Thomas Alsop
Research expert covering the global hardware industry
A graphics processing unit (GPU) is a type of processor that resides in computers and other electronic systems, as well as being deployed as a standalone unit. GPUs specialize in the acceleration of computational tasks, including the creation and rendering of images, videos, and animations. GPUs began as specialized application-specific integrated circuits (ASICs), developed to accelerate specific 3D rendering tasks. Today, GPUs have evolved to become more general-purpose processors, working in parallel to other computer components to handle a growing range of applications, notably artificial intelligence (AI) and machine learning (ML) tasks.
GPU market leaders include Nvidia, AMD, and Intel
GPUs are made available either as integrated or discrete components. Integrated graphics, where Intel dominates, refers to a computer where the GPU is built onto the same die as the central processing unit (CPU). The GPU is not part of the CPU in a discrete component and has a dedicated memory and power supply. They can also be referred to as a graphics add-in-board (AIB), with Nvidia and AMD having dominant control of this market. In smartphones and other mobile devices, GPUs feature as part of a system-on-chip (SoC) and can render graphics and help complete AI tasks on the device. Technologies and solutions from the leading GPU suppliers have been deployed to train and run a variety of large language models. A notable example is the development of ChatGPT, trained using tens of thousands of Nvidia GPUs, linked together in an AI supercomputer belonging to Microsoft.In addition to AI, GPUs are used for gaming and cryptomining
Beyond AI, a GPU provides the fastest graphics processing and for gamers, the GPU is a standalone card plugged into the computer via a peripheral component interconnect express (PCIe) interface. The more sophisticated the GPU, the higher the resolution, and the faster and smoother the motion. Improved framerates, or the number of frames per second the game runs at, means there is less stutter or freezing due to excessive load on the CPU, thus enhancing the gaming experience. The power potential of a GPU compared to a CPU has also proved useful in cryptocurrency mining due to its speed and efficiency in solving mathematical equations. Such is the widescale demand for GPUs, particularly with the soaring hunger for exploring generative AI applications, that a GPU shortage could well be on the cards down the line, a phenomenon that could see prices skyrocket as disruption to supply lines takes hold.
