Again, We Have Another New NVIDIA Driver

[phoronix]

Phoronix: Again, We Have Another New NVIDIA Driver

Last month there were five Linux driver releases from NVIDIA and we have already seen quite a few this month, but just days after releasing the 180.51 Linux driver, they have pushed out another new release. This time around NVIDIA released a new beta in the 185 series, which is called 185.18.04.This new NVIDIA Linux driver is known to work around a bug on Red Hat Enterprise Linux 5, memory management performance problems in a Composited environment, compatibility with the Linux 2.6.30 kernel, and improved DisplayPort support when using suspend-and-resume.The release announcement and x86/x86_64 Linux driver download links for this very latest NVIDIA build can be found at NvNews.net...

http://www.phoronix.com/vr.php?view=NzIyMw

[a7v-user]

Seems like Nvidia's new way of releasing beta and half-tested (pre-releases) drivers are here to stay so maybe we shouldn't be so surprised about it. The headline for this article could perhaps have mentioned which set of drivers (beta, pre-release or stable) that has been updated.

Hopefully this will spur on the other team making closed-source unix drivers, or at least fend of any talks about cutting back on driver development.

[deanjo]

Seems like Nvidia's new way of releasing beta and half-tested (pre-releases) drivers are here to stay so maybe we shouldn't be so surprised about it. The headline for this article could perhaps have mentioned which set of drivers (beta, pre-release or stable) that has been updated.

Hopefully this will spur on the other team making closed-source unix drivers, or at least fend of any talks about cutting back on driver development.

Time for a RSS feed!

[Milyardo]

I think the real news here is that its a driver update not made of mostly VDPAU fixes :|

[deanjo]

I think the real news here is that its a driver update not made of mostly VDPAU fixes :|

Not really, pretty much every release had non VDPAU fixes as well. There were just alot of VDPAU fixes included as well as would be expected in a newly developed feature.

[Kano]

Now i only wait for updates for 71, 96 and 173 series for 2.6.30. Then nvidia would be ready for next kernel round.

[Saist]

I guess when you don't have a real open source strategy, like that other company, releasing constant beta updates of closed-source stuff is all you can do.

[deanjo]

I guess when you don't have a real open source strategy, like that other company, releasing constant beta updates of closed-source stuff is all you can do.

Who cares, it works. In the end that's what most end users care about.

[Dragoran]

Phoronix: Again, We Have Another New NVIDIA Driver
but just days after releasing the 180.51 Linux driver,

You missed 180.53

[Saist]

Who cares, it works. In the end that's what most end users care about.

ah. you again. Why did I suspect if anybody was in need of getting shot down, again, it'd be you.

To directly answer your question: LOTS OF PEOPLE CARE

AMD didn't just up and start dropping their hardware and software documentation because they wanted to. They opened up their development tools because customers and end-users kept asking for literal years for access.

By the same token, while Intel's initial Open source 3d graphics code dump was largely a response to AMD buying up ATi, customer and end-user demand has seen a large quantity of their own code mature under open-licenses, not only their graphics chips, but their wireless cards as well.


To directly answer your implied statement, no it bloody doesn't work

There's a reason Nvidia lists half the stuff as beta. It's untested in a way that make's AMD's limited selection of tested distributions seem like a thorough and exhaustive test.

While I applaud Nvidia for putting a lot of their proprietary stuff up for end-user feedback before they finalize a driver set, it's nothing more than attempt to stay relevant when both competitors (Intel and AMD) are largely fostering genuine end-user built driver sets.

It's like a gilded cage. Yeah, it looks pretty, and might be pleasant to live in, but it's still a bloody cage.

Which leads into the counter for your final statement, end users don't just care about whether or not works.

If all end-users cared about was whether or not something worked, we'd all still be using Microsoft DOS with Corel Wordperfect. End users care about more. There's a reason the browser wars have re-ignited with 4 different relatively competent engines on the market (Gecko, Presto, IE, Webkit). There's a reason why Linux hasn't gone the way of the dodo in personal computer sales, and same with Apple.

If whether or not something worked was all most end-users cared about, we'd never be where we are today.

Btw, is there an ignore function or the like? I'm beginning to get the feeling that Dean is deliberately playing the part of flame-bait.