micro_benchmark_rdtsc.h File Reference

#include "micro_benchmark_common.h"
#include "micro_benchmark_rdtsc_impl.h"

Include dependency graph for micro_benchmark_rdtsc.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Functions
template<typename _Callable, typename... _Args>
void	micro_benchmarkRdtsc (double &ellapsedTimeNs, double &ellapsedTimeErrorNs, double &timePerElement, double &timeErrorPerElement, size_t nbTestPerf, size_t nbCallPerTest, size_t nbElement, _Callable &&__f, _Args &&... __args)
	Do the micro benchmarking of a given function and gives performance results in cy.
template<typename _Callable, typename... _Args>
void	micro_benchmarkRdtsc (double &ellapsedTimeNs, double &ellapsedTimeErrorNs, size_t nbTestPerf, size_t nbCallPerTest, _Callable &&__f, _Args &&... __args)
	Do the micro benchmarking of a given function and gives performance results in cy.
template<typename _Callable, typename... _Args>
void	micro_benchmarkRdtscPrint (const std::string &testName, size_t nbTestPerf, size_t nbCallPerTest, size_t nbElement, _Callable &&__f, _Args &&... __args)
	Do the micro benchmarking of a given function and gives performance results in cy and print the result.

Function Documentation

micro_benchmarkRdtsc() [1/2]

template<typename _Callable, typename... _Args>

void micro_benchmarkRdtsc	(	double &	ellapsedTimeCy,
		double &	ellapsedTimeErrorCy,
		double &	timePerElement,
		double &	timeErrorPerElement,
		size_t	nbTestPerf,
		size_t	nbCallPerTest,
		size_t	nbElement,
		_Callable &&	__f,
		_Args &&...	__args )

Do the micro benchmarking of a given function and gives performance results in cy.

Parameters

[out]	ellapsedTimeCy	: ellapsed time in cy
[out]	ellapsedTimeErrorCy	: error on the ellapsed time in cy
[out]	timePerElement	: time per element in cy
[out]	timeErrorPerElement	: error of time per element in cy
	nbTestPerf	: number of performance test
	nbCallPerTest	: number of calls per performance test
	nbElement	: number of element treated by the function __f
	__f	: function to be called and benchmarked
	__args	: parameter of the function to be benchmarked

Definition at line 58 of file micro_benchmark_rdtsc_impl.h.

{
        micro_benchmarkRdtsc(ellapsedTimeCy, ellapsedTimeErrorCy, nbTestPerf, nbCallPerTest, __f, __args...);
        timePerElement = ellapsedTimeCy/((double)nbElement);
        timeErrorPerElement = ellapsedTimeErrorCy/((double)nbElement);
}

References micro_benchmarkRdtsc().

Here is the call graph for this function:

micro_benchmarkRdtsc() [2/2]

template<typename _Callable, typename... _Args>

void micro_benchmarkRdtsc	(	double &	ellapsedTimeCy,
		double &	ellapsedTimeErrorCy,
		size_t	nbTestPerf,
		size_t	nbCallPerTest,
		_Callable &&	__f,
		_Args &&...	__args )

Do the micro benchmarking of a given function and gives performance results in cy.

Parameters

[out]	ellapsedTimeCy	: ellapsed time in cy
[out]	ellapsedTimeErrorCy	: error on the ellapsed time in cy
	nbTestPerf	: number of performance test
	nbCallPerTest	: number of calls per performance test
	__f	: function to be called and benchmarked
	__args	: parameter of the function to be benchmarked

Definition at line 22 of file micro_benchmark_rdtsc_impl.h.

{
        VecEllapsedTime vecTimeCy;
        for(size_t i(0lu); i < nbTestPerf; ++i){
                //Stating the timer
                size_t beginTime = phoenix_rdtsc();
                for(size_t j(0lu); j < nbCallPerTest; ++j){
                        __f(__args...);
                }
                //Get the time of the nbCallPerTest calls
                size_t elapsedTime(phoenix_rdtsc() - beginTime);
                double fullCy(elapsedTime/((double)nbCallPerTest));
                vecTimeCy.push_back(fullCy);
        }
        MapOrderedTime mapOrderTime;
        micro_benchmarkVecToMap(mapOrderTime, vecTimeCy);
        size_t nbValueUsed(vecTimeCy.size()*0.7 + 1lu);
        if(nbValueUsed > vecTimeCy.size()){
                nbValueUsed = vecTimeCy.size();
        }
        micro_benchmarkComputeTime(ellapsedTimeCy, ellapsedTimeErrorCy, mapOrderTime, nbValueUsed);
}

References micro_benchmarkComputeTime(), and micro_benchmarkVecToMap().

Referenced by micro_benchmarkRdtsc(), and micro_benchmarkRdtscPrint().

Here is the call graph for this function:

Here is the caller graph for this function:

micro_benchmarkRdtscPrint()

template<typename _Callable, typename... _Args>

void micro_benchmarkRdtscPrint	(	const std::string &	testName,
		size_t	nbTestPerf,
		size_t	nbCallPerTest,
		size_t	nbElement,
		_Callable &&	__f,
		_Args &&...	__args )

Do the micro benchmarking of a given function and gives performance results in cy and print the result.

Parameters

testName	: name of the performance test
nbTestPerf	: number of performance test
nbCallPerTest	: number of calls per performance test
nbElement	: number of element treated by the function __f
__f	: function to be called and benchmarked
__args	: parameter of the function to be benchmarked

Definition at line 76 of file micro_benchmark_rdtsc_impl.h.

                                                                                                                                                         {
        double ellapsedTimeCy(0.0), ellapsedTimeErrorCy(0.0), timePerElement(0.0), timeErrorPerElement(0.0);
        micro_benchmarkRdtsc(ellapsedTimeCy, ellapsedTimeErrorCy, timePerElement, timeErrorPerElement, nbTestPerf, nbCallPerTest, nbElement, __f, __args...);
        
        std::cout << testName << " : nbElement = "<<nbElement<<", timePerElement = " << timePerElement << " cy/el ± "<<timeErrorPerElement<<", elapsedTime = " << ellapsedTimeCy << " cy ± "<<ellapsedTimeErrorCy << std::endl;
        std::cerr << nbElement << "\t" << timePerElement << "\t" << ellapsedTimeCy << "\t" << timeErrorPerElement << "\t" << ellapsedTimeErrorCy << std::endl;
}

References micro_benchmarkRdtsc().

Here is the call graph for this function: