It's a GPU
One thing
that nVidia would like everyone to know right up front is that they no longer
call their units video cards. They like the term GPU and that stands for Graphic
Processing Unit. It is no coincidence that the term GPU closely resembles
CPU. nVidia would like to make it plainly obvious that their video cards move
much of the load of graphic rendering off the CPU and onto the GPU. While
the original GeForce chipset did a good job of improving performance with
this approach, it became obvious that higher core clock speeds and memory
performance were necessary to keep improving performance. You can only push
a core speed so far though, and it took a move to the .18 micron architecture
of the GTS chipset to keep improving performance. Avid overclockers already
know that a move to a smaller architecture has immediate benefits not only
in performance, since more transistors can be added in the same space, but
that lower heat production is also a benefit.
The core
of the GTS cards now can run 200 MHz from the factory. While that alone is
an improvement, nVidia improved the inherent design and added the ability
to have more operations processed on the GPU. While the terminology may seem
confusing, the GeForce GTS chipset is capable of such operations as bump mapping,
shadow maps, shadow volumes, volumetric explosion, vertex blending, waves,
refraction, and on and on. What is truly impressive here is that up to 7 operations
can be performed in a single pass. Want more information on all those possible
operations? nVidia has a great
area on their site devoted to the explanation of many of these concepts.
One great
thing that was available on the original GeForce DDR cards is their DDR memory.
While it gets harder to clock memory speeds higher and higher, nVidia chose
to go with memory that effectively double its performance. The memory on the
GTS cards is clocked at 333 MHz by default. DDR memory is definitely going
to be popping up on products as far ranging as even the system memory of future
motherboards. Another thing that nVidia would like people to take notice of
are the initials GTS. They stand for Giga Texel Shader. A Giga Texel of rendering
ability is a huge improvement over the original GeForce chipset. So it should
be plainly obvious that the GeForce2 GTS cards should be a large improvement
over the original GeForce cards. So what makes the Outrageous GeForce2 GTS
different from the competition?
