GPU

GPU stands for graphics processing unit. Its architecture distributes suitable work across many processing resources rather than treating every job as a single sequence. Graphics rendering, video processing and AI computation are examples of workloads that can use this parallelism. (Intel: CPU vs. GPU)

A GPU is not automatically faster than a CPU at every task. CPUs handle varied program and operating-system work, while GPUs are designed to accelerate suitable parallel operations. Use the workload, rather than the acronym, as the starting point for a comparison. (Intel: CPU vs. GPU)

An integrated GPU is built into a processor design; a discrete GPU is a separate processing device. Both are GPUs. The distinction describes how the graphics function is implemented, not a universal measure of capability. (Intel: CPU vs. GPU)

Consider an image-editing application as a learning example. Adjusting many image elements can involve repeated calculations, while opening files and responding to menus involves general application work. This illustrates why one application may use both CPU and GPU resources; it is not a claim that every filter runs on a GPU.

When a product description says “GPU,” ask whether it means the processing device or a complete board. For a real component, consult its documentation rather than assuming a board-level specification applies to a bare device.

Use these questions as a reading checklist, not as a device-selection recommendation. Exact manufacturer documentation takes priority over a general definition.

• Identify integrated versus discrete implementation.
• Check application and driver support for the intended operation.
• Read memory capacity, power and cooling requirements separately.
• Compare measured results for the same workload, not just product labels.

For the broader explanation, compare gpu, npu and cpu roles. This page defines the term; the companion guide connects it to the rest of the system.