Learning OpenMP (7) together -- critical section of Thread Synchronization

I. Introduction

The previous section describes the usage of the atomic mutex lock mechanism in OpenMP. This section describes the usage of the Critical Zone mutex lock mechanism. The critical section here is similar to the criticalsection in windows.

 

Ii. critical section declaration method

# Pragma OMP critical [(name)] // [] indicates the name is optional

{

// Parallel program block. Only one thread can access this parallel program block.

}

 

For example:

# Prgama OMP critical ()

A = B + C;

 

Iii. Example

# Include <iostream> <br/> # include <OMP. h> // header file to be included in OpenMP programming </P> <p> int main () <br/>{< br/> int sum = 0; </P> <p> STD: cout <"before:" <sum <STD: Endl; </P> <p> # pragma OMP parallel for <br/> for (INT I = 0; I <100; ++ I) <br/>{< br/> # pragma OMP critical (a) <br/>{< br/> sum = sum + I; <br/> sum = sum + I * 2; <br/>}</P> <p> STD: cout <"after: "<sum <STD: Endl; <br/> return 0; <br/>}

 

Iv. Summary

The usage of the critical section critical is introduced. The difference between critical and atomic is that atomic is only applicable to the two types of operations specified in the previous section, and atomic is only protected by a single code. Critical can protect a parallel program block. The next section describes the mutex functions provided by OpenMP.