NVIDIA Drops Linux Driver Feature Not Found On Windows

[blackout23]

Fuck you Nvidia, I buy AMD.

Have fun with SuperTuxKart.

[brosis]

Actually, there is an advantage to the user - better performance and all the features you initially paid for. If nvidia deemed this monitor limit part of the price, then that's your problem, not theirs. If you happen to find a way to work around it, fantastic - you just cheated them. If this is a problem that only linux has, it's PROBABLY because they encountered issues when trying to make it work. Sure being open source would give the community the opportunity to fix the problem, but nvidia doesn't want people messing with their drivers because they don't want customers complaining to them about a problem they didn't cause. HOWEVER, it wouldn't be such a bad idea if nvidia had a separate fork of their current drivers and open sourced those (much like chrome with chromium), but, I'm sure one of the other core reasons for not open-sourcing their drivers is because of trade secrets they don't want AMD or intel to know.

If you haven't noticed, I didn't respond to your posts from the Maxthon article because you keep focusing on certain conditions or scenarios, you misinterpreted half of what I said, and you seemed to try a little too hard to argue with me. Proprietary software has it's place.


I would have to agree that it is pretty crappy of nvidia to put software limitations on their products, especially when they're already a bit on the expensive side. But, for the past 10 years, I have never liked nvidia for multi-monitor setups anyway. That being said, I would much rather buy nvidia for performance and overall working-as-advertised hardware than to go for them for multi-monitor setups.

You forgot that performance was by hacks and bugs are still there. Intel and opensource amd basically reinvent the wheel. Due to proprietary. Weren't previous implementations proprietary, one would get everything one paid for zero day, with top performance. Hardware is paid, no licensing troubles. Don't think one would bother if patented S3C is released open or close - main thing patent is paid with every single sold hardware instance, using the technology in question.

Yeah, I noticed you not responding, but I am not here to build your own opinion. I strongly disagree about "place" for proprietary, thats my position. Had been there many times, gave compromisses, backed off, same crap returns, always.
They can do it without, I know it, they don't care. They start caring if people start boycotting - regardless how one calls it; for obvious reasons. Encryption for banking or private communication - different topic, its middle man protection, not "take it or forget it".

Sure, its part of the price. The hardware is capable - nvidia were the top-down builders (cutters) since G80 anyway; still ... AMD solves it always in a more elegant - reverse way, by pretending to "forget". First Athlons, situation with ati 9500/9800, HD6xxx reflashing. I don't understand how one cheats them - can't manage to get features at price - don't give them. Crippling is really pissing in the glass, cause customer had less money than required. Its wrong.

I think nvidia does not opensource not due to any trade secrets, but due to shier amount of IP licensed under NDA (as per agreement), that nvidia has to comply; multiplied by DRM tentacles; multiplied by crippling desires - partially from "other" agreements, f.e. with MS, partially from own cases - like this one. Poor guy uses consumer grade card, 4th monitor - too good for him. I can imagine how mad nvidia management was at their engineering - they make life too fast too easy with their inventions for the masses ahead of the time. Should have waited for the milking team. Good to know, the "bug" was fixed. Boy, its just like Balmers hate against Linux - they are inventing too fast and too cheap, too.

Have fun with SuperTuxKart.

SuperTuxKart was problematic on open drivers about two years ago. Now it plays very nice and the game itself looks nice. Benefits all the way. Got a problem MESA's still stuck at OpenGL3? Well, nvidia is stuck at monitor count 3. :3

[agd5f]

More likely there are bandwidth issues to consider. While technically 4 displays might work, you may run into bandwidth limitations after a certain level. It's less confusing and potentially frustrating for the consumer to see that 3 monitors are supported vs 4 monitors with a long list of restrictions. Also on the driver side, implementing the logic to handle all the corner cases and restrictions can get really complex really fast.

[GT220]

Kepler supports 4 display simultaneously, Fermi and older GPUs can only do 2 outputs simultaneously, this is a hardware limitation.

Everyone here acts like a retard, no wonder no companies want to work with Linux Nazis who think they're entitled to anything.

[johnc]

And just to clarify, you can do 4 monitors on a single Kepler card that has four outputs. So I guess that's a workaround of sorts.

[uid313]

I'm sure one of the other core reasons for not open-sourcing their drivers is because of trade secrets they don't want AMD or intel to know.

I doubt that Nvidia have some secret sauce or knows some secret that nobody else in the world knows.
All of that stuff AMD and Intel already know.
All that stuff is probably presented at SIGGRAPH and other graphics conferences anyway.

It's not like Nvidia is years ahead of everyone else and have figured out things that nobody in the world knows.
The graphics field is an old one, its existed more than 40 years, so most stuff is already well-known.

Also, its not like Intel and AMD couldn't reverse engineer it anyway if they really wanted to.

[Kivada]

Have fun with SuperTuxKart.

Riiight... http://wiki.x.org/wiki/RadeonProgram/

[kokoko3k]

If this is a problem that only linux has, it's PROBABLY because they encountered issues when trying to make it work.

Oh, please, BaseMosaic worked good on linux with 4 monitors.
There's an interesting reply from Plattner to another user:

You can use the BaseMosaic option to span the desktop to up to three screens, or use a Kepler-based graphics board with support for up to four displays on a single card.

Did a bell ring?

[combuster]

All of this is old news, both nVIDIA and AMD differentiate their products often by imposing artificial limitations through firmware (rarely through drivers). There are some benefits, like ability to get GTX_XXX for 50-100$ less then GTX_YYY wich is basically the same card with some parts of the chip disabled because of glitches during manufacturing process (OEM's like this also because they don't toss out much of the silicon). As long as limitations are known to the potential byer before he/she purchases the product - I don't have a problem with this. However, what pisses me off is when you purchase a piece of hardware, you know it is capable of a certain feature wich is btw advertised and then dropped by firmware update, driver update, or disabled on another piece of hardware wich is required for the feature to work.

ie, bought i5-2400, saw it supports VT-d, bought H67 based mainboard, saw that the chip itself supports VT-d, and then all of a sudden get a smack in the face when Intel (to differentiate their offerings) instructed the OEM's to disable VT-d through BIOS on H67 mb's. And I was a early adopter - my first (and only) mistake when I bought parts for my new desktop PC.

If multimonitor support for 4 displays was advertised by nVIDIA for the mentioned GPU - than this is something you could sue them for. If not - you can only spit on their window, and decide if you ever gonna buy their products again.

[kokoko3k]

Well, as stated in the driver documentation you can, with basemosaic, use all of the monitors, not just 4, not just 3...
You are still "free" to use that driver until it will be supported by Xorg, but you know, the next abi change may break everything and nvidia cannot be sued for an abi change in the xorg package.