What's NVIDIA up to...? v2

[Tabajara]

Continues from here: What's NVIDIA up to...? v1
==

Haha, well that's to be expected, but yet another rename of the 8800GT>9800GT (Plus GS/GSO derivatives) >GTS250>GT330 is just nuts...

The G80/92 refuses to die..!

I think it will be forever known as the little GPU that could

[Tabajara]

Jen-Hsun to Intel: we won't settle:

Sometimes companies spar out their differences behind closed doors, and sometimes they have guys like Jen-Hsun Huang at their helms and the whole world gets to know how they feel and what they intend to do about it. The Tegrasaurus Rex has taken a recent interview with Fortune magazine as an opportunity to eloquently lay out his side's case in the epic cross-licensing dispute between NVIDIA and Intel, and to let us all know that he sees "no reason" to settle with the Atom-making giant. Describing Intel's argumentation as "completely nonsense," NVIDIA's fearless leader tell us that he's eagerly anticipating the court clash scheduled for later this year. We can't yet confirm whether or not he finished it off with a "bring your popcorn" instruction, but all his recorded words await in video form just after the break.

Check the link, the video is interesting... The Fermi demos are running in the background while Jen-Hsun speals

http://www.engadget.com/2010/03/09/jen- ... -intel-se/

[powerarmour]

Haha, I totally agree with one of the comments, he's obviously slightly crazy, but he has a massive pair of balls...

But getting into court battles with Intel generally is like pulling on a Lion's tail...

[Tabajara]

But getting into court battles with Intel generally is like pulling on a Lion's tail...

While you are at his cave. And he's hungry.

If Jen-Hsun thinks he has a big company and pockets fat enough to fight intel, he might have to review his perception of reality soon

[rahul]

Haha, I totally agree with one of the comments, he's obviously slightly crazy, but he has a massive pair of balls...

But getting into court battles with Intel generally is like pulling on a Lion's tail...

Exactly, i think its wrong move by nvidia to go against the giant, Most probably Intel has more lawyers than the nvidia employee strength

[Blakhart]

they are both paying lawyers no matter if theyre busy or not so maybe this will still be in court in a hundred years

[thegrommit]

It seems nvidia is still busy messing with people who're using ATI/Nvidia hybrids for PhysX

http://www.geeks3d.com/20100422/hybrid- ... ebomb-fix/

A new version of the PhysX patch that allows a Radeon card (for 3D rendering) to be used with a GeForce card (for PhysX) is available.

If you read the release note there is a pretty shocking thing:

- added support for fixing a timebomb, introduced in 196.75 and 197.xx drivers (gravity reverse + overall slowdown after a few seconds of PhysX processing)

The video is amusing.

[Tabajara]

It seems nvidia is still busy messing with people who're using ATI/Nvidia hybrids for PhysX

http://www.geeks3d.com/20100422/hybrid- ... ebomb-fix/

A new version of the PhysX patch that allows a Radeon card (for 3D rendering) to be used with a GeForce card (for PhysX) is available.

If you read the release note there is a pretty shocking thing:

- added support for fixing a timebomb, introduced in 196.75 and 197.xx drivers (gravity reverse + overall slowdown after a few seconds of PhysX processing)

The video is amusing.

Seems PhyxX is an open architecture as long as you use a nVidia video card

[impar]

Greetings!

Computex time for Fermi 2

The new successor to Fermi is actually not completely based on a cut down Fermi architecture. The new GPUs, something that we know as GF104 and GF108, are based on a new architecture that has some Fermi basics, but doesn't have any cache. As you can imagine, the focus is on good tessellation support as well as on strong DirectX 11 performance. This is what Nvidia has been pushing for a while.

nVidia dropping the GPGPU and focusing on GPUs?

[powerarmour]

nVidia dropping the GPGPU and focusing on GPUs?

It's because they know that Fermi is too big to produce and scale economically, so it makes sense that they are stripping out all the best bits for something leaner and faster. The GTX480/470 was castrated of the full double precision performance of the Fermi Tesla anyway, and it doesn't really make any difference in most GPGPU apps.