NVIDIA Has Major New Linux Driver: Optimus, RandR 1.4

[phoronix]

Phoronix: NVIDIA Has Major New Linux Driver: Optimus, RandR 1.4

NVIDIA announced the 319.12 Beta Linux graphics driver announcement this morning. While the number may seem mundane, this is a very exciting NVIDIA Linux graphics driver update. It presents work that will benefit the agonized NVIDIA Optimus Linux users, provides new RandR functionality, and other new features...

http://www.phoronix.com/vr.php?view=MTM0NzE

[EvilTwin]

first oss uvd for radeons now optimus for nvidia
awsome

[schmidtbag]

That's actually a lot of features for 1 update, even without optimus. Just wondering though - which GPUs are supported by this?


Also, correct me if I'm wrong but I thought Linus said Nvidia was ONE of the worst companies he had to deal with, and Optimus wasn't the main reason he felt this way (it's been a while since I've seen the video). I'm not blaming just Michael here, this seems to be a very common misconception from multiple sources. While optimus is a feature advertised by nvidia hardware (and therefore should be available to all customers), it's more of a privilege than anything. I don't think Linus had endless months of sleepless nights because he couldn't switch his GPU when playing games, as does anyone else. You know what else Nvidia hardware supports that isn't on linux? Direct X, but you don't hear people whining to nvidia about that.

I respect and understand Linus' frustration toward nvidia, but I don't like it when people decide to put their opinion in someone else's mouth. Most people who fawn over Linus giving the finger didn't even watch the video.

[plantroon]

Tearing, TF2 loading time ...

Tearing, Stuttering, TF2 long loading time ... is that solved ?? Basically I only need to get rid of those things

[d2kx]

Tearing, Stuttering, TF2 long loading time ... is that solved ?? Basically I only need to get rid of those things

TF2 loading time is not a nVidia bug and will get fixed by April, 26th by Valve with the SteamPipe update.

[Calinou]

- Support for application profiles to the NVIDIA client-side GLX implementation.

Does this mean we can force anti-aliasing/anisotropic filtering per-application now? That'd be cool.

[enrico.tagliavini]

Can I just point out the README clearly says the other device must use the modesetting driver? This means if you have an intel card you can't use the intel driver with it to enable nvidia optimus. For now this support is just useless. Bumblebee is simpler, even more if used with primus.

I will be excited when I will see the same stuff they do on windows (so automatic swithing without user intervention and creative configurations). That said I hope this is a start..... a non exiting one.

[Alliancemd]

Re

Torvalds' criticism was actual at the moment. And he explained why he said that.
It's not just pure hatred as Michael puts it...

[Serafean]

Can I just point out the README clearly says the other device must use the modesetting driver? This means if you have an intel card you can't use the intel driver with it to enable nvidia optimus. For now this support is just useless. Bumblebee is simpler, even more if used with primus.

I will be excited when I will see the same stuff they do on windows (so automatic swithing without user intervention and creative configurations). That said I hope this is a start..... a non exiting one.

Why should THEY do it? Here is an entire framework getting ready for crazy stuff, and they're still in their sandbox.
modesetting driver... Well someone's having a good laugh How the hell does it actually work? Right now it sounds as binding the X server to one PCI device would achieve the same thing. (Well, except that the display isn't connected to the nvidia GPU...)

Serafean

[tuke81]

Does this mean we can force anti-aliasing/anisotropic filtering per-application now? That'd be cool.

Yep, it sure does among of other things:

http://us.download.nvidia.com/XFree86/Linux-x86/319.12/README/profiles.html#ListOfSupportedd38df