Helper functions and classes to simplify unit tests. More...

Namespaces
	defaults
	Define the default values for a microbenchmark config.

	detail

Classes
struct	create_random_sample
	Wrap a random number generator so it can be used for both real and complex numbers. More...

struct	create_random_sample< std::complex< T > >
	Specialize create_random_sample for complex numbers. More...

struct	extrapolate_periodic
	A functor to extrapolate a periodic timseries. More...

struct	extrapolate_with_zeroes
	A functor to extrapolate with zeroes. More...

class	has_resize
	Type trait to evaluate if a collection of numbers (e.g. More...

class	microbenchmark
	Run a micro-benchmark on a given class. More...

class	microbenchmark_base
	Refactor non-template parts of the microbenchmark template class. More...

class	microbenchmark_config
	Configure a micro-benchmark. More...

Typedefs
template<typename config >
using	microbenchmark_group = std::map< std::string, std::function< void(config const &cfg)> >
	Define a representation for a group of microbenchmarks. More...

Functions
template<typename value_t , typename std::enable_if< std::is_integral< value_t >::value >::type * = nullptr>
bool	check_close_enough (value_t num_a, value_t num_b, int tol)
	Given two numbers of the same integer type check if the difference is within a given tolerance. More...

template<typename value_t >
bool	check_close_enough (std::complex< value_t > com_a, std::complex< value_t > com_b, int tol=1)
	Given two complex numbers of the same value type check if the difference is within a given tolerance. More...

template<typename value_t >
bool	check_close_enough (value_t com_a[2], value_t com_b[2], int tol=1)
	Compare complex number (as arrays of size 2) values with some tolerance. More...

template<typename floating >
void	check_small (floating t, double small)
	Verify that a floating point value is "close enough" to a small number. More...

template<typename real >
void	check_small (std::complex< real > t, double small)
	Verify that a complex number is "close enough" to a small number. More...

template<typename precision_t >
std::complex< precision_t >	format (precision_t v[2])
	Wrap FFTW-style complex numbers in std::complex for iostreaming. More...

template<typename value_type >
value_type	format (value_type t)
	Allow generic treatment of FFTW-style complex numbers and other types. More...

template<typename real , typename std::enable_if< std::is_floating_point< real >::value >::type * = nullptr>
real	relative_error (real actual, real expected)
	Calculate the relative error between two float point numbers. More...

template<typename integral , typename std::enable_if< std::is_integral< integral >::value >::type * = nullptr>
integral	relative_error (integral actual, integral expected)
	Adapt relative_error() for integral numbers. More...

template<typename real >
real	relative_error (std::complex< real > actual, std::complex< real > expected)
	Calculate the relative error between two complex numbers. More...

template<typename real >
real	relative_error (real actual[2], real expected[2])
	Calculate the relative error between two complex numbers in FFTW representation. More...

template<typename collection_t >
bool	check_collection_close_enough (collection_t const &a, collection_t const &b, int tol=1, int max_differences_printed=JB_TESTING_MAX_DIFFERENCES_PRINTED)
	Given two collections of numbers of the same value type, find the differences that are out of a given tolerance and report them via Boost.Test functions. More...

template<typename T , std::size_t K>
bool	check_collection_close_enough (boost::multi_array< T, K > const &a, boost::multi_array< T, K > const &b, int tol=1, int max_differences_printed=JB_TESTING_MAX_DIFFERENCES_PRINTED)
	Specialization for boost::multi_array type. More...

template<typename value_t >
bool	check_collection_close_enough (std::size_t size, value_t const a, value_t const b, int tol=1, int max_differences_printed=JB_TESTING_MAX_DIFFERENCES_PRINTED)
	Given two collections of integers, floating point or complex numbers find the differences that are out of a given tolerance and report them via Boost.Test functions. More...

template<typename timeseries , typename generator >
void	create_random_timeseries (generator &gen, int nsamples, timeseries &ts)
	Create a simple timeseries where the values look like a random. More...

template<typename timeseries >
void	create_square_timeseries (int nsamples, timeseries &ts)
	Create a simple timeseries where the values look like a square. More...

template<typename timeseries >
void	create_triangle_timeseries (int nsamples, timeseries &ts)
	Create a simple timeseries where the values look like a triangle. More...

template<typename T , std::size_t K, typename A >
void	create_triangle_timeseries (int nsamples, boost::multi_array< T, K, A > &ts)
	Create a family of timeseries where the values look like parallel triangles shifted by a constant value. More...

template<typename timeseries_t , typename duration_t , typename extrapolation_functor >
jb::detail::array_traits< timeseries_t >::element_type	extrapolate_timeseries (timeseries_t const &ts, duration_t t, duration_t sampling_period, extrapolation_functor const &extrapolation)
	A function to get the extrapolated value of a timeseries. More...

template<typename timeseries_t , typename duration_t , typename extrapolation_functor >
timeseries_t	delay_timeseries (timeseries_t const &ts, duration_t delay, duration_t sampling_period, extrapolation_functor const &extrapolation)
	Delay a timeseries using a user-provided extrapolation policy. More...

template<typename timeseries_t , typename duration_t >
timeseries_t	delay_timeseries_periodic (timeseries_t const &ts, duration_t delay, duration_t sampling_period)
	Delay a timeseries using a periodic extension for early values. More...

template<typename timeseries_t , typename duration_t >
timeseries_t	delay_timeseries_zeroes (timeseries_t const &ts, duration_t delay, duration_t sampling_period)
	Delay a timeseries using zeroes for early values. More...

template<typename Generator >
Generator	initialize_mersenne_twister (int seed, std::string const &token)
	Initialize a Mersenne-Twister based generator. More...

std::ostream &	operator<< (std::ostream &os, microbenchmark_base::summary const &x)
	Stream the summary of a microbenchmark results in microseconds. More...

int	microbenchmark_group_main (int argc, char *argv[], microbenchmark_group< microbenchmark_config > const &testcases)
	Overload microbenchmark_group_main for jb::testing::microbenchmark_config. More...

template<typename config >
int	microbenchmark_group_main (int argc, char *argv[], microbenchmark_group< config > const &testcases)
	Implement the main() function for a multi-testcase benchmark. More...

template<typename timeseries >
void	resize_if_applicable (timeseries &ts, std::size_t newsize, std::true_type)
	Resize a timeseries to a newsize. More...

template<typename timeseries >
void	resize_if_applicable (timeseries &ts, std::size_t newsize, std::false_type)
	Dummy specialization for non-resizable timeseries collections. More...

template<typename timeseries >
void	resize_if_applicable (timeseries &ts, std::size_t newsize)
	Resize a timeseries to a newsize if it is resizable. More...

template<typename container_t >
auto	sum_square (container_t const &ts)
	Compute the sum square of a family of timeseries. More...

Variables
char const	uname_a []
	The kernel version, release, machine hardware, etc. More...

char const	compiler []
	The compiler version. More...

char const	compiler_flags []
	The compiler flags. More...

char const	linker []
	The linker version. More...

char const	gitrev []
	The git revision at compile time. More...

char const	default_initialization_marker [] = "__default__"

Detailed Description

Helper functions and classes to simplify unit tests.

Typedef Documentation

◆ microbenchmark_group

template<typename config >

using jb::testing::microbenchmark_group = typedef std::map<std::string, std::function<void(config const& cfg)> >

Define a representation for a group of microbenchmarks.

Each microbenchmark is encapsulated in a (type erased) std::function, which typically creates an instance of jb::testing::microbenchmark and runs it. The keys into the group are the names of the testcases.

Template Parameters

config the configuration class for the group of microbenchmarks.

Definition at line 33 of file microbenchmark_group_main.hpp.

Function Documentation

◆ check_close_enough() [1/3]

template<typename value_t , typename std::enable_if< std::is_integral< value_t >::value >::type * = nullptr>

bool jb::testing::check_close_enough	(	value_t	num_a,
		value_t	num_b,
		int	tol
	)

Given two numbers of the same integer type check if the difference is within a given tolerance.

Given two numbers of the same floating-point type check if the difference is within a given tolerance.

Tolerance is expressed as the difference between any pair of integer numbers.

Returns: true if the numbers are within tolerance.

Parameters

num_a	the first number to compare
num_b	the second number to compare
tol	tolerance that the pair of numbers have to be within

Template Parameters

value_t the type of the integral argument numbers

Tolerance is expressed in "number of epsilons" (i.e., the function will tolerate errors up to tolerance * std::numeric_limit<real>::epsilon()

Returns: true if the numbers are within tolerance.

Parameters

num_a	the first number to compare
num_b	the second number to compare
tol	tolerance that the pair of numbers have to be within

Template Parameters

value_t the type of the floating-point argument numbers

Definition at line 34 of file check_close_enough.hpp.

Referenced by BOOST_AUTO_TEST_CASE(), check_close_enough(), and check_collection_close_enough().

◆ check_close_enough() [2/3]

template<typename value_t >

bool jb::testing::check_close_enough	(	std::complex< value_t >	com_a,
		std::complex< value_t >	com_b,
		int	tol = `1`
	)

Given two complex numbers of the same value type check if the difference is within a given tolerance.

If value types are floating-point numbers, tolerance is expressed in "number of epsilons" (i.e., the function will tolerate errors up to tolerance * std::numeric_limit<real>::epsilon()

If value types are integer numbers, tolerance is expressed as the difference between any pair of integer numbers.

Returns: true if the both the real and imaginary parts are within tolerance

Parameters

com_a	the first complex number to compare
com_b	the second complex number to compare
tol	tolerance that the pair of numbers have to be within

Template Parameters

value_type the type of std::complex number

Definition at line 95 of file check_close_enough.hpp.

References check_close_enough().

◆ check_close_enough() [3/3]

template<typename value_t >

bool jb::testing::check_close_enough	(	value_t	com_a[2],
		value_t	com_b[2],
		int	tol = `1`
	)

Compare complex number (as arrays of size 2) values with some tolerance.

Returns: true if the both the real and imaginary parts are within tolerance.

Parameters

com_a	the first complex number to compare
com_b	the second complex number to compare
tol	tolerance that the pair of numbers have to be within

Template Parameters

value_type the type of std::complex number

Definition at line 115 of file check_close_enough.hpp.

References check_close_enough().

◆ check_collection_close_enough() [1/3]

template<typename collection_t >

bool jb::testing::check_collection_close_enough	(	collection_t const &	a,
		collection_t const &	b,
		int	tol = `1`,
		int	max_differences_printed = `JB_TESTING_MAX_DIFFERENCES_PRINTED`
	)

Given two collections of numbers of the same value type, find the differences that are out of a given tolerance and report them via Boost.Test functions.

Collections of numbers are any container representation (e.g. vector, deque), boost::multi_array, c-like array, etc.

Value types are integers, floating-point, and std::complex (integers or floating-point).

If value types are floating-point numbers, tolerance is expressed in "number of epsilons" (i.e., the function will tolerate errors up to tolerance * std::numeric_limit<real>::epsilon()

If value types are integer numbers, tolerance is expressed as the difference between any pair of integer numbers.

Returns: true if the numbers are within tolerance.

Parameters

a	the first collection of numbers to compare
b	the second collection of numbers to compare
tol	tolerance that each pair of numbers have to be within
max_differences_printed	how many differences will be printed out in full detail, some of the collections are large and printing all the differences can be overwhelming

Template Parameters

collection_t the type of the collection containing the numbers (e.g. std::vector<>, boost::multi_array)

Definition at line 259 of file check_close_enough.hpp.

References check_close_enough(), jb::error, format(), and relative_error().

Referenced by BOOST_AUTO_TEST_CASE(), and BOOST_AUTO_TEST_CASE_EXPECTED_FAILURES().

◆ check_collection_close_enough() [2/3]

template<typename T , std::size_t K>

bool jb::testing::check_collection_close_enough	(	boost::multi_array< T, K > const &	a,
		boost::multi_array< T, K > const &	b,
		int	tol = `1`,
		int	max_differences_printed = `JB_TESTING_MAX_DIFFERENCES_PRINTED`
	)

Specialization for boost::multi_array type.

Returns: true if the numbers are within tolerance.

Parameters

a	the first multi array of numbers to compare
b	the second multi array of numbers to compare
tol	tolerance that each pair of numbers have to be within
max_differences_printed	how many differences will be printed out in full detail, some of the collections are large and printing all the differences can be overwhelming

Template Parameters

T	type of the value stored on the timeseries
K	timeseries dimensionality
A	an Allocator type for type T allocator storage

Definition at line 299 of file check_close_enough.hpp.

References check_close_enough(), jb::error, format(), and relative_error().

◆ check_collection_close_enough() [3/3]

template<typename value_t >

bool jb::testing::check_collection_close_enough	(	std::size_t	size,
		value_t const *	a,
		value_t const *	b,
		int	tol = `1`,
		int	max_differences_printed = `JB_TESTING_MAX_DIFFERENCES_PRINTED`
	)

Given two collections of integers, floating point or complex numbers find the differences that are out of a given tolerance and report them via Boost.Test functions.

Returns: true if the numbers are within tolerance.

Parameters

size	the size of collection
a	the first multi array of numbers to compare
b	the second multi array of numbers to compare
tol	tolerance that each pair of numbers have to be within
max_differences_printed	how many differences will be printed out in full detail, some of the collections are large and printing all the differences can be overwhelming

Template Parameters

value_t value type of the number stored on collections

Definition at line 342 of file check_close_enough.hpp.

References check_close_enough(), jb::error, format(), relative_error(), and jb::testing::defaults::size.

◆ check_small() [1/2]

template<typename floating >

void jb::testing::check_small	(	floating	t,
		double	small
	)

Verify that a floating point value is "close enough" to a small number.

Definition at line 125 of file check_close_enough.hpp.

◆ check_small() [2/2]

template<typename real >

void jb::testing::check_small	(	std::complex< real >	t,
		double	small
	)

Verify that a complex number is "close enough" to a small number.

Definition at line 133 of file check_close_enough.hpp.

◆ create_random_timeseries()

template<typename timeseries , typename generator >

void jb::testing::create_random_timeseries	(	generator &	gen,
		int	nsamples,
		timeseries &	ts
	)

Create a simple timeseries where the values look like a random.

Definition at line 13 of file create_random_timeseries.hpp.

References jb::detail::nsamples().

Referenced by BOOST_AUTO_TEST_CASE().

◆ create_square_timeseries()

template<typename timeseries >

void jb::testing::create_square_timeseries	(	int	nsamples,
		timeseries &	ts
	)

Create a simple timeseries where the values look like a square.

Definition at line 13 of file create_square_timeseries.hpp.

References jb::detail::nsamples(), and resize_if_applicable().

Referenced by BOOST_AUTO_TEST_CASE(), and main().

◆ create_triangle_timeseries() [1/2]

template<typename timeseries >

void jb::testing::create_triangle_timeseries	(	int	nsamples,
		timeseries &	ts
	)

Create a simple timeseries where the values look like a triangle.

Definition at line 16 of file create_triangle_timeseries.hpp.

References resize_if_applicable().

Referenced by BOOST_AUTO_TEST_CASE(), and main().

◆ create_triangle_timeseries() [2/2]

template<typename T , std::size_t K, typename A >

void jb::testing::create_triangle_timeseries	(	int	nsamples,
		boost::multi_array< T, K, A > &	ts
	)

Create a family of timeseries where the values look like parallel triangles shifted by a constant value.

Parameters

nsamples	samples on each timeseries
ts	multi array containing timeseries to be filled as triangle

Template Parameters

T	type of the timeseries field
K	timeseries dimensionality
A	an Allocator type for type T allocator storage

The values stored in the input timeseries

Definition at line 38 of file create_triangle_timeseries.hpp.

References jb::detail::element_count(), and jb::detail::nsamples().

◆ delay_timeseries()

template<typename timeseries_t , typename duration_t , typename extrapolation_functor >

timeseries_t jb::testing::delay_timeseries	(	timeseries_t const &	ts,
		duration_t	delay,
		duration_t	sampling_period,
		extrapolation_functor const &	extrapolation
	)

Delay a timeseries using a user-provided extrapolation policy.

The values stored in the input timeseries ts

Definition at line 78 of file delay_timeseries.hpp.

References jb::detail::array_shape(), jb::detail::element_count(), and extrapolate_timeseries().

Referenced by delay_timeseries_periodic(), and delay_timeseries_zeroes().

◆ delay_timeseries_periodic()

template<typename timeseries_t , typename duration_t >

timeseries_t jb::testing::delay_timeseries_periodic	(	timeseries_t const &	ts,
		duration_t	delay,
		duration_t	sampling_period
	)

Delay a timeseries using a periodic extension for early values.

The values stored in the input timeseries ts

Definition at line 99 of file delay_timeseries.hpp.

References delay_timeseries().

Referenced by BOOST_AUTO_TEST_CASE(), and main().

◆ delay_timeseries_zeroes()

template<typename timeseries_t , typename duration_t >

timeseries_t jb::testing::delay_timeseries_zeroes	(	timeseries_t const &	ts,
		duration_t	delay,
		duration_t	sampling_period
	)

Delay a timeseries using zeroes for early values.

The values stored in the input timeseries ts

Definition at line 112 of file delay_timeseries.hpp.

References delay_timeseries().

Referenced by BOOST_AUTO_TEST_CASE().

◆ extrapolate_timeseries()

template<typename timeseries_t , typename duration_t , typename extrapolation_functor >

jb::detail::array_traits<timeseries_t>::element_type jb::testing::extrapolate_timeseries	(	timeseries_t const &	ts,
		duration_t	t,
		duration_t	sampling_period,
		extrapolation_functor const &	extrapolation
	)

A function to get the extrapolated value of a timeseries.

Parameters

ts	a timeseries, typically a std::vector-like data structure
t	the timestamp at which we want to extrapolate the timeseries
sampling_period	the sampling period of the timeseries
extrapolation	implement the extrapolation

Template Parameters

timeseries_t	the type of the timeseries
duration_t	the type used to represent timestamps, typically a std::chrono::duration<> instantiation
extrapolation_functor	the type of the extrapolation

Definition at line 62 of file delay_timeseries.hpp.

References jb::detail::element_count().

Referenced by delay_timeseries().

◆ format() [1/2]

template<typename precision_t >

std::complex<precision_t> jb::testing::format ( precision_t v[2] )

Wrap FFTW-style complex numbers in std::complex for iostreaming.

Definition at line 140 of file check_close_enough.hpp.

Referenced by check_collection_close_enough().

◆ format() [2/2]

template<typename value_type >

value_type jb::testing::format ( value_type t )

Allow generic treatment of FFTW-style complex numbers and other types.

Definition at line 146 of file check_close_enough.hpp.

◆ initialize_mersenne_twister()

template<typename Generator >

Generator jb::testing::initialize_mersenne_twister	(	int	seed,
		std::string const &	token
	)

Initialize a Mersenne-Twister based generator.

Often tests need to initialize a PRNG based on either a command-line argument (to make tests reproduceable), or from the std::random_device (to have good distribution properties in the PRGN). This is a fairly tedious initialization, so we have refactored it to a common function.

Parameters

seed	the command-line argument seed, 0 means initialize from std::random_device.
token	the parameter to initialize std::random_device, if default then the random device is default initialized. The semantics of token are platform specific, but on Linux it is usually the device to read random numbers from (/dev/urandom or /dev/random). It is very rare to need /dev/random, specially in tests, where this function is most commonly used.

Returns: a Generator initialized as described above.

Exceptions

std::exception if the token is invalid on the platform.

Template Parameters

Generator an instantiation of std::mersenne_twister_generator<>, most likely std::mt19937 or std::mt19937_64.

Definition at line 38 of file initialize_mersenne_twister.hpp.

◆ microbenchmark_group_main() [1/2]

int jb::testing::microbenchmark_group_main	(	int	argc,
		char *	argv[],
		microbenchmark_group< microbenchmark_config > const &	testcases
	)

Overload microbenchmark_group_main for jb::testing::microbenchmark_config.

A common idiom in complex benchmarks is to create multiple testcases that exercise different implementations, different template instantiations, or otherwise vary the test in a non-trivial way.

In such cases the main() function follows a well understood pattern:

Initialize the benchmark configuration.
Find which testcase is requested.
Call a function for that testcase.

Returns: the exit status for the program

Parameters

argc	the number of arguments in argv
argv	the command-line arguments
testcases	a group of microbenchmarks, executes the one selected via the –microbenchmark.test-case command-line argument.

Exceptions

nothing,all exceptions are captured and printed to stderr before the program exits.

Definition at line 6 of file microbenchmark_group_main.cpp.

References jb::info, JB_LOG, jb::config_object::process_cmdline(), jb::testing::detail::report_exception_at_exit(), jb::testing::microbenchmark_config::test_case, and jb::testing::microbenchmark_config::verbose.

Referenced by main(), and microbenchmark_group_main().

◆ microbenchmark_group_main() [2/2]

template<typename config >

int jb::testing::microbenchmark_group_main	(	int	argc,
		char *	argv[],
		microbenchmark_group< config > const &	testcases
	)

Implement the main() function for a multi-testcase benchmark.

A common idiom in complex benchmarks is to create multiple testcases that exercise different implementations, different template instantiations, or otherwise vary the test in a non-trivial way.

In such cases the main() function follows a well understood pattern:

Initialize the benchmark configuration.
Find which testcase is requested.
Call a function for that testcase.

This class refactors that pattern into a static function.

Template Parameters

config

the configuration object used by the benchmark. We expect that it has at least two configuration attributes:

log: of type jb::log::config
microbenchmark: of type jb::testing::microbenchmark_config

Returns: the exit status for the program

Parameters

argc	the number of arguments in argv
argv	the command-line arguments
testcases	a group of microbenchmarks, executes the one selected via the –microbenchmark.test-case command-line argument.

Exceptions

nothing,all exceptions are captured and printed to stderr before the program exits.

Definition at line 66 of file microbenchmark_group_main.hpp.

References jb::info, jb::log::init(), JB_LOG, microbenchmark_group_main(), and jb::testing::detail::report_exception_at_exit().

◆ operator<<()

std::ostream & jb::testing::operator<<	(	std::ostream &	os,
		microbenchmark_base::summary const &	x
	)

Stream the summary of a microbenchmark results in microseconds.

Definition at line 65 of file microbenchmark_base.cpp.

References jb::testing::microbenchmark_base::summary::max, jb::testing::microbenchmark_base::summary::min, jb::testing::microbenchmark_base::summary::n, jb::testing::microbenchmark_base::summary::p25, jb::testing::microbenchmark_base::summary::p50, jb::testing::microbenchmark_base::summary::p75, jb::testing::microbenchmark_base::summary::p90, jb::testing::microbenchmark_base::summary::p99, and jb::testing::microbenchmark_base::summary::p99_9.

◆ relative_error() [1/4]

template<typename real , typename std::enable_if< std::is_floating_point< real >::value >::type * = nullptr>

real jb::testing::relative_error	(	real	actual,
		real	expected
	)

Calculate the relative error between two float point numbers.

Returns: the relative error of actual compared to offset.

Parameters

actual	the number the test got
expected	the number the test expected

Template Parameters

real	the type of floating point number (float, double, long double)

Definition at line 163 of file check_close_enough.hpp.

Referenced by check_collection_close_enough(), and relative_error().

◆ relative_error() [2/4]

template<typename integral , typename std::enable_if< std::is_integral< integral >::value >::type * = nullptr>

integral jb::testing::relative_error	(	integral	actual,
		integral	expected
	)

Adapt relative_error() for integral numbers.

Returns: the absolute difference.

Parameters

actual	the number the test got
expected	the number the test expected

Template Parameters

integral the type of integral number

Definition at line 186 of file check_close_enough.hpp.

◆ relative_error() [3/4]

template<typename real >

real jb::testing::relative_error	(	std::complex< real >	actual,
		std::complex< real >	expected
	)

Calculate the relative error between two complex numbers.

Returns: the relative error of actual compared to offset, using the Manhattan metric.

Parameters

actual	the number the test got
expected	the number the test expected

Template Parameters

real	the type of floating point number (float, double, long double)

Definition at line 205 of file check_close_enough.hpp.

References relative_error().

◆ relative_error() [4/4]

template<typename real >

real jb::testing::relative_error	(	real	actual[2],
		real	expected[2]
	)

Calculate the relative error between two complex numbers in FFTW representation.

Returns: the relative error of actual compared to offset, using the Manhattan metric.

Parameters

actual	the number the test got
expected	the number the test expected

Template Parameters

real	the type of floating point number (float, double, long double)

Definition at line 224 of file check_close_enough.hpp.

References relative_error().

◆ resize_if_applicable() [1/3]

template<typename timeseries >

void jb::testing::resize_if_applicable	(	timeseries &	ts,
		std::size_t	newsize,
		std::true_type
	)

Resize a timeseries to a newsize.

Parameters

ts	timeseries are any container representation (e.g. vector, deque), boost::multi_array, c-like array, etc
newsize	size to resize the timeseries if possible

Template Parameters

timeseries type of the timeseries collection representation

Definition at line 46 of file resize_if_applicable.hpp.

Referenced by create_square_timeseries(), create_triangle_timeseries(), and resize_if_applicable().

◆ resize_if_applicable() [2/3]

template<typename timeseries >

void jb::testing::resize_if_applicable	(	timeseries &	ts,
		std::size_t	newsize,
		std::false_type
	)

Dummy specialization for non-resizable timeseries collections.

Parameters

ts	timeseries are any container representation (e.g. vector, deque), boost::multi_array, c-like array, etc
newsize	size to resize the timeseries if possible

Template Parameters

timeseries type of the timeseries collection representation

Definition at line 59 of file resize_if_applicable.hpp.

◆ resize_if_applicable() [3/3]

template<typename timeseries >

void jb::testing::resize_if_applicable	(	timeseries &	ts,
		std::size_t	newsize
	)

Resize a timeseries to a newsize if it is resizable.

Parameters

ts	timeseries are any container representation (e.g. vector, deque), boost::multi_array, c-like array, etc
newsize	size to resize the timeseries if possible

Template Parameters

timeseries type of the timeseries collection representation

Definition at line 72 of file resize_if_applicable.hpp.

References resize_if_applicable().

◆ sum_square()

template<typename container_t >

auto jb::testing::sum_square ( container_t const & ts )

Compute the sum square of a family of timeseries.

Returns: tde_result with the square sum of each timeseries

Template Parameters

container_t type of the collection of timeseries

Extract T out of std::complex<T>, otherwise simply T.

Definition at line 17 of file sum_square.hpp.

References jb::detail::element_count(), and jb::detail::nsamples().

Referenced by BOOST_AUTO_TEST_CASE(), and main().

Variable Documentation

◆ compiler

char const jb::testing::compiler[]

The compiler version.

Referenced by main().

◆ compiler_flags

char const jb::testing::compiler_flags[]

The compiler flags.

Referenced by main().

◆ default_initialization_marker

char const jb::testing::default_initialization_marker[] = "__default__"

Definition at line 11 of file initialize_mersenne_twister.hpp.

Referenced by main().

◆ gitrev

char const jb::testing::gitrev[]

The git revision at compile time.

Referenced by main().

◆ linker

char const jb::testing::linker[]

The linker version.

Referenced by main().

◆ uname_a

char const jb::testing::uname_a[]

The kernel version, release, machine hardware, etc.

Referenced by main().

Namespaces

Classes

Typedefs

Functions

Variables

Detailed Description

Typedef Documentation

◆ microbenchmark_group

Function Documentation

◆ check_close_enough() [1/3]

◆ check_close_enough() [2/3]

◆ check_close_enough() [3/3]

◆ check_collection_close_enough() [1/3]

◆ check_collection_close_enough() [2/3]

◆ check_collection_close_enough() [3/3]

◆ check_small() [1/2]

◆ check_small() [2/2]

◆ create_random_timeseries()

◆ create_square_timeseries()

◆ create_triangle_timeseries() [1/2]

◆ create_triangle_timeseries() [2/2]

◆ delay_timeseries()

◆ delay_timeseries_periodic()

◆ delay_timeseries_zeroes()

◆ extrapolate_timeseries()

◆ format() [1/2]

◆ format() [2/2]

◆ initialize_mersenne_twister()

◆ microbenchmark_group_main() [1/2]

◆ microbenchmark_group_main() [2/2]

◆ operator<<()

◆ relative_error() [1/4]

◆ relative_error() [2/4]

◆ relative_error() [3/4]

◆ relative_error() [4/4]

◆ resize_if_applicable() [1/3]

◆ resize_if_applicable() [2/3]

◆ resize_if_applicable() [3/3]

◆ sum_square()

Variable Documentation

◆ compiler

◆ compiler_flags

◆ default_initialization_marker

◆ gitrev

◆ linker

◆ uname_a