public void Benchmarking_for_loop_on_array_shows_that_foreach_is_the_winner()
{
var values = Enumerable.Range(1, 5000).ToArray();
BenchmarkResult rLinqCount = CK.Core.MicroBenchmark.MeasureTime(() =>
{
long sum = 0;
for (int i = 0; i < values.Count(); ++i)
{
sum += values[i];
}
});
BenchmarkResult rLength = CK.Core.MicroBenchmark.MeasureTime(() =>
{
long sum = 0;
for (int i = 0; i < values.Length; ++i)
{
sum += values[i];
}
});
BenchmarkResult rForEach = CK.Core.MicroBenchmark.MeasureTime(() =>
{
long sum = 0;
foreach (int i in values)
{
sum += i;
}
});
rLength.IsTotallyBetterThan(rLinqCount).Should().BeTrue("Count() is ovviously a perf killer.");
rForEach.IsTotallyBetterThan(rLinqCount).Should().BeTrue("Count() is ovviously a perf killer.");
rForEach.IsBetterThan(rLength).Should().BeTrue("foreach is better than for( i < Length ).");
}
public void Benchmarking_CPU_is_less_precise_that_time()
{
int Fib(int n)
{
return(n <= 0 ? n : Fib(n - 1) + Fib(n - 2));
}
BenchmarkResult r1 = CK.Core.MicroBenchmark.MeasureCPU(() => Fib(10));
BenchmarkResult r2 = CK.Core.MicroBenchmark.MeasureCPU(() => Fib(15));
r1.IsBetterThan(r2).Should().BeTrue();
r1.IsSignificantlyBetterThan(r2).Should().BeTrue();
r1.IsTotallyBetterThan(r2).Should().BeTrue();
}
public void Benchmarking_timed_operations()
{
int Fib(int n)
{
return(n <= 0 ? n : Fib(n - 1) + Fib(n - 2));
}
BenchmarkResult r1 = CK.Core.MicroBenchmark.MeasureTime(() => Fib(10));
BenchmarkResult r2 = CK.Core.MicroBenchmark.MeasureTime(() => Fib(11));
r1.IsBetterThan(r2).Should().BeTrue();
r1.IsSignificantlyBetterThan(r2).Should().BeTrue();
r1.IsTotallyBetterThan(r2).Should().BeTrue();
}
