Intimate Tech: Nvidia’s GF100


Over the course of the last couple of months, we’ve been hearing an awful lot from Nvidia. While a lot of that attention has been focused on the upcoming release of Nvidia’s second generation Tegra (for now still tentatively referred to as “Tegra 2”), there’s been more and more focus on the upcoming GF100.

The card itself looks a little bulky...

For now, there are a few fairly vital bits and pieces that we just don’t know about the GF100, but there’s still enough floating around to get a very solid picture indeed of what Nvidia is going with the new hardware. For those who are saying, as Shelton did, “Wait… what happened Fermi?” there’s at least the fact that we’re told the GF-100 is Fermi-derived, so we know it has that history behind it at least.

In Simple Terms:

For now, we don’t know the clock-speed of the GF100, and as a result there’s precious little to know about the power consumption for now, except that it’s already been noted to be a little heavier on power than Nvidia’s own GT200.

Still, before we get bogged down in technical terms, we can help to explain a bit about the GF100 and why there’s so much fuss about it since CES, in terms that just about anyone can understand. It’s relatively simple, but it seems as though very few people have actually tried to describe the GF100 in simple terms.

GF100 was initially just the codename, many have pointed out that it doesn’t stand for GeForce, as some expected, but for Graphics-Fermi. Still, it looks as though the moniker has stuck, and Nvidia could well launch the card as GF100. It was originally meant to be released last year, but rumour has it that due to unusually low yields it’s been bumped back, with many expressing fears that the release date is still looking a little hazy now, despite the fact that Nvidia is saying we’ll see it in March.


The folks at AnandTech seem to have had some more time to get acquainted with the GF100, providing an unusually in-depth breakdown of the device. Indeed, they point out in their description of it that,

“As a quick refresher, the GF100 is composed of 512 SPs, which Nvidia now calls CUDA cores. Each CUDA core is organised into a block of 32 cores which we call a Streaming Multiprocessor (SM), which in turn are laid out so that 4 of them are in a Graphics Processing Cluster (GPC). Finally, there are 4GPCs in a single GF100 chip. 32x4x4 = 512 CUDA cores.”

The piece then goes on to describe some of the key differences between it and some of Nvidia’s other hardware we might be familiar with, saying,

“While the GF100 resembles the GT200 in a number of ways as a complete GPU, as a gaming GPU it’s very close to being a complete departure from GT200. The big change here is that the single block of fixed-function hardware as we know it is gone. Virtually every piece of hardware has been split up and moved down a level, and can now be found as part of a GPC or as an SM.”

If you’d like to read anything very much more technical than that then you’d do well to check out the full and detailed technical breakdown over at HotHardware, where they seem to have teased out a tremendous amount of detail.

Where it all leads…

Of course, the upshot of all that fancy hardware, and power consumption, is that we can expect some faster next-generation cards out of Nvidia in the coming months. Indeed, given the current situation between Nvidia and ATI, that’s something that can’t come too soon really. Still, there are things happening with the GF100 that should give us a bit of a look towards the future.

One of those is the fact that the GF100 boasts 32x anti-aliasing, for those who just can’t deal with a single jagged line in anything they happen to play with it. Perhaps more interesting, despite its current status as in-development, is the fact that the GF100 features support for full hardware decode on the GPU for 3D Blu-Ray discs.

While the “3D revolution” is still in its infancy, the massive success of James Cameron’s Avatar seems to have spurred the world and its mother towards cinemas enough that Samsung and Sony are both betting that this is the year we’ll see 3D TVs take off in a big way. If we see 3D TV take off then it makes sense that we’ll see plenty of 3D releases on Blu-Ray too, so it makes an awful lot of sense for Nvidia to ensure that the GF100 is set for the

It’s nice to see it all coming from ‘distant future’ to just around the corner though. Fingers crossed that we’ll see something really interesting from Nvidia from a performance point of view, considering the fact that it’s been fairly mauled by ATI in those stakes in the last couple of years.


Tags: , , , , , , , , , , , , , , , , , , , , , , , ,

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: