What are a GPU and a graphics card?

A GPU is a graphics processing unit designed to perform many similar calculations in parallel. Besides rendering, it can accelerate video work and suitable scientific computations. A graphics card is the hardware board containing a GPU, memory, power components and connectors; the terms are not identical.

How does it differ from a CPU and integrated graphics?

A CPU handles general-purpose tasks with complex control flow, while a GPU excels at similar operations across many data items. Not every program runs faster on a GPU: software must be designed to use it. Integrated graphics may share a chip or package with a processor and use system memory. Discrete cards generally have their own memory for scene data and image buffers. More graphics memory alone does not establish greater processing capability.

What determines performance and compatibility?

Resolution, scene complexity and application settings change the workload. A game might be limited by the CPU, while another application runs short of graphics memory. Drivers, operating-system support and computing interfaces must work together. Power consumption, cooling, case dimensions and display connectors also matter. Acceleration does not improve latency or quality under every condition because transferring data has a cost. A benchmark for one workload should not be treated as a universal ranking across all applications.