NVIDIA Announces Intel nForce Chipsets Coming

ruiner5000 writes "NVIDIA has just made a surprise announcement about their cross license agreement with Intel to make chipsets. This means that the bragging rights AMD users have had about having the superior nForce chipsets is about to end, and it will also bring NVIDIA's superior Linux support to Intel users. We have a statement and press release from NVIDIA about planned shipment dates, and expected products NVIDIA will be aiming their chipsets at. With the nForce 4 NVIDIA is aiming for desktops, laptops, workstations, and servers."

Comments

Superior Linux Support?

[Taladar]

Does this mean Intel Mainboards will require proprietary, closed source drivers like nvidia graphics cards that are a total maintenance nightmare because they break with every other kernel version in the near future?

Re:Superior Linux Support?

[marsu_k]

I take it you don't have a nForce motherboard? Because they work quite fine out of the box with 2.6 .x kernels. As for the display drivers, yes, they take some extra fiddling upon install every now and then. While this is not ideal, at least you get good 3D performance (not like "that other manufacturer" in Linux ;-). And you can use the default X driver if you don't need 3d acceleration, which is open source.

Re:Superior Linux Support?

[Richard_at_work]

Im sick of this (and will probably get modded down) but this isnt the sole fault of the vendors, now is it. For whatever reason, they will not release their driver set under opensource licenses, and thats agreeable because its their code and their decision. On the other hand, the linux kernel devs wont supply a stable module API, because they dont like binary modules, which is also agreeable because its also their code and their decision. This does leave the end user in the unenviable position of recompilations, but IMHO nvidea seems to have found a suitable halfway point for this, only requiring a stub recompile. But from where Im standing, its not just nvideas fault, both sides are posturing and trying to make a good situation out of a less than good one, but the majority of people on slashdot seem to blame vendors for supplying closed drivers when they have no real need to.

Re:Superior Linux Support?

[bingo777]

well my nforce 2 works perfectly fine with 2.6x..it even has support for the integrated lan card n 6 channel sound... so what else r u lookin for???

Re:Superior Linux Support?

[EvilGrin666]

Intel publishes specs for its hardware where as nvidia does not. This means we get a closed source driver from nvidia thats often out of date and doesn't work where as with Intel hardware we have open source drivers written by 3rd parties (often the kernel devs) in the kernel tree itself.

The Linux kernel devs have no interest in a stable module API because they have no interest in backwards compatability. If they see a problem, they go in and fix it. Next kernel release all the open source modules in the kernel tree are using the new fixed api.

Its true Nvidia might not have a choice in the matter though. They might have cross licensing deals or patent royalities on some of the technology they are employing in their motherboard chipsets that prevents them using an opensource license or publishing open specs.

Re:Superior Linux Support?

[ImpTech]

Nforce boards do work OOB, its true... but AFAIK not any better than an Intel board. And up until relatively recently, the nForce's didn't work at all. Furthermore, since Nvidia's dumped soundstorm, I can't imagine why an Intel user would buy a nForce board over one of Intel's own.

Re:Superior Linux Support?

[Anonymous Coward]

Because they work quite fine out of the box with 2.6 .x kernels.

Only partially thanks to Nvidia, though. Their closed nvnet and nvaudio modules aren't great, and it took a while for them to release stable graphics drivers.

Did they offer any assistance with the open drivers for all these?

Re:Superior Linux Support?

[legirons]

"I'm sick of this (and will probably get modded down) but this isnt the sole fault of the vendors, now is it. For whatever reason, they will not release their driver set under opensource licenses, and thats agreeable because its their code and their decision. On the other hand, the linux kernel devs wont supply a stable module API, because they dont like binary modules, which is also agreeable because its also their code and their decision. This does leave the end user in the unenviable position of recompilations, but IMHO nvidea seems to have found a suitable halfway point for this, only requiring a stub recompile. But from where Im standing, its not just nvideas fault, both sides are posturing and trying to make a good situation out of a less than good one, but the majority of people on slashdot seem to blame vendors for supplying closed drivers when they have no real need to."

Well that's all nice and eloquent,

but...

If it doesn't run a Free Software operating sytem, we're not buying it.

What makes people think we'd go to the trouble of writing billions of dollars' worth of software which is Free, only to sacrifice those principles for a crappy graphics card, modem, or motherboard?

Re:Superior Linux Support?

[bconway]

I'm buying it. Some of us put the best tool for the job over ridiculous ideological fanatacism.

Re:Superior Linux Support?

[Donny Smith]

>If it doesn't run a Free Software operating sytem, we're not buying it.

It will run on a free OS, only the drivers will probably remain closed source.

Personally I couldn't care less, all I want is stable and well-performing X-Windows.

Re:Superior Linux Support?

[Richard_at_work]

If thats your arguement, then why not stop moaning about the situation. The simple fact of the matter is that you have two groups, one of which has a particular ideological view, and to insist that the other group follows the same view is nothing short of stupid. And IMHO, Nvidia cards are far from crappy.

Re:Superior Linux Support?

[Anonymous Coward]

The Linux kernel devs have no interest in a stable module API because they have no interest in backwards compatability. If they see a problem, they go in and fix it. Next kernel release all the open source modules in the kernel tree are using the new fixed api.

And there you have it, summed up in a nice little couple of paragraphs why Windows is dominant. Linux is arrogantly postering from a position at the back of the pack and if it's not careful vendors are going to go back to ignoring it. NVIDIA is undedr no obligation at all to support Linux... who are you going to switch to, ATI? Their support is even worse!

On the other hand, I'm not a huge fan of the Nforce chipset either... I've had nothing but problems with it under Linux. My Asus A7N8X Deluxe board regularly hard-locks-up or reboots itself under Linux... I'm talking like every couple of days, or up to 2 weeks if I'm lucky. Now, I'm not going to entirely blame it on the chipset, but this motherboard completely sucks ass. We have 3 of them at work and they're all experience the exact same lockups... either it's a bad batch or there is a design flaw with these boards and Linux. My other nightmare board was an Asus a7n266-vm that I was using in a MythTV system. That would lock up regularly as well and eventually it started getting so bad I had to reboot daily. I switched to another system and have had no problems since. The A7N266-VM however wasn't even stable running Windows!!! I just chucked the piece of shit in the garbage and moved on. Thankfully the MSI board that replaced it has been rock solid. Asus == teh suck.

Re:Superior Linux Support?

[Anonymous Coward]

Intel publishes its specs? Hello, Centrino anyone?

Nvidia drivers often out of date? They are probably the most current drivers for any gfx card on linux.

Fact is, nvidias drivers are probably the best for linux, bar none. Being open source wouldnt improve the driver, so what do I care.

Re:Superior Linux Support?

[Lisandro]

Well, you have a point. The main reason Linux has no stable API for drivers is because it's in constant developement; and it has changed quite a bit over the last few versions. It also "encourages" (forces?) the release of OSS drivers, or specs, at the very least, and Linus himself stated this was one of the main reasons behind this line of thinking.

Like you said, vendors have the right to keep their drivers closed, which is understandable. Sometimes writing drivers is not trivial (gfx cards, notably) and takes a lot of time and money. What i find idiotic is that most of the time not even the techincal documentation for the hardware is available - if you don't want to play OSS, fine, you're entitled to, but i bet developers would love to have some specs in return. It's not unrational, and it's not fanatical.
Why they don't release specs is a whole different discussion. But it can be done.

Anyway, nVidia atleast goes through the trouble of providing very decent Linux drivers for almost all their products, even while closed source. I just wish they would go the extra mile and release specs.

Re:Superior Linux Support?

[Anonymous Coward]

The main reason Linux has no stable API...

Make that "One of the reasons Linux has no stable API...". I need some sleep.

Re:Superior Linux Support?

[pe1rxq]

But is was not because of the 'superior linux support' from nvidia. Look up why the name of the lancard driver was choosen.

Jeroen

Re:Superior Linux Support?

[EvilGrin666]

My Linux desktop box (yes, they do exist!) has a Tyan Tiger motherboard with an Intel chipset. A matrox graphics card and a realtek network card. It works flawlessly with Linux 2.6.9 using drivers readily avalible in the offical tree. I don't need to switch, I'm already there.

I agree with your Asus comment. My gaming rig has an A7N8X deluxe which was a pig to setup and get running. Once its there with XP installed its fine though.

Re:Superior Linux Support?

[ztane]

Seen that. The integrated APU in the chipset of the board is claimed to be supported by Linux... but you have to use i8x0 drivers for it if you want to use ALSA... and then you'll get that horrible high pitch noise that's also synchronized with bus activity... and the noise is audible even without speakers or headphones.

And the nForce APIC&ACPI incompatibility issue then, without the bus disconnect my processor has 10C higher core temperature. Now, there's some kind of workaround for that in the recent 2.6 kernels... but unfortunately under heavy load my kernel is missing timer interrupts or something and the clock ends up drifting 5 minutes every hour...

Re:Superior Linux Support?

[shutdown -p now]

Depends on what you mean by "relatively recently". I have been running Linux (first 2.4 kernel, then 2.6) with nForce2 mobos for more than a year now. Never seen anything not work.

Re:Superior Linux Support?

[EvilGrin666]

Yes intel publishes specs, just like these Centino technical specs here.

As for Nvidia's drivers being out of date, how long did it take them to fix the 4k stacks issues? months wasn't it? I also suspect my Matrox cards drivers are updated faster than the Nvidia drivers due to the fact they live in the kernel tree and get updated when everything else in the kernel tree does.

Re:Superior Linux Support?

[Taladar]

And if the vendor ever decides not to update the driver any more a few kernel minor versions later it won't work anymore.

Re:Superior Linux Support?

[Anonymous Coward]

Well I guess you'd better start developing your own graphics chipsets, wireless chipsets etc and go broke in the process wondering why China just ripped off your R&D effort and you can't afford to pay your engineers anymore, because someone has just produced exactly the same thing as you that costs 50% less.

d'oh.

Re:Superior Linux Support?

[Anonymous Coward]

Release the docs? So you've just created thousands of hours work for someone sanitising internal only documentation (probably with rude comments about peoples design skills etc) supporting queries, updating information as new chip revs come out, etc. If they are able to release drivers that work well enough for the blue chips who want to run Linux on these components, why should they expend such effort to support the few die hards who want to stay in on saturday night recompiling their kernel?

Re:Superior Linux Support?

[Anonymous Coward]

The simple fact of the matter is that you have two groups, one of which has a particular ideological view.

Specifically, one group has the ideology that anyone who wants to buy their kit must submit to whatever contract they shove in front of their customers, you must do x, you must do y, you must not copy, you must not modify, you must not install without asking us first, you must not share, you must not help others - and to insist that the other group follows that same view is nothing short of stupid.

Read up on your history -- proprietary software is a radical ideological fanatacism, the concept didn't even exist until a few years ago, and the rate at which people have accepted the idea of forcing their own view onto others and keeping power and control over anyone who uses their software, is nothing short of frightening.

Re:Superior Linux Support?

[molnarcs]

The 'vendor goes AWOL' kinda argument in the close vs. open debate is getting really boring. Problem is, that it is true - but in this case: what do you think is the chance that nvidia will just screw its linux userbase?

None. The reason: they have excellent and commited developers. They use a unified code-base for all their drivers. Occasionally, they go out of their way to provide support even for less hyped operating systems (FreeBSD, for instance). This is an old thread, and FreeBSD with 5.3 has proper (proper for Nvidia's needs) tls implementation, but still, it is a good example of nvidia's commitment to work with open source developers on issues with their drivers.

So yes, the 'what if company X choses to no longer support opensource' has a ring of truth to it (and opensource advocates can always have it open in kwrite for copy & paste job for every newsbit about closed source drivers), but you always have to think about the specific case. Does nvidia have good support for their chipsets for linux/freebsd? Yes, it does, and these drivers are quite up to date with their windows counterparts. Are they willing to address various issues with their drivers? Yes they are. Are they willing to opensource their drivers? No. And that seems to bring out the worst of zealotry in opensource land. I'm not addressing this to parent post specifically btw, but to all who beat the 'company X suxorz cause they're proprietary' drum.

Re:Superior Linux Support?

[Lisandro]

Release the docs? So you've just created thousands of hours work for someone sanitising internal only documentation (probably with rude comments about peoples design skills etc) supporting queries, updating information as new chip revs come out, etc.

They don't have to, of course. But releasing specs it's a hell of a lot easier (and cheaper) for them than releasing a binary driver, which also requieres thousands of hours to develop and test. It's obvious that Linux is not an afterthought for nVidia; the quality of their drivers proves it, but their fixtation with binary only drivers causes endless headaches to the end users. Like i said, there's no real reason why at least specs can be published. That's all the excuse they need to separate themselves from the problem.
All of this, assuming of course the driver source itself can't be opened because of IP restrictions.

Re:Superior Linux Support?

[Jeff DeMaagd]

As far as I'm concerned, nVidia is a hardware company, not a software company. They can dick with little bits of semantics, but IMO, hardware companies should provide coding specs in how to use their hardware.

Re:Superior Linux Support?

[linefeed0]

I'm not really sure you understand the issues at stake here.

First, a stable module API doesn't matter. nV's drivers have the same fucking bugs for years' worth of releases. It's normal to expect some bugs for such a complicated piece of software when porting it to a new kernel major release, but these bugs happen in 2.4, 2.6, probably 2.2.

Even Microsoft is pissed off about this. I was told by a Microsoft engineer a couple years ago that drivers are the #1 cause of BSOD's on Windows by a long shot. If the Windows kernel is hobbled by bad drivers, what do you think happens to the Linux kernel, which is arguably better quality overall? The fact is, with all the code running in ring 0, and with privileged access to hardware that cannot be easily reinitialized if the driver crashes (ever tried to reload the NVdriver module?), it doesn't matter how good the kernel is. The Linux kernel developers are right to not support this madness because it dramatically lowers the delivered quality of their product. It's just plain old QA.

I do see a potential compromise that I'd embrace. NVidia could work with X.org to add hooks to the "nv" driver, so 2D could be rock solid all the time and I could use the same driver without rebooting for 3D. Even if 3D crashes every now and then, it matters a lot less, especially because I'm not going to get the same 3D elsewhere.

Re:Superior Linux Support?

[Anonymous Coward]

I, for one, *do not* welcome this stupid binary only driver overlord.

One day, Linux is nothing but thin layer of kernel abstraction just to support *all*, YES, *ALL* hardware, just because other hardware vendors distributing proprietary binary only drivers, following good karma of nvidia. Do you want to end up like that?

Re:Superior Linux Support?

[RR]

Im sick of this (and will probably get modded down) but this isnt the sole fault of the vendors, now is it. For whatever reason, they will not release their driver set under opensource licenses, and thats agreeable because its their code and their decision. On the other hand, the linux kernel devs wont supply a stable module API, because they dont like binary modules, which is also agreeable because its also their code and their decision.

Linux is not an x86-only program. While that's not a problem with a chipset, there's that handful of Mac users running Linux that can't get speedy graphics. Apparently, nVidia can't produce a PPC driver anytime soon, as they have let Apple do the driver development all this time. The only real option is to switch to ATI.

Even if nVidia produces a PPC driver, that wouldn't be any help to the (admittedly small) groups of people using SPARC or Alpha or MIPS as their desktops.

Re:Superior Linux Support?

[dnoyeb]

I been using a NVIDIA nforce board for my linux box for a couple years now. They have had some driver goofs, but the recent installer is a gem and they continue to get better.

Re:Superior Linux Support?

[Anonymous Coward]

i have the same motherboard, it works perfectly in 2.4.16 on up with no issues with any drivers using either the supplied forcedeth / i8x10 drivers or the official nVidia drivers.

Even debian will install and work perfectly with sound and network out of the box with no fugging.

you must not be as linux savey as you think

Re:Superior Linux Support?

[Anonymous Coward]

please, nVidias primary market is Gamers. Gamers dont run linux or SPARC or Alpha or MIPS because they suck for gaming.

nVidias market is people who run the fastest hardware and the latest version of windows to play the newest games.

There are bound to be more of them than linux, Mac, sparc, alpha, and MIPs desktop users combined.

nVidia is being more than generous offering an official Linux and BSD driver to the (small) linux using market.

Re:Superior Linux Support?

[Anonymous Coward]

The fact that the kernel devs change the module API at the drop of a hat is probably a good reason many vendors dont support linux with their own drivers.

I know if i was a hardware developer that couldnt release ALL of the specs on the hardware we develop because of IP issues, i'd consider releasing a closed source driver (cant open source with IP).

However, if i had to spend more development time fixing the driver because the kernels API changes every minor release, it wouldnt be top priority to keep up-to date with kernel releases.

If my main market was say, for server use. I know my users wont run the latest stable kernel because it hasnt been tested, so that way i can talior my driver to work with the API of kernel 2.4.21

Now, several of my users have other hardware that requires at least 2.6.4, which changes included an API change that breaks my kernel interface. So then i have to spend more money to tailor to that kernel, spend more Q&A finding and fixing bugs, and who knows what other issues i'd run into, just to support maybe a dozen users?

Now at the same time, the windows driver API changes what, once a decade? so i can get the driver SDK, write a rock solid interface that WONT CHANGE over the next 10 years and concentrate on making the drivers rock solid for my windows market, which is probably 90-95% of my sales.

Now the question i pose to you, which makes more sense for business.

Would you want to spend a lot of money on R&D, dev wages and Q&A to retool the driver every minor kernel release so that you support maybe 5% of your sales?
Keep in mind that the cost of R&D, wages and Q&A is greater than the total revenue provided by the non-windows market.

Re:Superior Linux Support?

[ImpTech]

I'm mostly talking about the rampant APIC issues, which were crashing people's systems all over the place up until maybe 4 months ago. And remember we're talking out of box install, so anything you got working with drivers from nVidia doesn't count. Not sure when they submitted their AGPGART patches, but their onboard NIC had to be reverse-engineered and that couldn't have been more than 9 months ago.

Now I'm not in the Intel camp anymore, so I don't know offhand how long it takes for their new chipsets to be supported. But as a nForce2 user, I can say that it took at least a year from the first boards being released to reasonably good support in the Linux kernel mainline.

Re:Superior Linux Support?

[niko9]

But from where Im standing, its not just nvideas fault, both sides are posturing...

Posturing? The kernel devolpers? Two way street? It's been said before on the LKML that Linus will not help out the compaines who are interested in providing only closed source drivers, 'cause they don't help him out, and ther rest of the gang, with specs. Sounds more like a one way street is you ask me. Compaines bemoaning the lack of cooperation from Linus while clinging to their NDA's and "third party trade secret" arguments.

I'll stick with Intel and VIA for now, thank you very much.

Re:Superior Linux Support?

[Anonymous Coward]

The key issue for Nvidia is that, just like ATI, all of their chips have various bugs that are worked around carefully in the binary drivers, and which they have no positive reason to expose to their competition.

Re:Superior Linux Support?

[IntergalacticWalrus]

Who's "we" anyway? The entire Free Software community? Please don't include me in, you fucking stallmanist zealot.

Re:Superior Linux Support?

[Ogerman]

And how exactly would "exposing" bugs in their own hardware help their competition? Afraid of the competition copying their bugs? C'mon now. And it's not like people don't already know that bugs are more or less inevitable, so there's no marketing spin available.

Re:Superior Linux Support?

[Anonymous Coward]

"It's the driver hell, stupid."

I think that the Linux community has a Love/Hate relationship with nVidia. We love their drivers when they work, and we hate their drivers when they don't. My 3D acceleration worked great with FC2. With FC3, it locks up and hard.

But I guess what strikes me is that this problem was known for a while since Fedora released Core 3 Test 3. It wasn't a big surprise that, yet again, there were going to be problems with nv video cards in another new release of a major distribution. I guess my question is: haven't we seen this before? Can't they track problems with their drivers on the pre-releases so they can fix them by the next version release?

The problem for Linux in the long term is that it becomes harder for Linux to continue to make inroads if people have to deal with driver hell. It's not an acceptable solution to tell people to limit themselves to 2D (and constant fiddling with the xorg.conf to try to get their 3D to work) when they've shelled out big bucks on a nice 3D card.

People were in dependency hell until yum/apt-get/etc. People were in GUI hell until gnome/kde. People were in nvidia driver hell until....? You get the idea.

Re:Superior Linux Support?

[Surt]

Perhaps more importantly in the particular case of nvidia, what do you do if they go out of business (or just get bought in a merger), and the buyer (say ATI) decides to bury their product line as a business decision. If the driver source has never been made public ....

Re:Superior Linux Support?

[EvilGrin666]

I'm aware that its possible to make it work. Thats not the point I'm trying to make. Its the missing features plus lower performance compared to the windows drivers and the closed source drivers with no published specs that IMHO is the problem.

Kudos to the guys for reverse engineering the hardware though but they really shouldn't have to waste their time doing that.

Re:Superior Linux Support?

[Ogerman]

And there you have it, summed up in a nice little couple of paragraphs why Windows is dominant. Linux is arrogantly postering from a position at the back of the pack and if it's not careful vendors are going to go back to ignoring it.

Driver support has precisely nothing to do with why Windows is dominant, unless you're talking niche markets like the PC extreme gaming crowd. (yes, it is sadly now a niche market for all practical purposes thanks to XBox and PS2) Other than 3D hardware (and some buggy non-standards compliant crap), most works perfectly fine with Linux.

When desktop application availability turns the tide in Linux's favor, every hardware company out there will be bending over backwards to make sure their products work perfectly in Linux out of the box. To do this they will have two valid options: 1.) provide open source drivers or 2.) provide adequate specifications so that other people can provide open source drivers. (usually this only means adding recognition + tweaks to existing drivers because of common chipsets, etc.) We can thank the kernel developer's "arrogant" stance for this. The final result will be the most stable PC platform ever known. Do you realize that many of Windows' stability problems are caused by bad drivers? (Even with it's "stable binary driver API") Guess what? That approach doesn't work and the kernel developers know it! The simple fact is, we cannot rely upon hardware developers to provide quality drivers for their own products or to keep their drivers up to date with the OS! It doesn't even work for Windows with its monopolistic share of the market!

Other reasons why open source drivers are imperative: 1.) Continued support if the hardware manufacturer disappears 2.) Security - can you trust downloaded 3rd party drivers from some random server in Taiwan? Remember, this is software that essentially has root on your machine.

Re:Superior Linux Support?

[Ogerman]

..but the majority of people on slashdot seem to blame vendors for supplying closed drivers when they have no real need to.

You don't really understand the issue. There are very good reasons why there is no "stable binary module API" for Linux. This stance has the power to transform the industry in a positive way.

When desktop application availability turns the tide in Linux's favor, every hardware company out there will be bending over backwards to make sure their products work perfectly in Linux out of the box. To do this they will have two valid options: 1.) provide open source drivers or 2.) provide adequate specifications so that other people can provide open source drivers. (often this only means adding recognition + tweaks to existing drivers because of chipsets commonality, etc.) We can thank the kernel developers' "arrogant" stance for this. The final result will be the most stable PC platform ever known. Do you realize that many of Windows' stability problems are caused by bad drivers? (Even with it's "stable binary driver API") Guess what? That approach historically doesn't work and the kernel developers know it! The simple fact is, we cannot rely upon hardware developers to provide quality drivers for their own products or to keep their drivers up to date with the OS! It doesn't even work for Windows with its monopolistic "unifying" share of the market!

It comes down to this: hardware companies are best at designing hardware. It is their core competancy. They should fully document that hardware and then mostly leave the software to people who are good at designing software to work with it. (Whether that's for Windows or Linux) What is inside that hardware doesn't matter. It can be a black box for all we care. But the *interface* should be well documented so that we know how to properly talk to the hardware. (This is how the whole rest of the electronics industry works.. computer hardware is currently an anomaly)

Other reasons why open source drivers are imperative: 1.) Continued support if the hardware manufacturer disappears 2.) Security - can you trust downloaded 3rd party drivers from some random server in Taiwan? Remember, this is software that has root on your machine! 3.) Platform support. Remember that open source OS'es are not just for 32-bit x86!

Re:Superior Linux Support?

[Cyno]

Well, one day a vendor will come along that supplies source code. Then we'll finally get some nice graphics hardware that fully integrates with the OS. When this happens the ball will be in nVidia's court to explain why their drivers don't fully integrate or why they have bugs and other problems that haven't been addressed.

Will it happen? Probably, but not today. So we'll see how it all plays out.

Re:Superior Linux Support?

[claar]

You must not have the nForce2 MB I do, because I have yet to find a distribution that even came close to working out of the box on mine. It's to the point that I've given up Linux on my desktop, and I'm not even a Linux newbie..

Even 2.6x kernel'd Gentoo fails to compile several core packages on my ASUS nForce2 MB.. I give up, I've never seen such horrible hardware support for any chipset, especially after said chipset has been out for years and years. Linux is wonderful when it works.. but when I have to spend 20 hours with APIC, X, sound, and network issues every time I install it, it makes Windows look better and better..

Re:Superior Linux Support?

[Richard_at_work]

Uhm, precisely what you describe IS a two way street, just with no traffic passing through it. You say "why should Linus et al provide a stable API when vendors arent supplying specs", well I come back with "why should the vendors supply specs when they are being given a moving target". As I said, posturing, on BOTH sides. Linus isnt helping the vendors while the vendors arent helping Linus, and both have their own valid reasons.

Re:Superior Linux Support?

[Richard_at_work]

When desktop application availability turns the tide in Linux's favor, every hardware company out there will be bending over backwards to make sure their products work perfectly in Linux out of the box.

Ahem, bollocks. This wont happen, Microsoft provided a driver signing program to ensure that drivers conformed to the windows specification. Funny that, the number of unsigned drivers that I come across when installing windows hardware, most jsut went to the bother of adding extra parts to their install documentation to bypass the unsigned driver warning. What makes you think vendors will be "bending over backwards" to support Linux?

To do this they will have two valid options: 1.) provide open source drivers or 2.) provide adequate specifications so that other people can provide open source drivers.

Well, 1) Why should they do this? Why do they have to subscribe to the same ideology that you do? Linux has been around for 14 years now, and the number of mainstream vendors releasing opensource drivers is still extremely low, what makes you think it will get better. 2) This has been discussed before, in a lot of cases hardware manufacturers are just plain unable to release specifications or driver code due to third party licenses. Stamping your feet and demanding it wont make the situation any better.

They have a third valid option, and that is to provide binary drivers. If the kernel devs cant supply a stable API due to their ideology, then thats their loss.

I can take issue with the 'arrogant stance' of the kernel developers. They demand something and think that everyone will fall into line. Well, that doesnt happen in the real world, and when it doesnt its the end users that suffer. As I said in my origional post, the vendors shouldnt solely take the blame for this, the kernel devs must take a fair share too. The Linux market just isnt big enough at the moment for such demands to really have an impact, theres no power behind it.

It really is arrogant to think that, just because linux exists under the ideology that it does, it should be given special treatment. Linux is the one breaking the status quo in this matter, but it gives little back in return for what it demands.

Re:Superior Linux Support?

[shutdown -p now]

I didn't use any drivers from nVidia. I just compiled and installed gentoo-dev-sources on my Gentoo box and it worked.

Re:Superior Linux Support?

[Ogerman]

What makes you think vendors will be "bending over backwards" to support Linux?

Because it will be the most popular desktop platform and they'll have to support it properly if they want to sell hardware. Nothing complicated about that argument.

Well, 1) Why should they do this? Why do they have to subscribe to the same ideology that you do?

Because it works maybe? That's the whole point. When Linux begins to dominate, it will have pulled the entire industry towards Open Source. At that point, the vendors will be fighting against the tide to try to stay binary only. Maybe you don't think that will ever happen. If so you're wrong.

Linux has been around for 14 years now, and the number of mainstream vendors releasing opensource drivers is still extremely low, what makes you think it will get better.

Linux has only entered the mainstream in the last 2-3 years. The number of mainstream vendors with supported hardware is already very high, whether or not they themselves actually wrote the drivers. What makes me think it will get better? Look at the track record thus far and refer to the first argument.

2) This has been discussed before, in a lot of cases hardware manufacturers are just plain unable to release specifications or driver code due to third party licenses. Stamping your feet and demanding it wont make the situation any better.

This is a BS argument in the majority of cases. The only major field where there is a significant amount of 3rd party licensing is 3D graphics chipsets. Even then, they don't have to release details of internal operation or actual software. We're talking about register level documentation. The interface.. It is doubtful that there are many cases of 3rd party licensing issues here. In other cases, once Linux takes off, vendors will simply have to resolve their licensing issues to get the drivers out. Maybe this will tear down a lot of the "IP" cross-licensing stupidity that exists today. Good riddance. Everyone will be better off for it, including the vendors.

They have a third valid option, and that is to provide binary drivers. If the kernel devs cant supply a stable API due to their ideology, then thats their loss.

If vendors are obsoleted by their reluctance to open specifications and therefore be properly be supported in Linux, then that's there loss. Again, good riddance.

I can take issue with the 'arrogant stance' of the kernel developers. They demand something and think that everyone will fall into line. Well, that doesnt happen in the real world, and when it doesnt its the end users that suffer.

There have been dozens of cases where hardware manufacturers have succumbed to pressure from customers and released docs to get their products working in Linux. And many have realized that it makes sense to just jump on the bandwagon across the board with all product lines. Note: pressure from customers. The market will drive the opening process. Not kernel developers. But the ramifications of their philosophy in proper software design is helping to improve the industry.

As I said in my origional post, the vendors shouldnt solely take the blame for this, the kernel devs must take a fair share too.

The binary only driver model is a blatant failure on all platforms. You don't quite seem to get this do you? That means regardless of whether we're talking Linux or Windows, it is the vendor's fault.

The Linux market just isnt big enough at the moment for such demands to really have an impact, theres no power behind it.

At the moment. And yet, most common hardware is already supported. It will only get better from now forward. Go down to your local electronics discounter and look on hardware boxes. You'll find that about 90% of them mention support of Linux now. No power behind it? Right.

Linux is the one breaking the status quo in this matter, but it gives little back

Re:Superior Linux Support?

[justins]

The Linux kernel devs have no interest in a stable module API because they have no interest in backwards compatability. If they see a problem, they go in and fix it. Next kernel release all the open source modules in the kernel tree are using the new fixed api.

That's not really the issue. The problem for driver writers is that the binary interface into the kernel can change between minor versions of the kernel, or just with different sorts of kernel patches of the same kernel version.

I'd actually love to hear an explanation of why Linux doesn't offer a binary interface into the kernel for driver writers (which isn't total bullshit). It really seems as though they're too lazy to do it, or they just hate binary drivers. If it's the latter, criticizing the kernel developers a little isn't out of line.

Re:Superior Linux Support?

[Asphixiat]

I have an ASUS A7N8X-E Deluxe and have used every 2.6 kernel since 2.6.0 on it, never had any lockups etc.

The only downside was using the i8x0 driver for sound for a long time, but now I'm using the driver from nvidia it's fine, I even have lm-sensors working (since 2.6.7 or something).

he downside of using the nvidia sound drivers is they are OSS....but I have at least 6 HW channels - so it's better than the i8x0 ALSA drivers that truely sucked (only one channel unless you went and setup the DMIX plugin shite).

At work I set up a GIGABYTE nforce2 MB for a dev server, and it's perfect also. Both systems are gentoo :)

If you're getting lockups from the APIC, I suggest you look into that patch (love sources used to have it), but I think it got merged into the kernel proper a while back.....maybe you have a HW issue that windows doesn't know about (yet).

I wish NVIDIA would give me ALSA drivers, but the OSS code is open, and I don't think it links to any libs....so I am surprised noone in the ALSA team has just "liberated" the code yet. Guess they don't care. The only thing it doesn't do yet is the HW AC3 encoding, but I guess that'd be up to ALSA & the guys who do those sort of projects to figure out.

Re:Superior Linux Support?

[Ryan Huddleston]

source: http://lwn.net/1999/0211/a/lt-binary.html

To: linux-kernel@vger.rutgers.edu
From: torvalds@transmeta.com (Linus Torvalds)
Subject: Re: Kernel interface changes (was Re: cdrecord problems on
Date: 7 Feb 1999 08:15:24 GMT

In article <79g5bu$spd$1@palladium.transmeta.com>,
H. Peter Anvin <hpa@transmeta.com> wrote:
>
>* Linus Torvalds has no interest whatsoever in developing such a
> plug-in ABI. Someone else is welcome to do it.

No, it's even more than that.

I _refuse_ to even consider tying my hands over some binary-only module.

Hannu Savolainen tried to add some layering to make the sound modules
more "portable" among Linux kernel versions, and I disliked it for two
reasons:

- extra layers decrease readability, and sometimes make for performance
problems. The readability thing is actually the larger beef I had
with this: I just don't want to see drivers start using some strange
wrapper format that has absolutely nothing to do with how they work.

- I _want_ people to expect that interfaces change. I _want_ people to
know that binary-only modules cannot be used from release to release.
I want people to be really really REALLY aware of the fact that when
they use a binary-only module, they tie their hands.

Note that the second point is mainly psychological, but it's by far the
most important one.

Basically, I want people to know that when they use binary-only modules,
it's THEIR problem. I want people to know that in their bones, and I
want it shouted out from the rooftops. I want people to wake up in a
cold sweat every once in a while if they use binary-only modules.

Why? Because I'm a prick, and I want people to suffer? No.

Because I _know_ that I will eventually make changes that break modules.
And I want people to expect them, and I never EVER want to see an email
in my mailbox that says "Damn you, Linus, I used this binary module for
over two years, and it worked perfectly across 150 kernel releases, and
Linux-5.6.71 broke it, and you had better fix your kernel".

See?

I refuse to be at the mercy of any binary-only module. And that's why I
refuse to care about them - not because of any really technical reasons,
not because I'm a callous bastard, but because I refuse to tie my hands
behind my back and hear somebody say "Bend Over, Boy, Because You Have
It Coming To You".

I allow binary-only modules, but I want people to know that they are
_only_ ever expected to work on the one version of the kernel that they
were compiled for. Anything else is just a very nice unexpected bonus if
it happens to work.

And THAT, my friend, is why when somebody complains about AFS, I tell
them to go screw themselves, and not come complaining to me but complain
to the AFS buys and girls. And why I'm not very interested in changing
that.

Linus

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

Re:Superior Linux Support?

[BobaFett]

Good point. Hardly unique to closed source though. Remember when Fedora Core 2 was released, they disabled Firewire in the kernel? Firewire maintainer stopped supporting the drivers, for whatever reason, and they were broken in 2.6 kernels for long long time. And the best thing the FC developers could do is to disable the feature alltogether.

In theory, there are all these nice things about OSS and its support and how if it gets dropped someone else can pick it up and so on. Sometimes it happens this way, and sometimes it does not. And lately it's more often the latter. One example: say, I have some strange error message in the log, or some module won't work. I search google, net and news. Two years ago, if I found an article describing this problem, most of the time there was a followup. Not always resolution, but at least followup. Someone was at least answering. Today, at least 80% of hits the search finds are one-article threads. Nobody bothers answering. It's really depressing, but I just don't see the current OSS support model scaling and keeping up with the rate of Linux adoption.

Re:Superior Linux Support?

[MojoStan]

Furthermore, since Nvidia's dumped soundstorm, I can't imagine why an Intel user would buy a nForce board over one of Intel's own.

1. SLI graphics (2 graphics cards powering 1 monitor)
2. 3Gb/s SATA (Intel uses 1.5Gb/s SATA)
3. NVIDIA ActiveArmor (firewall)

I'd like to see NVIDIA release a low-end chipset with integrated graphics that competes with Intel's DirectX 9 compatible GMA 900 graphics. Even with shared system and graphics memory, integrated GeForce 6200 graphics would smoke GMA 900. Plus, NVIDIA would probably include a PCI Express x16 slot, unlike Intel's low-cost 915GV and 910GL chipsets.

Unfortunately, like you said, it seems like NVIDIA dumped SoundStorm. Meanwhile, Intel added Intel HD Audio, which apparently encodes Dolby Digital 5.1 in real time.

Re:Superior Linux Support?

[Donny Smith]

> And if the vendor ever decides not to update the driver any more a few kernel minor versions later it won't work anymore.

It's not like we're buying 100TB of enterprise storage with virtualization that requires proprietary drivers so servers can see that storage - it's just a PC.

Even if nVidia stops updating drivers within a year (late 2005), people can stick with the then-current 2.6 kernel which will be good for at another year or so. And by late 2006, nForce4 systems will be outdated so badly that people (at least those who like the word "latest") will want to replace them anyway.

And what is the probability of discontinoued Linux support now that they'll triple shipments of nForce4 chipsets (after having reached a cross-licensing agreement with Intel)? Miniscule.

What is the worst case scenario? You get stuck with a so-so kernel 2.6 which can't be updated to latest something. Not exactly a disaster. Not to mention that one will be able to run _any_ Linux kernel on top of VMWare Workstation running on Windows XP or 2003. Compared to time-wasting fooling around with other chipsets/drivers, that's an acceptable risk to take.

Re:Superior Linux Support?

[HuguesT]

Yes and if the vendor suddenly goes belly up open-source drivers ensure the hardware will continue to run.

Re:Superior Linux Support?

[Trejkaz]

My VIA motherboard also worked out of the box, and I didn't have to install any extra, closed source drivers, either.

Re:Superior Linux Support?

[SenorCitizen]

Furthermore, since Nvidia's dumped soundstorm, I can't imagine why an Intel user would buy a nForce board over one of Intel's own.

I don't see why Soundstorm would have been the killer app. The APU was a broken design that tried to implement Sensaura's EAX algorithms in hardware, but failed. EAX support *never* worked correctly, in spite of dozens of audio driver versions. No wonder they dropped the whole thing.

Way to go NVIDIA...

[WIAKywbfatw]

Not. Talk about screwing over the average guy just to make a fast buck.

Re:Way to go NVIDIA...

[Seven001]

How exactly are they screwing over the average guy? Granted I dislike Intel greatly and refuse to use their products, but I don't see how its screwing anyone over. I don't like the partnership either, but unforunately theres not a damn thing anyone can do about it except hope the Intel version of nForce flops.

Re:Way to go NVIDIA...

[Anonymous Coward]

Why exactly do you refuse to use their products? Are you another hypocritical Slashdotter who believes in "good for-profit companies" and "bad for-profit companies"?

Re:Way to go NVIDIA...

[BCW2]

Some of us refuse to take the hit in the wallet for Intel. AMD is still the best bang for the buck. I know that my reasons are strictly money. Intel is overpriced for what you get.

Re:Way to go NVIDIA...

[Seven001]

There are several reasons why I refuse to use their products. For one, I just don't like a company that relies more on marketing than actually having a superior product. I think their marketing is intended to promote ignorance, like AOL's "make the internet better" bullshit. Another reason, same as BCW2 gave, I don't like their pricing. It's way too high for what you get. Then of course theres the part where I like to root for the underdog when I think they are deserving of more success. AMD has relied far more on the quality of their products than on marketing, and that makes them deserving if you ask me. I have NEVER seen an AMD commercial, and I very scarcely see any other kinds of ads from them.

Am I the only one who saw this coming?

[Valarauk]

Intel and NVIDIA cut a deal with Sony Online Entertainment to get their products given "premier" status, complete with back to back advirtisments while loading games like Everquest2.

Re:Am I the only one who saw this coming?

[EvilGrin666]

if you mash the keys really hard you can skip the adverts in the EQ2 load screens.

Good News for Nvidia and Intel

[blueZhift]

I guess this is good news for both Nvidia and Intel. This should help Nvidia make up for being shut out of the Xbox 2 graphics game, though they may have lost money on the original Xbox deal anyway. And this should bring some gamer cred back to Intel who may have been using some gamer sales to AMD because of the nForce chipset. Of the two though, I think Nvidia gets the best part of the deal since they will now have an easier entry into the wider PC market which is dominated by Intel based systems. Intel will only see marginal gains since gamers are not a big part of the market, though they do buy a good proportion of high end systems I would guess.

Re:Good News for Nvidia and Intel

[MrWim]

I don't understand how it would affect the XBox deal. The XBox 2 will be based on the PowerPC architecture from IBM and will probably have little to do with intel.

Re:Good News for Nvidia and Intel

[nonmaskable]

>>

They've lost gamer sales because the AMD 64 processors are a much better value for gaming than Prescott spaceheaters. The mainboard chipset is a pretty marginal contributor to framerates.

Re:Good News for Nvidia and Intel

[blueZhift]

Nvidia provided the graphics chipset for the original Xbox. Then they had some sort of falling out with Microsoft over how much they were supposed to be paid. This probably contributed to the Xbox 2 graphics chipset being done by ATI. So Nvidia will be missing the Xbox 2 party, but given that the Xbox deal probably wasn't good for them anyway, it may not be a big loss.

Re:Good News for Nvidia and Intel

[Anonymous Coward]

Intel lost gamer cred b/c their chips are out-freaking-rageously priced and AMD has a MUCH better price/prformance ratio. Nvidia lost gamer cred b/c ATI simply built a better product, the same reason Nvidia HAS cred for chipsets.

Re:Good News for Nvidia and Intel

[teg]

They've lost gamer sales because the AMD 64 processors are a much better value for gaming than Prescott spaceheaters. The mainboard chipset is a pretty marginal contributor to framerates.

The performance difference between the CPUs isn't that big... the performance increase gained by SLI on the new nforce boards, OTOH, seems significant.

Whoops...

[WIAKywbfatw]

Either the story summary just changed or I misread it completely as saying that nForce chipsets would no longer be available for AMD users. It's probably the latter, so feel free to call me an idiot, but given that it's Slashdot, I wouldn't be surprised if it was the former.

Regardless, even as it stands now, it's hardly the least ambiguous story summary I've ever read.

Re:Whoops...

[Anonymous Coward]

Third sentence of TFA.

They just can't allow AMD to have all that nForce 4 goodness on their own...

I know this is Slashdot and people don't actually bother to read the articles but let me ask you this - how, if you don't give the articles so much as a cursory glance, do you know what the hell you're talking/flaming about?

Do you just glean all your information from what the editors summarize? <Shudder>

Do you just regurgitate what you think will be popular opinions in an attempt to win approval?

Oh... suddenly the veil lifts. You're probably no different from many other people here. But, unfortunately, that makes you just as worse.

Re:Whoops...

[Anonymous Coward]

Did you read my F post? I said that, although I can't prove it, there's a good chance that the story summary was changed between me posting the first time and me posting the second time. DUH.

heh, I was going to ask just that... [nt]

[cockroach2]

nt, i say!

meh

[grumpygrodyguy]

Since nvidia dropped Soundstorm I no longer have any reason to care.

Re:meh

[the unbeliever]

Soundstorm isn't really necessary anymore.

And did you happen to notice the dearth of nForce boards with Soundstorm? It wasn't nVidia that killed it, it was the mobo makers. They didn't want to pay for it.

Re:meh

[Anonymous Coward]

Really?

Blimey! Where can I find real time on the fly dolby digital compression all done in hardware then?

And yes, I don't need it. But once I've tasted it I miss it, a lot.

Curse you Asus/nVidia/The ant-god of hardware sound/whoever, curse you to hell

Re:meh

[foldedspace]

Someone needs to mod the parent up. Soundstorm was one of the big things nVidia had over VIA and SiS. The other thing was not being VIA or SiS. Not that VIA hasn't gotten tons better lately. I've never really been interested in SiS.

Yeah, you could just get an Audigy 2 ZS or whatever, but it's not the same AND it's another $80+. The nForce motherboards were about the same price as the VIA motherboards, but they had far superior integrated audio.

I don't see a lot of people moving from Intel chipsets to nVidia chipsets without a slight regret, unless nVidia is keeping another secret. The integrated firewall on the NIC might be a nice plus though.

The press release says (to me) that it's just so Intel gets to have SLI available to its customers. SLI is going to be a very small niche market. 6600GTs in SLI cost the same as a 6800GT and are roughly the same speed in some tests. Twin 6800GTs or Ultras are about all that would be worth it and it's a LOT of money, heat and power for one computer.

Re:meh

[mp3phish]

Mobo mfg's not picking that feature up is nVidia's fault, not the mobo mfg's.

If they didn't want to pay for it, it means they have real research to show that their end users weren't willing to pay over 50$ more for integrated sound. They were correct. nVidia was overcharging. period.

not to mention that soundstorm works like shit in linux.

Re:meh

[the unbeliever]

Funny, my motherboard only cost $92 plus shipping and has soundstorm (Abit NF7-S) -- are you saying that without soundstorm this would have cost only $42? Don't think so.

Soundstorm is only necessary for media boxes anyway. I got the board because it's a good performer.

Re:meh

[batkiwi]

Soundstorm IS necessary.

Name another audio "solution" which is BUYABLE TODAY which does real time digital 5.1 encoding. DTS or DD, I don't care as my receiver does both, but I want 5.1 through a digital interface (coax or optical).

I'll give you a hint:
There are zero on the market. The only possible contender is that new intel pro-audio onboard thing, but NO ONE has plans to impliment it yet.

NOTHING by creative does this, nor do those "7.1" fortissimo/etc cards.

Re:meh

[MojoStan]

Name another audio "solution" which is BUYABLE TODAY which does real time digital 5.1 encoding. DTS or DD, I don't care as my receiver does both, but I want 5.1 through a digital interface (coax or optical).

I'll give you a hint: There are zero on the market. The only possible contender is that new intel pro-audio onboard thing, but NO ONE has plans to impliment it yet.

I'm not sure if this is what you're referring to, but from the Product Brief (pdf) on the Intel Audio Studio page:

Dolby* Digital Live
This real-time interactive content encoder allows 5.1 audio streams to be transported over an optical connection to your digital speakers or entertainment center.

Also:

Real-time 5.1 digital encoding: Send Dolby Digital 5.1 (AC3) encoded content to your digital speakers or home theater over an optical connection.

Is this the feature you're looking for? If it is, then expect to see it implemented on many boards soon (Intel 925/915/910 chipsets). For now, note that the info page says: "Intel Audio Studio comes exclusively with Intel® Desktop Boards."

Re:meh

[Noehre]

Why do you care about it being 'through a digital interface'?

Its going to sound better through 6 analog connections. Compression isn't a good thing if you can avoid it, regardless of what Dolby marketing has led people to believe.

The only advantage AT ALL is saving a few dollars on cables. This might be a great advantage if you also buy $50 interconnects, but not if you buy cheap and equally good cables from companies not rhyming with Ronster Bable.

Re:meh

[Anonymous Coward]

What this means is, like any other sound card, is that your dvd will magicly send dolby digital 5.1 (ac3) content out. However with noramal software and games that use surround sound, your stereo will still see only a pcm 2 channel signal.
Soundstorm was the ONLY sound system that did real dobly digital out.

This is only of the reasons I have not bothered to upgrade to an athlon 64 yet.

Re:meh

[batkiwi]

That's exactly what I'm referring to. I can't find anyone with plans to stick it on a board.

Re:meh

[batkiwi]

Few things you're missing:

1. There isn't much/any compression on a real time AC3 feed, as you can use up the entire bitrate. It's not like a dvd where you have to worry about space constraints.

2. I can guarantee you that the D/A converter in my receiver is MUCH higher quality than the shitty ones in the sound card, not even counting all the electrical noise inside a computer case.

I've NEVER been able to get rid of that list tiny bit of hiss coming out of any sound card's analog output.

3. having 3 1/8" to dual RCA splitters plugged out the back, then 6 line level RCA cables going to my receiver, is NOT the way to carry a clean signal.

Remember, I'm mainly talking about an HTPC, going into a REAL sound system (not a shitty klipsh/whomever puts out a 300$ "all in one" pc system with tiny plastic speakers) used for watching movies and playing games on a big screen.

Re:meh

[SenorCitizen]

Yeah, you could just get an Audigy 2 ZS or whatever, but it's not the same AND it's another $80+. The nForce motherboards were about the same price as the VIA motherboards, but they had far superior integrated audio.

Well, yes, if the feature you were after was just DD encoding. But since the VIA mobos started coming with ALC655 (or newer) codecs that were bundled with the Sensaura suite, VIA's 3D sound support was way beyond Nvidia. Oh yes, I know, the APU does the same in hardware -- except that it doesn't. It's terminally broken, no amount of driver upgrades has been able to fix EAX support in games. Plain DS3D does work great however.

Superior ?

[matt-larose]

What the hell makes commercial closed source drivers superior ?.

I agree that nForce is supported well on linux, but its mostly because the sound and ethernet are handled by opensoure projects now. The stupid AGP gart is another issue. /me curses nVidia.

Re:Superior ?

[ben_rh]

And more to the point, in my experience Intel chipsets are the single best supported chipsets under Linux.

Every Intel-based machine I've owned has operated beautifully under Linux.

Re:Superior ?

[plupster]

What the hell makes commercial closed source drivers superior ?

They exist. ;-)

Re:Superior ?

[13Echo]

Hmm... I've never had problems with my "inferior" VIA chipsets. I was running a KT600 almost right after it was released, with both ATI and nVidia graphics cards. Benchmarks of the KT600 were practically the same of as a comparable nForce 2 at the time... And EVERYTHING worked with kernel-level drivers.

Re:Superior ?

[IntergalacticWalrus]

> I agree that nForce is supported well on linux, but its mostly because the sound and ethernet are handled by opensoure projects now. The stupid AGP gart is another issue. /me curses nVidia.

WTF are you talking about? There's a fully GPL driver for nvidia AGP in the kernel, and it's even made by nvidia themselves!!

Re:Superior ?

[JThundley]

I've have an nForce2 board (Asus A7N8X Delux) and some things have only worked in Linux for half the time I've had it. I couldn't get on the network at all until the reverse-engineered unstable testing hacked forcedeth driver came about in the 2.6 kernel.

Proprietary drivers a PITA

[xrepete]

Proprietary drivers are nothing but a PITA. They are totally unsupportable, and you have no idea how they will affect your system. Don't get me wrong. At least nVidia creates Linux drivers. However, until they create open-source drivers it is not something that I cannot have confidence in because if there are problems caused by the drivers I have zero support.

Re:Proprietary drivers a PITA

[molnarcs]

"Proprietary drivers are nothing but a PITA. They are totally unsupportable, and you have no idea how they will affect your system."

I have a fairly good idea how nvidia's binary drivers will affect my system. It makes neverwinter nights playable on linux/freebsd. It makes those fancy screensavers work. It provides 3d accelaration. etc.

This is sooo old, I don't even think where to begin. I know that _my_ ideas of what the binary drivers do are the layman's ideas. But do you have an idea of how [insert random opensource driver here] will affect your system? Do you read the source code of every opensource driver you happen to have on your install? And even if you do, how many of us does that regurarly?

Unsupportable ... by whom? Ah, I know, opensource developers, cause they are proprietary. But is that really so important, or is that really a problem? This can only be perceived as a problem if you automatically assume that opensource devs could provide better support ... because... well, why? Because they are opensource devs, and therefore are better? Because of the many-eyes argument (which was never really proved, because we tend to automatically assume that someone is looking at the code, but this someone often proves to be hypothetical). Nvidia drivers are not unsupportable ... they are just supportable and supported by its developers - just like most opensource software is (even though there is the possibility for Joe Random to begin supporting software X, it rarely happens). And they do an excellent job at providing regular updates, bugfixes, improvements... they even do a better job of supporting their software than some open source projects.

Re:Proprietary drivers a PITA

[Anonymous Coward]

I'll get modded down for this, but oh well...

"Do you read the source code of every opensource driver you happen to have on your install? And even if you do, how many of us does that regurarly?"

if you don't have the source, you can't fix it yourself. How the hell do you think most kernel dev's got to be kernel dev's? By fucking intuition? No, asshole, they had the source code.

nVidia for my video, not my chipset

[dlZ]

I've worked with the nForce chipsets before, but for my AMD needs I still prefer VIA. And on the Intel side, it'll take me a lot to sway away from Intel boards with the Intel chipsets. I'm mostly working with Windows machines on the Intel side, and the Intel chipset has never given me an issue, unlike SIS and even Via.
The average machine going out of my shop is going into a home with no real PC experience. I have to worry more about spyware than maybe some odd chipset incompatibility.

Re:nVidia for my video, not my chipset

[MarcoAtWork]

as much as we agree on going intel chipsets for intel CPUs, why do you prefer via to nforce on amd? I've always been under the impression that recent nforce products are much better than via's offerings in features *and* compatibility/stability

Re:nVidia for my video, not my chipset

[ilyanep]

nVidia cards are better supported than ATI in Linux, and they kick butt in Windows (imo). However, there's no way ever I'm using an nForce chipset.

Re:nVidia for my video, not my chipset

[dlZ]

My home machine is a Athlon 64 3200+ running on an MSI board with the VIA KT800 Chipset. I've had nothing but amazing luck with it, under XP and Fedora. But it seems anytime I deal with the nForce with the 64 chips, I see nothing but stupid problems. Mind you, these are computers brought into my shop, and many times they have low rated motherboards (usually home builds from friends and family who knew enough to make it tick, but not enough to make it tick properly.) On the other hand, they seem okay with the XP/Semperon chips. I just prefer a particular MSI board here that happens to be VIA for them.

Re:nVidia for my video, not my chipset

[Anonymous Coward]

Funny, I've had the exact opposite experience. Some of the early VIA Socket A motherboards had a serious problem with their PCI bus. This resulted in stuttering sound, flaky network, and corrupted filesystems. I know this firsthand, having experienced this with a Soyo Dragon+ motherboard. It was my first PC build, and it was a nightmare. Finally I figured out it wasn't my fault, bought an nVidia nForce board - and everything worked smoothly.

YMMV

Re:nVidia for my video, not my chipset

[dougsk]

You me the same via chip that is so horribly broken to be totally unusable for importing video over firewire on the pci bus? I'll give you a windows gamer box were the only thing that isn't integrated is the video card, but for huge i/o along the pci bus, I'd run far away. intel, sis,amd in that order.

Nvidia's Linux support superior to Intel

[EvilGrin666]

Since when? AFAIK Intel publishes its sepcs and Nvidia doesn't. Hows that superior exactly? Granted Nvidia release drivers, but there performance and features pale in comparison to the windows version or indeed similar Intel hardware with open source drivers written from the published specs.

Re:Nvidia's Linux support superior to Intel

[Anonymous Coward]

I'm not really sure how you can have superior linux support with closed binary drivers. The submitter probably doesn't know that all new and relevant Intel hardware made has an open source driver. Yes, even the Centrino wireless chips.

Re:Nvidia's Linux support superior to Intel

[AstroDrabb]

but there performance and features pale in comparison to the windows version

Just making things up to try to prove a point? The NVidia chipsets work out-of-the-box under Linux with Open Source drivers, no binary drivers required. For NVidia graphics cards, NVidia uses a unified driver that shares most code cross-platform. The Linux and Windows NVidia drivers are feature for feature compatible. The performance numbers are all about the same. I get just as good FPS under Linux with NVidia that I do on the same hardware under MS Windows. Even the installation is about the same. For Windows you run the executable and reboot. For Linux you just log out of X and run the executable without a reboot.

Re:Nvidia's Linux support superior to Intel

[EvilGrin666]

I don't think anyone has done any benchmarks comparing Linux vs Windows performance of the nForce chipsets. However on the graphics card front there is this Anandtech article using Doom 3 and Nvidia graphics cards. Although the article is primarly concerned with image quality, they do note a 25% performance drop in several configurations. Thats hardly "just as good FPS" now is it? IIRC Nvidia claim they use the same driver core for their graphics card drivers too. If the anandtech results are anything to go by, it doesn't bode well for any chipset benchmarks anyone might try.

No, my main beef with the binary drivers is that firstly they are binary drivers. It kind of goes against the grain of the distro I use (Gentoo) that I have to install a binary blob to make my motherboard work respectably. It doesn't help with the compatability either, remeber the whole 4k stacks saga for instance? What happens next time? we have to wait several months again for nvidia to fix it?

Granted there are opensource drivers avalible in the offical Linux tree now, but this begs the question why? If the nvidia drivers are as great as you profess why does forcedeth even exist?

As to features, where's my 6.1 audio? where's my Gigabit Ethernet?

Re:Nvidia's Linux support superior to Intel

[Anonymous Coward]

You are missing some important stuff regarding your comments on the graphics.

"Nvidia claim they use the same driver core for their graphics card drivers too." Ok then, if that's true, for what system was this core developed? Does it behave identical on different operating systems?

You also use Doom3 as a benchmark, well, guess what system Doom3 was developed for... that could perhaps explain your 25%.. Point is, just because some app will run on one OS, doesn't mean it will work as well as on the one it was designed for.

Re:Nvidia's Linux support superior to Intel

[EvilGrin666]

"Ok then, if that's true, for what system was this core developed? Does it behave identical on different operating systems?"

I guess windows and yes being the answers, however I'm not nivida and I can't tell without looking at the code. Which was the original point wasn't it?

You want some more application benchmarks, ok then.

Re:Nvidia's Linux support superior to Intel

[AstroDrabb]

95% of NVidia's code is shared across Linux, Windows and Mac. That Doom 3 performance drop is do to the Doom 3 code, not the NVidia driver. I read that in one of the Linux Doom3 vs. Windows Doom3 comparisons. The Linux Doom 3 code is not as optimized yet as the Windows Doom3 code. Again, this has nothing to do with NVidia which as I stated shares 95% of the code between different platforms. That other 5% is different due to the way the code needs to interface with the kernel/graphics system.

Granted there are opensource drivers avalible in the offical Linux tree now, but this begs the question why? If the nvidia drivers are as great as you profess why does forcedeth even exist?

You answered your own question. Some people don't want to use a binary driver. I personally use the Open Source drivers for my NForce mobo and I use the binary driver for only my NVidia grapichs card.

Note that I am only talking about the graphics drivers as I don't use the binary chipset drivers. I will agree that NVidia was dragging on that and they didn't work with my newer kernel under Fedora Core.

Re:Nvidia's Linux support superior to Intel

[mp3phish]

"NVidia drivers are feature for feature compatible"

BZZZZZZT!!! Wrong answer.

You know as good as everyone else here that this is just completely falsified and made up. But good try.

Maybe with the 6-12 month lag time between nvidia and windows drivers, if you can jump forward in time, this would be true. But it isn't.

Re:Nvidia's Linux support superior to Intel

[AstroDrabb]

BZZZZZZT!!! Wrong answer.

Wow, did you think that up yourself?

You are wrong. The NVidia Unified Driver Architecture (UDA) shares 95% of its code base between all supported operating systems. That other 5% obviously needs to be different because each OS has a different way to interface with the kernel/graphics system.

Re:Nvidia's Linux support superior to Intel

[mikael]

The Linux and Windows NVidia drivers are feature for feature compatible

Explain this problem question from nvnews.net...

Re:Nvidia's Linux support superior to Intel

[AstroDrabb]

See my post here. NVidia uses a unified code base that shares 95% of the code across Linux, Windows and Mac. That other 5% is for differences with each OS and how to interact with the kernel/graphics layer. I didn't write the NVidia drivers so I cannot tell you what those extensions do. However if I were to take a stab at it, I would say they deal with GL things that are specific to X under Linux?

Re:Nvidia's Linux support superior to Intel

[mikael]

I checked up on this. You have to download and run 'nvemulate' to switch on the extra extensions. For early cards, the emulation is done in software emulation. For later cards, this makes the extensions visible. In 6xxx cards this is done automatically, but anything below this has to be done manually - ugh!

Not too hot for AMD.

[SnowCrashed]

The NForce boards were really one of the best things about AMD's chips for people who wanted a good all in one mobo for a basic system (esp. the NForce2 with the onboard GF4mx which totally rocked Intel's "extreme" garbage). While AMD rigs will still probably be cheaper, this is quite a low blow. However, maybe it won't make much difference with ATI's XPress chipsets coming out. They like they will be the ultimate solution for a while, at least until newer things come.

Re:Not too hot for AMD.

[gid]

While nforce is popular, it's not really vital. They no longer have the good sound, and the onboard video isn't an option anymore, which pretty much makes it just like any old chipset now, except it's a bit faster, but quite a bit more expensive. Granted the onboard video wasn't anything special, but it was more than fast enough for normal 2d usage and casual gaming, as long you don't expect to play hl2 or doom3 at a decent frame rate. I ended up going with a AMD64 system, Asus K8V SE Deluxe, (K8T800 based) over an NForce option, because it was cheaper, had linux compatibility, and the disk performance was benchmarked a few places to be better. Which really is more what I cared about than squeezing out a few extra cpu cycles.

Of course this is mostly my opinion over what I've read, but my machine's rock solid and haven't had a single problem with the sound or onboard lan. And it works like a charm under linux and xp with no extra special drivers (xp sp2 did need extra special drivers actually). Just use the onboard promise sata controller, not the onboard via, because the via has a special feature where it doesn't blink the HD led.

Re:Not too hot for AMD.

[Jeff DeMaagd]

For a long time, the big driver for Intel v. AMD was that AMD CPU chips were often cheaper than Intel's CPU chips. The price differences of the boards they go onto wasn't so wide, but I hadn't looked at board prices lately.

To be honest, I'd prefer to stick to an AMD chipset for an AMD chip, and an Intel chipset for an Intel chip. Other than price, I'm not sure what nVidia has going for it, especially with the demise of SoundStorm.

Does anyone know why SoundStorm is gone? That would seem to be nVidia's biggest chipset selling points.

Re:Not too hot for AMD.

[Anonymous Coward]

>Does anyone know why SoundStorm is gone? That would seem to be nVidia's biggest chipset selling points.

It was a good selling point among geeks, but not too many other people cared. Motherboard manufacturers saw this and decided since no one cared it wasn't needed in the boards. The manufacturers axed it. nVidia just finished it off by not requiring it in nForce3/4.

Thankfully, I have a Compaq branded nForce2 with SoundStorm and I'll have to double check my other home-built tower -- it might have it too. Good on board audio is nice! Now if I just had hardware mixing in Linux....

INSIDE INFO: Nvidia will not succeed.

[Anonymous Coward]

I have a friend who actually works at Nvidia. Like many companies headed by a Chinese, software development is a hodge podge. There is no documentation. The simulator for each new chip often crashes, leaving a "core dump".

Does this story sound familiar? It should. It is the same story that plagued Sun Microsystems before it crashed in flames started by the UltraSPARC III.

Eh?

[MiniChaz]

"NVIDIA's superior Linux support"

Are you on drugs? Since when did binary only modules constitute "superior Linux support"?

Re:Eh?

[NewStarRising]

I think it refers to superior to ATI's Linux support .
I have had major troubles getting my ATI card to run correctly under Linux, and from trying to find support on it, have found I am not the only one .. the main advice i get is "Use a Nvidia card".

Re:Eh?

[pe1rxq]

I think most people (the article submitter included) still can't distinguish between drivers for the nforce chipset and the graphics drivers. They just hear 'NVIDIA' and start screaming....

Jeroen

Re:Eh?

[MikeBabcock]

As the poster of comment 10874482 pointed out, Intel chipsets are very well supported in Linux.

Does this mean

[Anonymous Coward]

Does this mean NVIDIA won't continue to produce the NForce boards for AMD?

If NVIDIA is going to continue to produce boards for AMD, what's the problem?

I'll wait and see how much AMD support suffers from this (I'm hoping it doesn't suffer much...)

Chipset drivers

[Derf_X]

Am I te only who thinks it's not normal for chipset to require drivers to be installed. Every chipset I tried works out of the box without drivers needing to be installed, be it on Windows or Linux.

Iwas very surprised the first time I did troubleshooting on my friends computer who has an nForce chipset. I was actually troubleshooting his GeForce video card (in Windows) so I uninstalled the drivers, but I actually uninstalled the chipset drivers because I didn't know the chipset required some. Same when I installed Linux on his computer: I knew the videocard required drivers but the chipset?!?!

This is ridiculous! It conviced me never to buy a motherboard with an nForce chipset. Now if I could only find a videocard with open source drivers that support accelerated TV-out under Linux...

Re:Chipset drivers

[Slimcea]

Speaking from experience (Windows XP), the chipset doesn't really require drivers per se. The only thing that becomes non-functional when you don't use the nForce drivers is the on-board LAN (which I assume XP doesn't have built-in drivers for).

The chipset drivers supposedly improve overall performance though, although there have been niggles with the IDE drivers at one point of time.

Re:Chipset drivers

[Jeff DeMaagd]

Intel chipsets often don't need drivers but that's because basic drivers for Intel chipsets are usually included with Windows.

With my current computer, Windows 2000 did not have an AGP driver, so my AGP video card was running in PCI mode. Chipset drivers also enable performance features. IIRC, Windows defaults to PIO modes, chipset drivers allow users to enable UDMA.

Now, I'd probably never use a chipset with built-in graphics.

Re:Chipset drivers

[Anonymous Coward]

Not normal? People have been doing it for years... Why is it any different than any other device? Using chipset drivers on a 440 BX/LX/GX/etc running Windows98 did wonders for its performance and such.

Windows XP came out in 2001... the Intel 865PE came out after that, so unless it has no new functionality accessable by software, its going to need drivers... Installing chipset drivers is great considering the chipset is arguably the most important component of a computer.

What support?

[Anonymous Coward]

What is this? "NVIDIA's superior Linux support".

Is this how they describe NVIDIAS's lack of information sharing? Is it a joke?

Best tool for the job

[bug1]

The "best tool for the job" depends on the timeframe your looking at.

In the short term the easiest method is the best tool for the job as you cant justify overheads of learning a new tool.

In the long term retooling costs are insignificant and the best tool for the job may even involve making a custom tool for yourself.

Software Freedom enables long term solution, plans to be made in regard to maintenace and development to ensure long availablility and reliability.

Closed source software is a consumable, its not reusable, its long term maitanance and reliability is beyond your control, its nothing more than a short term solution.

Free Software should be seen as a form of infrastructure that indirectly benefits all of society.

Re:Best tool for the job

[shutdown -p now]

This "short term solution" has been unavoidable for many activities for last twenty years. In this case, until before Open Source community comes up with its own "free" hardware, I don't see how we can possibly get away from using those "crappy graphics cards", for instance, and proprietary drivers for them.

Re:Best tool for the job

[nomadic]

The "best tool for the job" depends on the timeframe your looking at

I'd like to play PC games, and my timeframe I'd like to do is "before the next ice age".

Re:Best tool for the job

[cowbutt]

I'd like to play PC games, and my timeframe I'd like to do is "before the next ice age".

Whereas I don't really care about any games that came out after 1999 or so (hence my Radeon 7500 still performs fine for me), but I /do/ care about being able to use the hardware I've paid for for as long as I want to (rather than as long as the hardware vendor doesn't want me to buy new kit) and I like having the fallback position of being able to self-support if there are problems. So I try and by documented hardware supported by Free drivers these days.

Horses for courses though.

--

...but it might not be the chipset that matters...

[codefreez]

I just read an interesting article last night that claims it is not chipset that matters so much, because the Intel CPUs stick to the traditional north/south bridge design that limits I/O, while AMD64 processors have multiple hypertransport interfaces on-chip. http://www.samag.com/documents/s=9408/sam0411b/041 1b.htm

Doesn't matter

[SQLz]

Intel is still missing HyperTransport and on chip memory control. Who cares what chipset the motherboard is, its an inferior platform.

Re:Doesn't matter

[Anonymous Coward]

AMD is missing hyperthreading, making AMD systems not as responsive on the desktop. Also, having an on-chip memory controller is not necessarily a good thing. You can get an intel system that supports dual channel ddr with only spending ~$100 on the processor, while with AMD you'd have to spend over $500 on an AMD-64 FX. I love AMD, but blind zealotism will get you nowhere.

Re:Doesn't matter

[doublebackslash]

Ok, I'm responding to an AC, but I can't resist.
Hyperthreading can SLOW DOWN your computer. Why? Because the two tinstructions that are supposed to execute similtaniously aren't allways on the same set of data, due to poor scheduling. This means that data is 'thrashed' in and out of low level cache to higherlevel cash and main system memory. Even with dual channel support on cheap chips it dosn't make up for the fact that their processor can't keep up with AMD's on operations (read: they could, in theory, get more data to the cpu, but can't do much with it fast enough), and as far as hyperthreading making the desktop more responsive, well, this MAY be true, in similarly performing chips, but save the cash from an intel HT chip, and buy a higerperforming AMD at the same price (most likely a 64 w/ hypertransport) and your latency issues will go DOWN.
Intel hasn't made a decent chip since the P3, and I never even bothered to invest in a p4 because of the price to performance disparity between AMD and intel.
I don't expect intel to be the lagger for much longer, but they have a lot of work to do to get the price down, and the performance up (across the board, not just in synthetic benchmarks).
Thats just the way I sees it.

--
There are more things in Heaven and on Earth than attempted in our philosphy...

Re:Doesn't matter

[bersl2]

I think you got your Hyper* products mixed up. The AC said nothing about HyperThreading. He said that since Intel does not employ HyperTransport and the on-chip memory controller (you missed that), he will not use their parts.

Re:Doesn't matter

[Wesley Felter]

You can get a Socket 939 processor for $150, not $500. FUD will get you nowhere.

Re:Doesn't matter

[Anonymous Coward]

I've gone with Intel and AMD quite a bit...I had a dual celeron box (BP6), and an athlon/duron setup too..

Although AMD definitely gets the performance crown, I went with Intel, because they have always done better as far as heat issues... the house I live in gets HOT... and the athlon boxes (there are 4 of them) just get too hot and crash a lot... now these don't have Cool'n'Quiet because there a bit older, but still.. AMD's chips are just terrible with heat.... but for me, stability is far more important (and Intel chips are still pretty fast) than raw performance.

I fried one of my athlon chips on accident because I turned the fan too low... it was still blowing air, but not enough and the chip hit over 200F... thats crazy... a Pentium 4 will just downclock itself so it doesn't crash as easily.

If AMD improves in this area I might switch back to AMD... Which brings up a question:

From what I understand, the new Cool'n'Quiet technology that the Athlon64 and Semprons has is supposed to throttle down the cpu when its not being used at high demand, but what happens if your doing something like RC5 or SETI or whatever... will it downclock when it overheats like P4? I haven't read anything one way or another, and I don't know anyone with an A64.

Re:Doesn't matter

[SQLz]

Blind zealotism? $500? A decent Socket 939 chip is like $150 to $200.

The HyperTransport bus and on chip memory controller and the ability for the Operons to have seperate banks of memory as well as 2,4, and 8 way systems implemented on the chip itself, not being controlled by the mobo chipset, is huge. Its why you see dual and quad opteron systems making a mockery of Xeons in Database server performance.

Hyperthreading does nothing unless the application itself is multi-threaded and while many are, the gains or losses you see are not even noticable the naked eye.

"screen frozen, but mouse pointer moves" bug

[zby]

The drivers NVIDIA produces are not the best quality: http://www.nvnews.net/vbulletin/showthread.php?t=3 1858

Would it be better if it was all open source? I don't know, but possibly with more eyes the bug would be eventually find?

Re:"screen frozen, but mouse pointer moves" bug

[linefeed0]

Fucking A -- that's exactly the bug I have all the time with my (old) nvidia card, and I was beginning to suspect bad hardware. It is actually somewhat reproducible when loading very large images with mozilla or xpdf, too.

I was willing to give nVidia the benefit of the doubt since they fixed other bugs (like the infamous kernel oopsing bit back in the 3xxx series), but that does it. This bug has been around for fucking ever. My next video card will be a Radeon 9200.

Re:"screen frozen, but mouse pointer moves" bug

[Wolfrider]

--Tell me about it. Up till now I haven't had a problem, but their latest Linux driver ( NVIDIA-Linux-x86-1.0-6629-pkg1.run ) breaks compat with my PCI TNT2 (Guillemot) card. X freezes so hard that you have to reboot to get the display back. SSH still works however; but there is a huge slowdown on all processes, even during ' reboot '.

--I'm about to yank teh stupid card out and just use the mobo Intel at this point; Nvidia's driver has been broken since November 5!

Re:"screen frozen, but mouse pointer moves" bug

[Wolfrider]

( BTW, this is using the latest vanilla kernel 2.6.9, so backlevelling the driver to NVIDIA-Linux-x86-1.0-6111-pkg1.run doesn't work! )

Re:"screen frozen, but mouse pointer moves" bug

[zby]

the linke once again:
"screen frozen, but mouse pointer moves" bug

Bragging rights

[Lando Griffin]

This means that the bragging rights AMD users have had about having the superior nForce chipsets is about to end...

Maybe so, but we still have cheaper and better CPUs.

whats wrong with intels chipsets?

[Spellbinder]

IMHO amd users are those with the crooked chipset and the intel users had the good stuff (even linux wise)
no offence i use amd myself but i feel that when there is a place where intel is superior it is with their chipsets

So...

[Dragoon412]

It's fairly well-understood that these days, the Athlon 64 is utterly dominant in terms of sheer performance and price/performance. At least in the gaming market, which is the nForce boards' target market. It seems the only reason Intel is even still competetive in the high-end home PC market is due to uneducated users buying from the likes of Dell and Alienware, and their success at branding themselves, which, as of late, seems to be falling by the wayside.

So while I don't think this is a bad thing at all, I think the gaming community, specifically the enthusiast builders as a whole (who actually care what chipset their system uses) will have a reaction along the lines of "Meh..."

So, I doubt this'll have much of an effect on anything. Enthusiasts are buying AMD, and the uninformed will keep spending money on Dells and the like regardless of who made the mainboard.

Re:So...

[Anonymous Coward]

Meh

dear Nvidia, you can just dream about it!

[ratta]

Unluckily, if you want to have good 3D accel on linux you are almost forced to use an Nvidia GPU. But luckily there are open spec chipset (es Via), so if you @nvidia think that i will use a binary only module just for the MB, well, this is unthinkable, you are just kidding!

nForce superior?

[EvilIdler]

I'm not so sure the nForce chipset is superior, at least not the nvnet
bit. I've tried both the NVidia driver and the reverse-engineered
forcedeth driver, but neither lets me use the network at all.

From what I can tell, I am not alone, either. Google has many stories
involving non-functional nForce 2/3 integrated NICs.

Re:nForce superior?

[mAriuZ]

works for me out of the box (debian sid)

here are the machines specs

http://http//forums.amd.com/index.php?showtopic=69 2/

Re:nForce superior?

[Anonymous Coward]

Dont most nforce motherboards have dual network cards. you might be plugging the cable into the wrong one.

Also, nVidia dropped nvnet drivers from their driver package because forcedeth was better and they didnt want to have two different drivers that offer the same performance. They've also been contributing to the forcedeth project fixing bugs and adding missing features.

looking at benchmarks between the via chipsets and the nforce ones, nforce comes out on top in almost everyone.

Re:nForce superior?

[EvilIdler]

All Deluxe cards have dual, yes, and the second NIC is a wonky 3Com
variant that still is problematic. The nvnet driver is still in NVidia's
latest downloadable nForce kernel drivers, too.

Next?

[deconvolution]

Nvidia have handled video chipset and motherboard chipset -- is "nvidia inside" coming?

If you're so ashamed of your product...

[Anonymous Coward]

Then make it better! Refusing to release specs just screws your own users. I'm a long-time Measurement Computing customer, and those guys are producting such crap now, they're embarassed to release specs on their new equipment. We've got over 400 of their USB ADC's that we can't use since they're so ashamed of their hardware. Come-on, if it's that bad, then work to make it better. Don't just continue to screw your customers. nVidia is in the same boat. They know their hardware is crap so they know they'll hurt their sales if they show the world just how bad they really are. It's a difficult position for a company to be in, but the only way you're going to get better is to admit your problem.

My take the deal

[TheRealBlueEAGLE]

The real question is wether or not this agreement will make nVidia (graphics) more DirectX orientated or not. Microsoft and Intel has had a long history of aliances, and now that nVidia (company) couples with Intel there is a real possibillity that it will make them more DirectX-orientated. This is ofcourse a BAD thing for us linux users.

As for AMD it will see their partner nVidias chipset get a bigger userbase. This ought to lead to a drop in overall cost of the chipset wich will make both Intel and AMD mobos cheaper. (Intel mobo-chipsets will drop in price simply due to more competition).

Intel might see a takeup of their mobos and processors becaues many people I know, myself included, have had trouble with both Intel and VIA mobos for the intel processor before. If they can get a decent (rad: nForce :)) mobo for a fair price (and if Intel could/would cut abit in the price for their processor) they might see a big gain in sales of both chipsets and processors.

So who is the big (long term) looser here? It all depends on how it plays out really. If nVidia gets the bigger edge with gains in sales, it might strengthen the support for OpenGL (since nVidia has got a strong OpenGL focus) and more games will probably be made with that library as it is the best supported from a big gfx company and it will make it easier to port games to linux. This means that Microsoft (DirectX stands to loose ground and with more games running on linux many gamers migth start using linux (Then again most gamers I know pirate windows, so it wouldn't influence the bottom line in windows sales)) and ATi (better with DirectX, porer on OpenGL) will loose ground. This is good news for linux users since DirectX is poorly (understatement?) supported in linux due to you-know-who.

How ever if nVidia gets influenced by the Wintel-alliance and starts focusing more in DirectX and less on OpenGL then Microsoft will win because their library will have supperiour support from both major GFX players and OpenGL might dwindle into a the murky shades of history.

As for more mobo-spesific components. I have yet to experience any unfixable problems with my NF7-S (nForce2) mobo on linux. The biggest problem was finding the right ethernet (100MB) driver in the list. As for the 4kb stack, it was well documented on the web when I hit it. Then again I don't recompile the kernel just because a new version has been released. I recompile if I need new features or to fix any security issues.

So it looks as only time will tell what this means for the end user.

Finally I will put up a few words about releasing specs vs releasing binary modules. I can see why nVidia might want to keep their property to them selves. This is all well and good since their driver architecture is as good as it is. Since I use Gentoo I am comfortable with compiling that little stub that needs to compile to get the drivers working. For people not familiar with the consept (those asking "How do I execute a .bin file?" all the time) it might be more of a hassle. Atleast I get to use the latest xorg with my nVidia graphics card.

Wow, are you still reading this post?? I am honored. :)

Re:My take the deal

[Anonymous Coward]

I think it would be foolish for nVidia to move into DirectX. They have firmly positioned themselves as the OpenGL leader, letting ATI take the DirectX crown. They have also proven to be a reliable(if controversial) supporter of Linux.

I think the real change is on Intel's part. They have been cornered by AMD for at least the next year, and they're also recognizing growing support for open-source in general. A deal with nVidia is a way for them to gain more ground in the Linux world, and possibly a move towards a marketing push in future years:

"Game on an Intel Linux system with nVidia graphics!"

It could happen, and it's in their interest to be ready when it does.

Re:Bragging rights - oh yea?

[StormyWeather]

Have you been to the store recently? I can't really say that the AMD chips are the cheaper of the two anymore from my price comparisons. It seems that AMD has decided that they are good enough not to take the second tier on pricing anymore. While this is probably good for AMD, I wish they hadn't done this, because I was always happy to stay a bit behind on speed and get a chip for half the cost allowing me to upgrade more often and have a herd of old computers to play around with. That way I never had the fastest chip in the herd on my game machine or development box, but I always had something in the top percentile.

I won't get into who makes the better chip because I honestly don't know =). I have AMD on my windows laptop, windows game machine, and linux developer workstation. I also have a linux server running on a broken scren laptop using an Intel, and my M10000 for a linux multimedia center in the living room. I find them all to get the job done nicely.

Wonder how NVidia swing this deal

[Anonymous Coward]

If I remember correctly, when NVidia was negotiating with Intel for the first time, it didn't go well. Intel wanted every NVidia MB shipped have branding of "Intel Inside". NVidia at the time didn't want to share that so the deal was off. That was why NVidia had to turn to AMD.

I wonder what changed NVidia's mind? Was it because Intel has always been king of integrated gfx chipset and NVidia has finally caved to Intel in exchange for that piece of the pie?

I see this deal would only further Intel's CPU attach rate. Also, it would also be interesting in seeing how Intel's gfx chipset division fair against NVidia. Either way it's a big win for Intel's CPU division.

NVIDIA SUPERIOR LINUX SUPPORT?

[7-Vodka]

OMFG, Go FUCK yourself and your 'nvidia superior linux support'.

Superior linux support would be, you know, if they actually released the specs to their hardware AND organized and contributed to GPL-licensed drivers for everything AND publicised this and encouraged linux use. THAT would be superiour linux support.

NOT releasing the specs to their hardware and just releasing buggy, outdated, *blackbox* drivers with no source code is the federal pound-you-in-the-ass version of linux support.

Re:NVIDIA SUPERIOR LINUX SUPPORT?

[Anonymous Coward]

Not everything should be open source...there I said it.

If NVidia released their software drivers freely on the internet, it would screw them over.... secrecy has its place... ATI and others could take their technology or even get new ideas from it and make their products even better without releasing anything back... The fact is, NVidia's linux graphics drivers (although I haven't used them in a while) are awesome... better than the others... because they aren't open source doesn't make them any less great. It's nobodies business but NVidia's whether or not they release them or not.

The longer NVidia has a technological edge the longer they stay in business... but if they release all their secrets they lose that edge... its really that simple.

Oops

[bersl2]

I need to stop posting when I first wake up...

you guys don't pay much attention, do you?

[ruiner5000]

If you had then you would have read my report on Nvidia working on Linux system utilities, and continuing to improve their graphic drivers particularly focusing on DCC. Yeah, Nvidia can not open source their drivers due to licensing issues. What are they supposed to do? I think I will take the GiGE, advanced SATA RAID, advanced firewall, and best in class performance. Yeah, we are all pulling for Soundstorm to come back, but because the motherboard vendors didn't want it we lost it in nForce 3. You really need to learn more about what is going on with Nvidia before you criticize. I suppose most of you slashdotters complaining in this thread aren't paying attention.

Forceware ported to Linux is good.
Support in the Kernel is good.
Support for 64 bit in Linux is good.
Support for FreeBSD is good.
Advanced SATA RAID far past what Intel has is good.
GigE superior to Intel or any other chip maker is good.
Hardware and software firewall superior to what Intel or any other chipmaker has is good.

Yeah, real big deal having to install closed drivers, and miss out on those features. I think plenty of people will, and Nvidia is the preferred solution for AMD users running Linux exactly for these reasons. Is Slashdot behind the times?

Re:you guys don't pay much attention, do you?

[Wesley Felter]

Advanced SATA RAID far past what Intel has is good.

It's all software RAID; who cares?

GigE superior to Intel or any other chip maker is good.

Is the nVidia GigE really better than e1000? And even if it was, would anyone notice?

Hardware and software firewall superior to what Intel or any other chipmaker has is good.

I'd rather keep the firewall in the OS where it belongs. If I used a firewall at all, that is.

Re:you guys don't pay much attention, do you?

[ruiner5000]

No, it is not all software raid. Yeah, plenty of people do. Wow, so software firewalls are good, but software raid is bad now? Well I would think if you were deploying a load of systems in enterprise you would like the good ethernet, and firewall, not to mention unified drivers. But that is why no one uses nForce amongst the linux community right?

Re:you guys don't pay much attention, do you?

[k8to]

I think the point is not really that software raid is bad, but that software raid is pointless since Linux already does that just fine thankyouverymuch.

The idiots at Measurement Computing

[Anonymous Coward]

We designed a product around their PMD-THERM. They screwed us hard after refusing to release specs. I guess you're right that they're so embarrassed by the crap they sell now that they don't want to release specs. We're scrambling now to find a different solution. Our largest and oldest customer is depending on us to get something working by Christmas vacation when they'll have everything shutdown so we can install new monitoring systems. After getting screwed and wasting so much time with Measurement Computing, I don't know if we can do it. They might have cost me my job.

That company is doing so badly that they've removed contact info from their management team's web site. The only e-mail address I now have is for a guy named Ilya Tatar. All he'll tell us is that he isn't allowed to help customers. WTF good is he?

Bendrix Bailey

[Anonymous Coward]

is the problem as I understand it. I've used Measurement Computing's products for a long time, even before when they had to shed their old name out of shame and left town. They used to be known as ComputerBoards before they moved out into the woods. They were founded by using stolen schematics & board layouts from MetraByte. They used to release specs, but that got them in trouble considering the work they sold as their own, simply wasn't. I guess they learned their lesson. If you're going to steal, don't make it so easy to get caught. I wonder how many of the other companies that only release Windows drivers are also thieves like this.

Re:Because NVIDIAs work better than ATI

[Elladan]

"The driver" for a Radeon 9500?

The driver for a Radeon 9500 is a closed source binary module, just like NVidia's is. ATI won't release any specs either on cards newer than the 8500.

Re:...but it might not be the chipset that matters

[Wesley Felter]

Pretty much all the chipset reviews on the hardware sites confirm this; the different in performance between chipsets is a few percent at most.

FUCK OFF

[msimm]

Mod me whatever you want, but I'm am sick to death of hearing OSS users whine about closed source drivers.

Nvidia has excellent support for Linux whether you like their policy or not. If you want opensource drivers you can reverse engineer them your goddamned self. But to be honest, if your that much of a OSS zealot you should start by engineering and manufacturing your own OPEN HARDWARE PLATFORM and stop worrying about what Nvidia is doing.

No one owes you anything. OSS is a choice.

Re:Bragging rights - oh yea?

[Dr. Spork]

Your post makes no sense. If you prefer to stay behind the cutting on speed, you should not buy from the top of AMD's line. You don't need to be an economist to figure that out. The fact that AMD has something better than what you (or I) can afford should have nothing to do with your decision about what chip to buy. The question is: can you get identical performance for less from AMD, or, better performance for the same price? And I think the answer to both questions is a pretty clear "yes". The parent suggested that you can get both at the same time, a better price and better performance, and honestly, I think this is right. For example, if you look at the 3500+ I'm almost sure there are more expensive Intel chips that don't perform as well.

The one reason I get Intel CPUs is Intel chipsets

[wernst]

It seems to me that the mainboards I regularly service and the systems I maintain tend to be less trouble-prone when they have Intel chipsets. I'm not saying that they are trouble-free, but supporting an Intel board is far easier in terms of downloads and drivers and compatibility than VIA and other chipsets, especially after a few years pass. That's why my last two personal PCs were made with Intel CPUs -- I can drive them with a mainboard with an Intel chipset.

Remember when you could use an AMD CPU in a board with an Intel chipset? The last time I could do that was with an Intel TX-chipset equipped board and an AMD K6-3. That system is STILL running just great for me, even 5 or 6 years after it was made. I don't generally see similarly aged VIA (or other chipset equipped) boards still working well.

If somehow I could get a modern AMD processor working in a modern mainboard with an Intel chipset, I would buy and build that system in a heartbeat.

So forgive me if I don't care that I can use anotehr chipset with an Intel CPU.

About fricking time. Who care about linux.

[gumbysworld]

About fricking time. Who care about linux.

I was totally sick about then dam Nforce boards and Intel users being left out. It was crap BS.

Only a dummy make a product for less then half of the market.

Oh and who cares about linux users??? I sure don't. I can care less if you get good drivers for linux. Linux is to small of a market for any company to bend over backwards for. They do how ever bend for us windows users.

It's people claiming "Superior Linux Support"

[anti-NAT]

I completely agree. I don't care what Nvidia do, if they don't provide open specs, I won't be buying their hardware. That is their decision and that is their business loss.

The issue people are talking about is the claims of "superior Linux support", how "wonderful" Nvidia for "supporting" Linux, when they are only paying lip service to the ideals (open specifications) that created Linux in the first place.

I'm not going to get into an argument about the "value" of open specifications, and how they "don't really matter because Nvidia provide wonderful support." I first ran Linux in 1992. Nvidia didn't exist at that time, nor did binary modules. I wouldn't have been able to run Linux on my hardware if open specifications didn't exist.

To put the whole open specification argument into a true context, the PC as we know it wouldn't exist either if it wasn't due to open specifications. The only specification IBM didn't publish openly of their IBM PC was the BIOS, although technically they did open publish it, they just retained copyright on it, preventing people from copying it. Compaq reverse engineered that, which made the IBM PC hardware specification fully open.

If you don't understand the importance of open specifications, then realise that the PC you are (probably) sitting in front of would be a very different machine if it weren't for them, and would have probably cost way, way more than you paid for it, as proprietory products always cost more. Even Apple users have benefitted, as Apple decided to use open, industry specifications such as PCI, SATA, USB, Firewire and a number of others, which reduced the cost of an Apple PC.

Re:It's people claiming "Superior Linux Support"

[sweede]

I thought Apple created USB, Firewire, SCSI and a couple others that we use in the PC world ?

Re:It's people claiming "Superior Linux Support"

[anti-NAT]

I think you might be right about Firewire aka IEEE 1394, although I think Sony had some significant involvement. I'm pretty sure USB was primarily created by Intel, and SCSI was originally created by Shugart Associates' Alan Shugart as "SASI". SASI was further developed by Shugart Associates and NCR to become the ANSI SCSI standard. (Alan Shugart then went on to start Seagate).

where's my real-time dobly digital encoding?!!!

[majid_aldo]

...damnit!!

I think you mean...

[BumpyCarrot]

... *AMD* gives *nForce* users bragging rights. My next system will be using something other than nForce, but I'll still be using an AMD CPU.

The nForce was designed for gaming, something that the AMD CPUs excel at. I'd rather use a VIA board with an AMD chip than an Intel board with an nForce, particularly since the removal of Soundstorm.

No soundstorm, no sale

[AbRASiON]

If it can't encode dolby 5.1 on the fly (they recently dropped support for this in all future chipsets) then I don't want it, period.

The primary reason their chipsets rocked they dropped, speed is good but that kind of audio is out there - nothing else does it.

I'm sure it was already said and the thread is a little old but that's how it is, at least for me.

fools