parallel_for_overhead_test.cpp

#include <Windows.h>
#include <cstdio>
#include <cstdint>

#define LOOP_REPEAT											10
#define LOOP_COUNT											48000 * 6

int main()
{
	int8_t buffer1 [ LOOP_COUNT ];
	int16_t buffer2 [ LOOP_COUNT ];

	LARGE_INTEGER freq, start, end;
	QueryPerformanceFrequency ( &freq );

	QueryPerformanceCounter ( &start );
	for ( int c = 0; c < LOOP_REPEAT; ++c )
		for ( int i = 0; i < LOOP_COUNT; ++i )
			buffer2 [ i ] = buffer1 [ i ] * 256;
	QueryPerformanceCounter ( &end );

	printf ( "Plain For: %lfms\n", ( end.QuadPart - start.QuadPart ) / ( double ) freq.QuadPart );

	QueryPerformanceCounter ( &start );
	for ( int c = 0; c < LOOP_REPEAT; ++c )
#pragma omp parallel for
		for ( int i = 0; i < LOOP_COUNT; ++i )
			buffer2 [ i ] = buffer1 [ i ] * 256;
	QueryPerformanceCounter ( &end );

	printf ( "OpenMP For: %lfms\n", ( end.QuadPart - start.QuadPart ) / ( double ) freq.QuadPart );

    return 0;
}

LOOP_REPEAT: 10
Plain For: 0.010821ms
OpenMP For: 0.002598ms

LOOP_REPEAT: 1
Plain For: 0.000709ms
OpenMP For: 0.000903ms

Test Environment
CPU: AMD Ryzen 5 2600X
RAM: DDR4 32GB(Dual Channels) 2400MHz