testquark 0.1.0

testquark

This is a benchmarking framework. Use as follows

int[] generate(size_t x) ///Sanity cheque!
    out(v; v.length == x)
{
    auto tmp = new int[x];
    foreach (ref i; tmp)
    {
        import std.random;
        i = uniform(int.min, int.max);
    }
    return tmp;
}
//                                                              V You specify the range over which to test here
@Parameterizer("Measure [our new favourite] sorting algorithm", iota(1, 300), &generate)
//                                                                            ^ Generate your own data, e.g. array[N]
int[] sort(int[] data)
//         ^Generic, but built for arrays
{
    import std.algorithm : sort;
    return data.sort.array;
}
//This is then triggered elsewhere, i.e. 
//mixin BenchmarkInfrastructure!(No.Expose);

You will soon be able to use benchquark to measure performance counters of your benchmarks (e.g. cache, Instructions executed and what the branch predictors been up to) by choosing Yes.Expose

Using PAPI:

Pass Yes.UsePapi to runAndPrint(). You must link PAPI 5 in some way, i.e. in "libs" in dub.json or `pragma(lib, "papi")'. 

PAPI 5 can be obtained through apt.