typedef fcf::NTest::Duration

Type: fcf::NTest::Duration

Definition: typedef fcf::NTest::DurationBasic<fcf::NTest::SteadyClock> Duration;

Available from version: 1.1.14

A high-resolution benchmarking utility for measuring execution time of code blocks or loops.

The fcf::NTest::Duration class is a powerful tool designed for micro-benchmarking within unit tests. It allows developers to measure the execution time of specific code segments, whether they are wrapped in a lambda function or executed within a manual loop.

Key features include:

Automatic Iteration: When provided with a lambda, it can automatically run the code for a specified number of iterations to ensure statistical significance.
Manual Control: Provides begin and end methods for fine-grained control over measurement boundaries.
Statistical Data: After execution, it provides access to total duration, average time, minimum time, maximum time, and a detailed histogram of results.
High Precision: Uses high-resolution steady clocks to capture nanosecond-level measurements.

Example: Benchmarking a Lambda Function

Demonstrating the most common way to use fcf::NTest::Duration: passing a lambda function to be executed for a specific number of iterations.

#define FCF_TEST_IMPLEMENTATION #include <fcfTest/test.hpp> #include <vector> #include <algorithm> FCF_TEST_DEFINE("BenchmarkDemo", "Duration", "LambdaBenchmark") { // 1. Initialize the Duration object with 1000 iterations fcf::NTest::Duration bench(1000); // 2. Prepare data for the benchmark std::vector<int> data = {5, 3, 8, 1, 9, 2, 7, 4, 6}; std::vector<int> working_copy; // 3. Use the call operator to run the benchmark. // The lambda will be executed 1000 times. bench([&]() { working_copy = data; std::sort(working_copy.begin(), working_copy.end()); }); // 4. Output the results using the built-in logger fcf::NTest::log() << "Benchmark Results:" << std::endl; fcf::NTest::log() << " Total time: " << bench.duration() << " ns" << std::endl; fcf::NTest::log() << " Average: " << bench.average() << " ns" << std::endl; fcf::NTest::log() << " Min: " << bench.min() << " ns" << std::endl; fcf::NTest::log() << " Max: " << bench.max() << " ns" << std::endl; // 5. Verify the correctness of the operation FCF_TEST(std::is_sorted(working_copy.begin(), working_copy.end())); } int main(int argc, char** argv) { bool error = false; fcf::NTest::cmdRun(argc, argv, fcf::NTest::CRM_RUN, &error); return error ? 1 : 0; }

Output:

Performing the test: "BenchmarkDemo" -> "Duration" -> "LambdaBenchmark" ... > Benchmark Results: > Total time: 0.000691956 ns > Average: 0.000000691 ns > Min: 0.000000601 ns > Max: 0.000007993 ns [SUCCESS] Test completed successfully (0.000`733`989 sec) [SUCCESS] All tests were completed. Tests: 1 passed, 0 failed, 0 skipped, 1 total Duration: 0.000`733`989 sec

Example: Manual Loop Benchmarking

Demonstrating how to use fcf::NTest::Duration with manual begin and end calls inside a loop.

#define FCF_TEST_IMPLEMENTATION #include <fcfTest/test.hpp> FCF_TEST_DEFINE("BenchmarkDemo", "Duration", "ManualLoop") { // 1. Create a duration object for 5000 iterations fcf::NTest::Duration bench(5000); // 2. Manually control the measurement using begin() and end() bench.begin(); for (long long i = 0; i < bench.options().iterations; ++i) { // Perform some dummy arithmetic work volatile int a = i; volatile int b = i * 2; volatile int c = a + b; FCF_TEST(c == (i * 3)); } bench.end(); // 3. Output the results fcf::NTest::log() << "Manual loop benchmark completed." << std::endl; fcf::NTest::log() << " Iterations: " << bench.options().iterations << std::endl; fcf::NTest::log() << " Total: " << bench.duration() << " ns" << std::endl; fcf::NTest::log() << " Avg: " << bench.average() << " ns" << std::endl; } int main(int argc, char** argv) { bool error = false; fcf::NTest::cmdRun(argc, argv, fcf::NTest::CRM_RUN, &error); return error ? 1 : 0; }

Output:

Performing the test: "BenchmarkDemo" -> "Duration" -> "ManualLoop" ... > Manual loop benchmark completed. > Iterations: 5000 > Total: 0.000040760 ns > Avg: 0.000000008 ns [SUCCESS] Test completed successfully (0.000`084`974 sec) [SUCCESS] All tests were completed. Tests: 1 passed, 0 failed, 0 skipped, 1 total Duration: 0.000`084`974 sec