Grafikprocessor

The grafikprocessor, commonly referred to as a graphics processor or GPU, is a specialized electronic circuit designed to accelerate the creation and manipulation of images for display. GPUs are optimized for parallel processing, enabling them to perform the same operation on many data elements simultaneously, which is well suited for rendering 3D graphics, image processing, and other data-parallel workloads.

Modern GPUs contain hundreds to thousands of processing units grouped into streaming multiprocessors or compute units. They include shader units, texture units, rasterizers, and memory controllers, and implement a rendering pipeline that handles vertex processing, shading, texturing, rasterization, and final color output. Many GPUs also feature dedicated hardware for ray tracing and AI-based tasks and support high-bandwidth memory interfaces such as GDDR6 or HBM.

GPUs can be discrete cards installed in a system or integrated on the same chip as the CPU. Discrete GPUs offer higher performance and memory, while integrated GPUs share system memory and are more power-efficient. GPUs expose graphics and compute APIs such as DirectX, OpenGL, Vulkan, and Metal; and compute frameworks such as CUDA, OpenCL, or vendor-specific interfaces for general-purpose computing.

Applications include gaming, professional visualization, scientific computation, and machine learning. In graphics, GPUs render complex scenes in real time; in compute workloads they perform large-scale matrix operations and simulations due to their parallel architecture.

Trends emphasize real-time ray tracing, AI-assisted rendering, energy efficiency, and scalable multi-GPU configurations. The GPU market is dominated by a small number of manufacturers that compete on raw performance, memory bandwidth, and software ecosystems.