Bench: fft vs jit

src/Benchmark/Transforms/FFT.cs

@@ -11,15 +11,17 @@ namespace Benchmark.Transforms
     {
         readonly Dictionary<int, Complex[]> _data = new Dictionary<int, Complex[]>();
 
-        [Params(32, 64, 128, 1024, 8192, 65536)]
+        // 64=1.5ms, 128=2.9ms, 2048=46.4ms, 8192=185.8ms, 65536=1.5s, 524288=11.9s (44.1 kHz)
+        //[Params(32, 64, 128, 1024, 2048, 4096, 8192, 65536, 1048576)]
+        [Params(128, 1024, 1048576)]
         public int N { get; set; }
 
         public FFT()
         {
-            var realSinusoidal = Generate.Sinusoidal(65536, 32, -2.0, 2.0);
-            var imagSawtooth = Generate.Sawtooth(65536, 32, -20.0, 20.0);
+            var realSinusoidal = Generate.Sinusoidal(1048576, 32, -2.0, 2.0);
+            var imagSawtooth = Generate.Sawtooth(1048576, 32, -20.0, 20.0);
             var signal = Generate.Map2(realSinusoidal, imagSawtooth, (r, i) => new Complex(r, i));
-            foreach (var n in new[] { 32, 64, 128, 1024, 8192, 65536 })
+            foreach (var n in new[] { 32, 64, 128, 1024, 2048, 4096, 8192, 65536, 1048576 })
             {
                 var s = new Complex[n];
                 Array.Copy(signal, 0, s, 0, n);