Unidade de Procesamento Gráfico

Na Galipedia, a wikipedia en galego.

Este artigo necesita tradución, a súa axuda é benvida. Se o artigo non se traduce ao Galego no prazo de trinta días, pasará á lista de páxinas para borrar.

GeForce 6600GT (NV43) GPU

Image:Thermalgrease.jpg

Radeon 9800 Pro (R350) GPU

A Unidade de Procesamento Gráfico (Graphics Processing Unit) ou GPU (tamén coñecida como Unidade de Procesamento Visual (Visual Processing Unit) ) e un microprocesador (ou acelerador gráfico) para computadores persoais, videoconsolas e outros dispositivos que requiran de gran cantidade de imagexes xeradas por computador. As GPUs modernas son moi eficientes manipulando e mostrando gráficos por computador, e a súa estructura de alto paralelismo fainas moito máis efectivas que as tipicas CPUs, para un amplio rango de algoritmos complexos.

Unha GPU implementa un numero de operacións gráficas primitivas nun modo que as fai correr moito máis rapidas que debuxar directamente na pantaia co CPU da máquina. As operacións máis comuns para os primeiros gráficos 2D inclúen o copiado rectangular de zoas de memoria (como as xanelas ou os videoxogos), coñecido polo termino ingles de "blitter", e tamén operacións de recheo de rectangulos, triangulos, circulos, e demáis. As GPUs modernas levan soporte para gráficos 3D, como ordeación dos objectos, iluminación, transformacións xeométricas, etc, e tamén capacidades de vídeo dixital.

[editar] Historia

As GPUs modernas descenden dos monolíticos chips gráficos dos últimos anos 70 e anos 80. Estos chips estaban limitados a axuda do blitting na forma de sprites, se equipaban axuda algunha, e non tiñan xeralmente ningunha axuda para o debuxo. Algúnhas GPUs podían levar a cabo varias operaciones de display list, e poderían utilizar acceso directo a memoria para reduci-la carga no procesador. Un primeiro exemplo foi o ANTIC o co-procesador usado no Atari 800 e Atari 5200. O final dos anos 80 e comezando os 90, ca alta velocidade dos microprocesadores de uso xeral fíxose popular implementándose en varias tarxetas gráficas (moi costosas) para PC y estacións de traballo usandos coprocesadores matemáticos (como a Serie de TMS340 de Texas Instruments), para por a dispor do programador as funcions de debuxo rápidas, e moitas impresoras laser equipadas con PostScript (un caso especial de un GPU) funcionando con CPUs RISC como AMD 29000.

As chip process technology improved, it eventually became possible to move drawing support and BitBLT onto the same board (and, eventually, into the same chip) as a regular frame buffer controller such as VGA; these cut-down "2D accelerators" weren't as flexible as microprocessor-based GPUs, but were much easier to make and sell. The Commodore Amiga was the first mass-market computer to include a blitter in its video hardware, and IBM's 8514 graphics system was one of the first PC video cards to implement 2D primitives in hardware.

By the early 1990s, the rise of Microsoft Windows sparked a surge of interest in high-speed, high-resolution 2D bitmap graphics (which had been the domain of Unix workstations and the Apple Macintosh before then.) For the PC market, the dominance of Windows meant PC graphics vendors could now focus development effort on a single programming interface, GDI.

In 1991, S3 Graphics introduced the first single-chip 2D accelerator, the S3 86C911 (which its designers named after the Porsche 911 as an indication of the speed increase it promised). The 86C911 spawned a host of imitators; by 1995, every major PC graphics chip maker had added 2D acceleration support to their chips. By this time, fixed-function Windows accelerators had surpassed expensive general-purpose graphics coprocessors in terms of Windows performance, and coprocessors faded away from the PC market.

With the advent of the DirectX version 8 API and similar functionality in OpenGL, GPUs added programmable shading to their capabilities. Each pixel could now be processed by a short program that could include additional image textures as inputs, and each geometry vertex could likewise be processed by a short program before it was projected onto the screen. By 2003, with the introduction of the NVIDIA GeForce FX (a.k.a. NV30), pixel and vertex shaders could implement looping and lengthy floating point math, and in general were quickly becoming as flexible as a CPU for image-array-style operations.

Today parallel GPUs have begun making computational inroads against the CPU, and a subfield of research, dubbed GPGPU for General Purpose Computing on GPU has found its way into fields as diverse as oil exploration, scientific image processing, and even stock options pricing determination.

[editar] As Capacidades das GPUs modernas

Modern GPUs use most of their transistors to do calculations related to 3D computer graphics. They began by accelerating the memory intensive work of texture mapping and rendering polygons, and later added units to accelerate geometry calculations such as mapping vertex into different coordinate systems. Recent developments in GPUs include support for programmable shaders which can manipulate vertices and textures with many of the same operations supported by CPUs, oversampling and interpolation techniques to reduce aliasing, and very high-precision color formats. Because most of these computations involve matrix and vector operations, engineers and scientists have increasingly studied using GPUs for non-graphical calculations. Because all those applications are way beyond actual GPU's usage target, a new term, GPGPU is usually employed. While GPGPUs and GPUs are the same thing, there's some pressure by "GPGPU users" to improve hardware design. This is usually biased towards adding more flexibility in the programming model.

Although modern PC GPUs feature progammability in the form of 3D-shaders, this should not be confused with general software progammability. Instead, these units operate as SIMD or sometimes MIMD parallel processors. For rectangular arrays of colored pixels, this is an optimal design, and many algorithms -- but not all -- can be adapted to use a GPU for extremely high throughput.

In addition to the 3D hardware, today's GPUs include basic 2D acceleration and frame buffer capabilities (usually with a VGA compatibility mode). In addition, most GPUs made since 1995 support the YUV color space and hardware overlays (important for digital video playback), and many GPUs made since 2000 support MPEG primitives like motion compensation and iDCT.

The typical modern stand-alone GPU sits on a separate graphics card from the motherboard, connected to the CPU and main RAM through the AGP or PCI Express bus. It has access to RAM on the card which is usually faster but lower-capacity than the main RAM. On the other hand, many motherboards have a GPU integrated into the Northbridge that uses the main memory as a frame buffer. This will usually be a cheaper solution than an independent GPU but will have dramatically lower performance. Integrated motherboards may or may not have a slot for a stand-alone graphics card.