Appendix A: Retrospective
Over the past decade, NVIDIA’s graphics processing units (GPUs) have evolved
from specialized, fixed-function 3D graphics processors to highly programmable,
massively multithreaded, parallel-processing architectures used for visual computing
and high-performance computation.
NVIDIA GeForce GPUs enable incredibly realistic 3D gaming and outstanding
high-definition video playback, while NVIDIA Quadro
®
GPUs provide the highest
quality and fastest workstation graphics for professional design and creation. For
high-performance computing tasks in various engineering, scientific, medical, and
financial fields, NVIDIA’s new Tesla
™
GPUs and CUDA parallel programming
environment enable supercomputing-level performance on the desktop, at a fraction
of the cost of comparably performing CPU-based multiprocessor clusters.
The GeForce 8800 GPU was launched in November 2006. It was the world’s first
DirectX 10 GPU with a unified shader architecture. This was important as each of
the unified shader processing cores could be dynamically allocated to vertex, pixel,
and geometry workloads, making it far more efficient than prior-generation GPUs,
which used a fixed number of pixel processing units and a fixed number of vertex
processing units. This same unified architecture provided the framework for
efficient high-end computation using NVIDIA CUDA software technology.
The GeForce 9 Series GPUs were introduced in 2007, offering a vastly improved
price-performance ratio and advanced PureVideo
®
features. Its smaller chip allowed
dual-GPU GeForce 9800 GX2 graphics boards to be built more efficiently, while
offering up to twice the performance of the GeForce 8800 GTX.
As of May 2008, over 70 million NVIDIA GeForce 8 and 9 Series GPUs have
shipped and each supports CUDA technology, allowing greatly accelerated
performance for mainstream visual computing applications like audio and video
encoding and transcoding, image processing, and photo editing. These GPUs also
support the new NVIDIA PhysX technology for enabling real-time physics in
games.
GPUs are the most important and most powerful processors in the new era of
visual computing. High-end GeForce GTX 200 GPUs provide the best user
experience when running intensive DirectX 10-based games like
Crysis
at high
quality and high resolution settings. Very capable motherboard and mid-range
GPUs are also needed for stutter-free, high-definition video playback on the PC
while simultaneously displaying the Aero 3D user interface of Windows Vista.
May 2008 | TB-04044-001_v01
21
