What is parallel computing and what is its use?

(parallel computing)

Parallel computing refers to the process of solving computing problems by using various computing resources. In order to perform parallel computing, computing resources should include computers with multiple processors (parallel processing), the number of computer-specific computers connected to the network, or both. The main purpose of parallel computing is to solve large and complex computing problems quickly. In addition, it also includes: using non-local resources to save costs-using multiple "cheap" computing resources to replace large computers, while overcoming the memory limit on a single computer.

Traditionally, serial computing refers to performing software writing operations on a single computer (with a single central processor). CPU uses a series of instructions to solve problems one by one, but only one instruction can be used at any time. Parallel computing evolved from serial computing, which attempts to simulate the transaction state in nature: many simultaneous, complex and related events occur in sequence.

In order to take advantage of parallel computing, the usual computing problems are characterized as follows:

Dividing the work into discrete parts is helpful to solve at the same time;

Execute multiple program instructions at any time and in time;

Using multiple computing resources to solve problems takes less time than using a single computing resource.

The following matrix lists four possible classifications of computing structures:

S I S D S I M D M I S D M I M D

Single instruction stream, single data stream, single instruction stream, multiple data streams, multiple instruction streams, multiple instruction streams, multiple data streams.

The following websites have related pictures and schematic diagrams.

/Chinese/software/parallel computing . PHP