benchmarks.hpp File Reference

Benchmarking framework for Monte Carlo integration algorithms. More...

#include <montecarlo/domains/hypercylinder.hpp>
#include <montecarlo/domains/hyperrectangle.hpp>
#include <montecarlo/domains/hypersphere.hpp>
#include <montecarlo/geometry.hpp>
#include <montecarlo/integrators/MCintegrator.hpp>
#include <montecarlo/integrators/MHintegrator.hpp>
#include <montecarlo/integrators/ISintegrator.hpp>
#include <montecarlo/proposals/uniformProposal.hpp>
#include <montecarlo/utils/plotter.hpp>
#include <montecarlo/optimizers/PSO.hpp>
#include <montecarlo/optimizers/GA.hpp>
#include <fstream>
#include <iostream>
#include <chrono>
#include <iomanip>
#include <vector>
#include <string>
#include <functional>
#include <thread>
#include <cmath>
#include <cstdint>

Include dependency graph for benchmarks.hpp:

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

Go to the source code of this file.

Classes
struct	results
	Stores benchmark result for a single sample count. More...

Functions
void	saveResults (const std::string &filename, const std::vector< results > &results, const std::string &function_expr)
	Save benchmark results to a CSV file.
void	uniDimIntegration ()
	1D integration test: f(x) = x²
void	circleIntegration ()
	2D circular integration: f(x,y) = x²+y² over unit disk
void	sphereIntegration ()
	3D spherical integration: f(x,y,z) = x²+y²+z² over unit ball
void	rectangularIntegration ()
	2D rectangular integration over axis-aligned box
void	cylinderIntegration ()
	3D cylindrical integration: hypercylinder base with height
void	parallelepipedIntegration ()
	3D parallelepiped (rectangular box) integration
void	fiveDimIntegration ()
	5-dimensional integration benchmark
void	fourDimIntegration ()
	4-dimensional integration benchmark
void	eightDimIntegration ()
	8-dimensional integration benchmark
void	twelveDimIntegration ()
	12-dimensional integration benchmark
void	runBenchmarks (bool useGnuplot)
	Run integration benchmarks with hardcoded integrands.
void	runBenchmarks (const std::string &expression, bool useGnuplot)
	Run integration benchmarks with custom mathematical expression.
void	runBenchmarksMH ()
	Run Metropolis-Hastings MCMC integration benchmarks.
void	runOptimizationBenchmarksPSO ()
void	runOptimizationBenchmarksGA ()

Variables
const std::vector< size_t >	n_samples_vector
	Global benchmark configuration.
unsigned int	n_threads
	Global OpenMP thread count (set at runtime)

Detailed Description

Benchmarking framework for Monte Carlo integration algorithms.

Author

Domenico

Date

12/3/25

Provides infrastructure for evaluating integration accuracy and performance across multiple algorithms (uniform MC, importance sampling, MCMC) and dimensionalities.

Key Features:

• Multiple test domains: 1D, circle, sphere, rectangle, cylinder, parallelepiped, 5D-12D
• Configurable sample counts and thread counts
• Results export to CSV for analysis
• Gnuplot visualization support
• Both hardcoded integrands and parser-based expression support

Benchmark Types:

• Integration benchmarks: convergence and accuracy across sample sizes
• MH-specific benchmarks: burn-in, thinning, and proposal parameter effects
• Optimizer benchmarks: GA and PSO performance (in benchmarks/ subdirectory)

Definition in file benchmarks.hpp.

Function Documentation

circleIntegration()

void circleIntegration

(

)

2D circular integration: f(x,y) = x²+y² over unit disk

cylinderIntegration()

void cylinderIntegration

(

)

3D cylindrical integration: hypercylinder base with height

eightDimIntegration()

void eightDimIntegration

(

)

8-dimensional integration benchmark

fiveDimIntegration()

void fiveDimIntegration

(

)

5-dimensional integration benchmark

fourDimIntegration()

void fourDimIntegration

(

)

4-dimensional integration benchmark

parallelepipedIntegration()

void parallelepipedIntegration

(

)

3D parallelepiped (rectangular box) integration

rectangularIntegration()

void rectangularIntegration

(

)

2D rectangular integration over axis-aligned box

runBenchmarks() [1/2]

void runBenchmarks

(

bool

useGnuplot

)

Run integration benchmarks with hardcoded integrands.

Parameters

useGnuplot

If true, generate gnuplot script for visualization.

Definition at line 292 of file integration_benchmarks.cpp.

Here is the call graph for this function:

runBenchmarks() [2/2]

void runBenchmarks	(	const std::string &	expression,
		bool	useGnuplot
	)

Run integration benchmarks with custom mathematical expression.

Parameters

expression	Mathematical function string (muParserX syntax)
useGnuplot	If true, generate gnuplot script

Definition at line 449 of file integration_benchmarks.cpp.

Here is the call graph for this function:

runBenchmarksMH()

void runBenchmarksMH

(

)

Run Metropolis-Hastings MCMC integration benchmarks.

Definition at line 309 of file integration_benchmarks.cpp.

Here is the call graph for this function:

runOptimizationBenchmarksGA()

void runOptimizationBenchmarksGA

(

)

Definition at line 352 of file ga_benchmarks.cpp.

Here is the call graph for this function:

runOptimizationBenchmarksPSO()

void runOptimizationBenchmarksPSO

(

)

Definition at line 338 of file pso_benchmarks.cpp.

Here is the call graph for this function:

saveResults()

void saveResults	(	const std::string &	filename,
		const std::vector< results > &	results,
		const std::string &	function_expr
	)

Save benchmark results to a CSV file.

Parameters

filename	Output CSV filename
results	Vector of result structures
function_expr	String describing the integrand for reference

Save benchmark results to a CSV file.

Parameters

filename	Output CSV filename
results	Vector of benchmark results {n_samples, integration_result, duration_ms}
function_expr	String description of the integrated function

CSV Format:

• Header: Function description and column labels
• Each row: n_samples \t result \t duration_ms

Useful for:

• Plotting convergence curves: error vs sample count
• Computing empirical convergence rate
• Analyzing computational efficiency
• Comparing algorithm scalability

Definition at line 49 of file benchmarks.cpp.

sphereIntegration()

void sphereIntegration

(

)

3D spherical integration: f(x,y,z) = x²+y²+z² over unit ball

twelveDimIntegration()

void twelveDimIntegration

(

)

12-dimensional integration benchmark

uniDimIntegration()

void uniDimIntegration

(

)

1D integration test: f(x) = x²

Domain: [0,1], Expected: 1/3 ≈ 0.333...

Variable Documentation

n_samples_vector

const std::vector<size_t> n_samples_vector

extern

Global benchmark configuration.

Global benchmark configuration.

Definition at line 28 of file benchmarks.cpp.

n_threads

unsigned int n_threads

extern

Global OpenMP thread count (set at runtime)

Definition at line 31 of file benchmarks.cpp.