benchmark.hpp File Reference

Wall-clock + cycle-accurate benchmarking for performance profiling. More...

#include <chrono>
#include <functional>
#include <iostream>
#include <string>

Go to the source code of this file.

Functions
void	benchmark (const std::string &name, int trials, std::function< int()> func)
	Benchmark wrapper that logs wall time and CPU cycles.

Detailed Description

Wall-clock + cycle-accurate benchmarking for performance profiling.

Wraps a simulation function and logs:

• Wall time in nanoseconds + seconds
• CPU cycles (via rdtsc or cntvct_el0)
• π estimate from number of hits

Features

• Cross-platform CPU cycle counting (x86 + ARM)
• High-resolution std::chrono timer
• Printable results for logging or scripting

Example

benchmark("SIMD", 10000000, [&]() {
    return monteCarloPI_SEQUENTIAl(10000000);
});

Definition in file benchmark.hpp.

Function Documentation

benchmark()

void benchmark	(	const std::string &	name,
		int	trials,
		std::function< int()>	func )

Benchmark wrapper that logs wall time and CPU cycles.

Times the execution of a function using both std::chrono. Outputs hit count and estimate of π. duration in seconds and nanoseconds

Parameters

name	Name of the benchmark (e.g., "SIMD")
trials	Total number of Monte Carlo trials
func	Function that returns number of hits

Definition at line 44 of file benchmark.hpp.

                                                                                  {
    auto start = std::chrono::high_resolution_clock::now();
 
    int hits = func();
 
    auto end = std::chrono::high_resolution_clock::now();
 
    double piEstimate = 4.0 * hits / trials;
    auto elapsed_ns = std::chrono::duration_cast<std::chrono::nanoseconds>(end - start).count();
 
    std::cout << name << ":\n"
              << "  Trials: " << trials << "\n"
              << "  Hits: " << hits << "\n"
              << "  Estimate: " << piEstimate << "\n"
              << "  Time: " << (elapsed_ns / 1e9) << "s (" << elapsed_ns << " ns)\n";
}