t02_TransformArgvRangeSizeObufBackIns.cpp File Reference

#include <assert.h>
#include <stdlib.h>
#include <iostream>
#include <string>
#include <list>
#include <set>
#include <rpa/testing.h>
#include <rpa/containers/list.h>
#include <rpa/algorithms/transform.h>

Classes
struct	Char2Str
	This because we want to use the algorithm 'transform'. More...
Defines
#define	NB_THR   20
	Our thread pool. The threads are 'create'-d by the algorithm.
#define	NB_SAMPLES   ( ( ( 1 << 20 ) + 1 ) * TST_RATIO )
Typedefs
typedef std::list< std::string >	LstStrT
	Used for storing the results of the algorithm 'transform'.
typedef std::back_insert_iterator< LstStrT >	BackInsT
	This is passed as input the the algorithm 'transform'.
Functions
static void	badRetVal (const char *funcName, const char *typeName)
	Tells that the return value is different from the output iterator.
template<class ThrType>
static void	main_list_unbounded (int aArgC, const char *const *aArgV, rpa::thr_nb_t aNbThr, ThrType *aThrArray, LstStrT &myLstStrs)
template<class ThrType>
static void	main_list_bounded (int aArgC, const char *const *aArgV, rpa::thr_nb_t aNbThr, ThrType *aThrArray, LstStrT &myLstStrs)
template<class ThrType>
static int	mainIntern (size_t aArgC, const char *const *aArgV, rpa::thr_nb_t aNbThr, ThrType *aThrArray)
static void	mainThr (int aArgC, const char *const *aArgV)
int	main (int aArgC, const char **aArgV)
Variables
static char	myStr [NB_SAMPLES][RPA_INT_STR]
	Global variable because too big for the stack.
static char *	myPtr [NB_SAMPLES]
	Global variable because too big for the stack.

---

Define Documentation

#define NB_SAMPLES   ( ( ( 1 << 20 ) + 1 ) * TST_RATIO )

Power of two, plus one. Multiplied by a ratio for adjusting the tests duration.

#define NB_THR   20

Our thread pool. The threads are 'create'-d by the algorithm.

---

Typedef Documentation

typedef std::back_insert_iterator< LstStrT > BackInsT

This is passed as input the the algorithm 'transform'.

typedef std::list< std::string > LstStrT

Used for storing the results of the algorithm 'transform'.

---

Function Documentation

static void badRetVal	(	const char *	funcName,
		const char *	typeName
	)			[static]

Tells that the return value is different from the output iterator.

int main	(	int	aArgC,
		const char **	aArgV
	)

Loops on internal data and on the command line. This ensures that a minimal and identical set of data is run, with the extra flexibility of adding other test samples on the command line.

Setup a buffer of random strings.

Tries the algorithm with various sizes, much smaller and much bigger than the number of threads and the size of buffers. Division by two at each step.

For commodity, tests with the command line.

template<class ThrType>

static void main_list_bounded	(	int	aArgC,
		const char *const *	aArgV,
		rpa::thr_nb_t	aNbThr,
		ThrType *	aThrArray,
		LstStrT &	myLstStrs
	)			[static]

This function receives an array of const char *, plus its size. It receives too an array of threads, and its size. This function must apply the transform algorithm in these input data, with these threads, an return the result. Note: Tests with a buffer of LIMITED size.

If the buffer has a limited size, we need a mutex for synchronization. That is, the sub-threads buffers must be periodically flushed into the output, and this must be protected.

* 'buffer_back_inserter' does the same is 'stdback_inserter', but allows more parallelization. 'make_thread_tree' creates, out of any threads iterator, a 'thread_tree' object which is necessary for parallel execution. A mutex is necessary because buffers have a limited size and will be periodically flushed into the output iterator.

Temporary variable to check the return value.

template<class ThrType>

static void main_list_unbounded	(	int	aArgC,
		const char *const *	aArgV,
		rpa::thr_nb_t	aNbThr,
		ThrType *	aThrArray,
		LstStrT &	myLstStrs
	)			[static]

This function receives an array of const char *, plus its size. It receives too an array of threads, and its size. This function must apply the transform algorithm in these input data, with these threads, an return the result. Note: Test with an output buffer of UNLIMITED size.

* 'obuf_iterator' does the same as 'stdback_inserter', but allows more parallelization. 'make_thread_tree' creates, out of any threads iterator, a 'thread_tree' object which is necessary for parallel execution. No mutex is necessary because the buffers have an unlimited size and will be flushed at the end of the execution.

Temporary variable to check the return value.

template<class ThrType>

static int mainIntern	(	size_t	aArgC,
		const char *const *	aArgV,
		rpa::thr_nb_t	aNbThr,
		ThrType *	aThrArray
	)			[static]

Given a set of data, compares the results calculated with the parallel algorithms, with output buffers, when the buffers have a size limit and when they do not have one.

The order of the results is not deterministic, so we must sort before comparing. The input data are all different, so it must be for the results.

Both conditions must be true.

static void mainThr	(	int	aArgC,
		const char *const *	aArgV
	)			[static]

This function receives an array of input data, and its size. It applies the tests with these input data, and different kind of threads.

Intermediate tests, to eliminate race conditions. This thread type executes the code 'created'.

In sequential test mode, this will be a 'fake' thread.

Having no threads is equivalent to single-threaded behaviour.

---

Variable Documentation

char* myPtr[NB_SAMPLES] [static]

Global variable because too big for the stack.

char myStr[NB_SAMPLES][RPA_INT_STR] [static]

Global variable because too big for the stack.

---