!!!

Re:And you people say Linux is ready for the deskt

[Anonymous Coward]

Hey Fuckie, The driver infrastructure is there for anyone who wants to use it. Don't blame linux for the faults of NVidia and their fucked up install method. They could have easily created a download that you just double click on to install, but they did not. Can I blame microsoft for having to jump through hopes to install a xerox printer? No? Okay, then let me serve you up a nice tall cool glass of shut the hell up.

Re:REPORT YOUR EXPERIENCES BELOW

[Anonymous Coward]

I suspect the real reason cheat protection broke was Linux's automatic 'gay detector' turned on as soon as the cheat protection loaded.

Re:Once again, Linux displays it's flaws.

[Anonymous Coward]

Interesting.

I have a Gainward card as well (a GeForce2MX) and while it works flawlessy with the default drivers as supplied during the installation of either Win2K or WinXP. However, installing one of either:

i) Drivers from "Windows Update" (Win2K)
ii) Gainward's drivers (Win2K or WinXP)
iii) nVidia's drivers (Win2K or WinXP)

will result in my graphics subsystem being horribly broken (with the Gainward drivers being the worst - lol!) - although out of those 3 choices above, nVidia's drivers appear to be the "best" ones.

Go figure eh?

Because I wasn't the boss?

[A nonymous Coward]

Keerist in a bucket buddy, what makes you think I had any say in the matter?

Ahem..

[cybrthng]

Windows drivers have worked fine for years.

What is so difficult about drivers on linux/X11? Is it THAT bad?

Re:This is how I solved the problem

[The Bungi]

ooohh, and he made me a foe. how fucking quaint. and "M$"? hahahaha!! brilliant.

Re:REPORT YOUR EXPERIENCES BELOW

[0x1337]

Not sure what.. you are... talking about... But VALVe patched the servers to not recognise Wine/WineX as a hack - lol. I think there was a patch for WIne floating around before as well... meh.

incompitent boobs

[Anonymous Coward]

wow i cant beleive all the crying and suffering going on over this.... if you cannot get these drivers to work, then either you should use another OS that does everything for you, read some docs, or use another video card

its so easy to use the installer, or rpms, or even install from the old tar.gz files they used to provide my 10 year old son does it... whats your problem??? :P

Hoser!

[pigscanfly.ca]

Theres actually a reason why .
Linus doesnt like binary only drivers , makeing it easier for people to make binary only dirvers would not be a very good idea .By makeing in damm hard to have make binary only drivers companies that might have gone binary only "just to be safe" go the OO route because it costs less $ (and thats one language all companies speek $) .
Then again linux is open source , if you want to write a wrapper to load binary modules written to your own API specs , go do it your self (you might find some interested people , but as you can guess I would not be on that list) .
People who use bianry only drivers are more of the , the computer just has to work crowd (coincidentlly they dont code a lot) ; and the people who preffer OO drivers tend to be those who want to know WTF your doing on there system , and maybe even tweek your drivers.

In conclusion:Binary only drivers sux0r uber , so why make people make more?

Re:[OT] I'd define bloat differently.

[Anonymous Coward]

> OTOH, if you consider KDE to be pure bloat, then you are right, you probably should go back to using twm, resurrect your old 486.

Ever heard that there might be things inbetween bloat and completely outdated ;-)
I use BlackBox, and for me KDE adds absolutely nothing (except a very few progs that need the libs) and thus is bloat IMO.
Of course others might think differently, but finding KDE bloat and liking twm certainly isn't the same!