Multicore

Multicore refers to a single processing component that contains multiple cores on one integrated circuit. Each core can independently fetch, decode, and execute instructions, enabling parallel execution of multiple threads. Modern CPUs and mobile system-on-chips commonly use multicore designs, from dual-core to many-core configurations, to increase throughput and responsiveness. Software can exploit this with multithreading and parallel algorithms.

Cores usually share memory and I/O resources. Typical architectures include per-core caches (L1 and often L2) and a larger shared cache (often L3). An interconnect links cores to memory controllers and peripherals. Cache coherence protocols maintain a consistent view of memory across cores, while synchronization mechanisms manage inter-core communication and contention.

Advantages include higher aggregate performance and improved energy efficiency for parallel tasks, better multitasking, and responsiveness. However, to realize gains software must be parallelizable; overhead from synchronization, contention for shared resources, and memory bandwidth limits can reduce speedups. Amdahl's law captures the potential limits of parallelization.

Variants include homogeneous multicore designs, where all cores are identical, and heterogeneous designs such as big.LITTLE that mix high-performance and power-efficient cores. Multicore CPUs began to dominate in the mid-2000s and became mainstream in the 2010s, with desktop, mobile, and server processors widely adopting the approach.