The 7th of September, filled with high hopes. Today was the day that Creative Labs would give us the first glimpse of their 3D Blaster GeForce256, or "Annihilator" (in the US). Prior to this day, all sorts of information on the GeForce256 have been appearing on the web. There were many speculations going on and even some rumors. Creative Labs set out to show us just what the GeForce256 was all about, its ups and its flaws. In the following feature we will present exactly what we saw at the Creative Labs GeForce256 Press event!

Live the Experience

To the beat of Jimi Hendrix' 'Experience' the entire Scandinavian computer press lobby got introduced to Creative Labs' latest slogan - Live the experience. A few minutes of 'rad' video clips, and the press event could begin.

Everyone there knew pretty much what they wanted to see, but, as always with corporate events, there were a couple of minutes of pep-talk from the marketing team assuring us that Creative Labs rocks. With the smell of the GeForce 256 in the air we were quite excited, so demonstrations of less 'important' products caught our attention pretty well too. Anyway, let's skip the irrelevant babble and get down to the wares.

GeForce256

As everyone and their grandma now know, nVIDIA uncovered something they called a revolutionary (rather than evolutionary) step in computer graphics, by releasing their GeForce 256 processor. Originally called NV10, the processor had been rumored to incorporate all kinds of mad features, but the key specs people were interested of were high fill rates and most importantly, on board T&L (polygon transformation and lighting). Easily put, T&L moves all the polygon transformation and lighting calculation from the computers CPU, to the graphics processor (or GPU- Graphics Processing Unit), making computer 3D graphics something rather processor independent (an upside down situation to what we had until today...).

In this sense, it's true, the chip is a revolution, and we can tell you this, nothing we saw at the show hinted at anything else...

What? Does it stop there? Do we look like fools? Read on for pageloads of GEForce 256, and some focus (well okay, loads of it) on Creative Labs' GEForce 256 based card!

Transform Engine

With the introduction of nVidia's GeForce256 "GPU", Transform & Lighting (T&L) have now been moved on-chip. No longer do we require the CPU to do T&L calculations prior to rendering a scene. Basically, transformation & lighting use a lot of CPU power and this of course will slow down performance (on polygon intensive games).

With GeForce256, the chip does all T&L calculations itself, thus rendering a fast CPU useless for gameplay. This has set a new standard with 3D gaming as we see it, no longer will you have to use the latest CPU in order to get fast framerates. The downside of this is that when you now decide to upgrade your CPU, say from a PII 400 to a PIII-550, you will not see a large performance shift, rather, it is your GeForce256 that will set the limit, upgrading a CPU won't affect end performance.

Exactly how much CPU power is freed now that T&L is moved onboard? We asked that very same question and Creative Labs confirmed that roughly 20 - 50% of CPU time can be given back, highly depending on the application in question. They also said that 50% maybe was a very optimistic number, but 20% should be expected when dealing with games. Not bad at all, you get 20% of you CPU power back which can be used to perform other tasks such as physics, inverse kinematics, sophisticated artificial intelligence etc. This mean that game developers now have more CPU time to play with when doing their computer AI in games, and you thought Quake2 enemies were stupid because they're programmed badly. ;)

The Lighting Engine

The lighting engine is similar to the transform engine because it has a set of mathematical functions that it must perform. The GeForce256 have separate transform and lighting engines, this way each engine can run at maximum efficiency. Without these engines, the transform performance would be limited by having to share compute time with the lighting calculations.

The lighting engine task is to calculate distance vectors from light sources to objects in the 3D scene as well as distance vectors from objects to the viewer's eyes. A vector contains information about direction and distance. The engine must also separate the distance information from the direction information as that simplifies future steps in the 3D pipeline. Lighting calculations are mainly used for vertex lighting, but they are also used for other effects such as advanced fog effects, which are based on the eye-to-object distance rather than just the Z-value of the object.

Why We Want Integrated T&L Engines on a Graphics Chip

Separate transform and lighting engines integrated into one chip are necessary for graphics processors to increase the user experience and are absolutely necessary for graphics processors to continue to advance the user experience. 3D graphics performance have grown rapidly over the past years, with primary emphasis on the pixel fill rate and texture mapping capabilities. This gave gamers fast frame rates but left the task of geometry transform and lighting calculations for the CPU to do. Advances in 3D scene complexity and object detail will move slow as long as the T&L calculations are done by the CPU as CPUs double in speed every 18 months or so while graphics processors advance much more than that during the same period of time. Geometry processing (T&L) performance is now become the most important barrier to break in order to get more sophisticated 3D graphics on a typical PC. With the GeForce256, nVidia presents the first graphics chip that has T&L integrated inside the chip itself, thus freeing the CPU and breaks the geometry performance bottleneck.