A 3D Accelerator is a device that manipulates and displays computer graphics for workstations, game consoles, and personal computers. It is also called the Graphics Processing Unit (GPU).
The 3D Accelerator’s Composition
The graphics card dedicated to the examination of floating point operations carries the GPU. These graphic accelerators contain microchips with special mathematical processes used in the effective and efficient rendering of graphics. These GPUs also apply graphics that have primitive operations, thereby making operations run faster than directly drawing onto the host’s screen.
3D Accelerators Today
With the rise of OpenGL API and similar functions in DirectX, GPUs are now capable of programmable shading. Pixels and geometric vertices can now be processed by short programs. Additional image texture can also be included as inputs. GeForce3, introduced by NVIDIA, was the first chip with this capability. ATI Radeon 9700 followed later.
3D Accelerator’s Computational Functions
GPUs use transistors to perform calculations related to 3D graphics. At first, these were used to speed up the work of texture mapping and in the production of images of polygons. Later, these were also used step up geometric calculations.
More recent developments include support for programmable shaders used to manipulate textures and vertices with the same operations that are supported by CPUs. Most computations now involve vector and matrix operations.
3D hardware today also contains the basic 2D framebuffer capabilities and its accelerations. Most of these support hardware overlays and YUV color spacing, which is essential for the playback of digital videos. GPUs made in the year 2000 support MPEG formats, like motion compensation and iDCT.