1.1.2c Multicore and Parallel Systems

Multicore systems allow multiple instructions to be processed simultaneously by using separate cores. 

The cores are integrated into the same chips. Chip multiprocessor (CMP) 

Their performance is improved by: 

• Shared cache 

• Inter-core communication 

Double the cores doesn’t mean double the speed because it can be hard for the cores to communicate with each other as well as some programs not being designed to make use of the extra cores. 

The improvement of performance depends on parallel processing software so that the programs can be split among the cores. 

Can be achieved by processing the same instruction among multiple cores by splitting it up vias the FDE cycle so one processor does each part of the cycle. 

Can also be achieved by one processor fetches and decodes the instruction then hands it off to be executed 

Parallel processing is limited by how many cores can be used at one time