Hardware Architecture for Parallel Processing - CS8791 Cloud Computing - Unit 1

Hardware architecture for parallel processing

• The core elements of parallel processing are CPU's.
• It is based on the number of instructions and data streams.
• It can be processed simultaneously.

Computing systems are classified into four categories

1. Single Instruction Single Data (SISD) System
2. Single Instruction Multiple Data  (SIMD) System
3. Multiple Instructions Single Data  (MISD) System
4. Multiple Instructions Multiple Data  (MIMD) System

SISD - Single Instruction Single Data

• It is a uni-processor machine capable of executing a single instruction which operates on a single data stream
• Machine instructions are processed sequentially
• All instructions and data to be processed have to be stored in primary memory. 
• This is limited by the rate at which the computer can transfer information internally.

SIMD - Single Instruction Multiple Data

• It is a multiprocessor machine capable of executing the same instructions on all the CPU's but operating on different data streams
• this model is well suited for scientific computing which involves lots of vector and matrix operations.

MISD - Multiple Instructions Single Data

• MISD computing system is also a multiprocessor machine so it is capable of executing different instructions on different processors that operate as a single data.
• Few machines are built but none of them are available commercially.

MIMD - Multiple Instructions Multiple Data

• It is a multiprocessor machine capable of executing multiple instructions on multiple data sets.
• Each processor in the MIMD separates instruction and data streams.

MIMD machines are broadly classified into two types.

1. Shared memory MMD machines
2. Distributed-memory MIMD machines

1. Shared memory MIMD machines

• All processors are connected to a single global memory.
• This is also called a tightly coupled multiprocessor.
• Communication between processor through by shared memory.
• Modification of the data stored in the global memory by one processor is visible to all other processors.

2. Distributed Memory MIMD Machines

• All processes have a local memory.
• Systems based on this model are also called loosely coupled multiprocessor systems.
• Communication between processes through the interconnection network.
• Each processes are operated by asynchronously and also it exchanging messages synchronously.