AI engine

EN Texte en anglais Pas encore d'article dans ta langue : extrait en anglais.

An AI engine is the computational core of artificial intelligence systems and is used for AI inference. It can also refer to a specific computing architecture created by AMD (formerly by Xilinx, acquired by AMD in 2022). For AI inference, an AI engine consists of software layers that run AI workloads and manage memory and scheduling, simplifying application development. The AMD AI engine is commonly used for streamlining linear algebra operations (such as matrix multiplication), for artificial intelligence algorithms, digital signal processing, and more generally, for high-performance computing. The first products containing AI engines were the Versal adaptive compute acceleration platforms, which combine scalar, adaptable, and intelligent engines connected through a Network on Chip (NoC). AI engines have evolved significantly alongside modern computing workloads, with developments specifically focused on accelerating AI applications. The basic architecture of a single AI engine integrates vector processors and scalar processors to implement Single Instruction Multiple Data (SIMD) capabilities. AI engines are integrated with other hardware such as FPGAs, CPUs, and GPUs to support high-performance, heterogeneous computation across various domains.