-------------------------------------------------------------------------------
scramble_fft

Testing FFT of size 256 (rank = 8)...
Time = 30.0 s, iterations = 30707000, performance = 1023533.9 [i/s], average time = 0.00098 [ms/i]
Testing FFT of size 512 (rank = 9)...
Time = 30.0 s, iterations = 16442000, performance = 548052.6 [i/s], average time = 0.00182 [ms/i]
Testing FFT of size 1024 (rank = 10)...
Time = 30.0 s, iterations = 8535000, performance = 284481.6 [i/s], average time = 0.00352 [ms/i]
Testing FFT of size 2048 (rank = 11)...
Time = 30.0 s, iterations = 4002000, performance = 133392.5 [i/s], average time = 0.00750 [ms/i]
Testing FFT of size 4096 (rank = 12)...
Time = 30.0 s, iterations = 1368000, performance = 45568.6 [i/s], average time = 0.02194 [ms/i]
Testing FFT of size 8192 (rank = 13)...
Time = 30.0 s, iterations = 605000, performance = 20133.5 [i/s], average time = 0.04967 [ms/i]
Testing FFT of size 16384 (rank = 14)...
Time = 30.0 s, iterations = 298000, performance = 9931.3 [i/s], average time = 0.10069 [ms/i]
Testing FFT of size 32768 (rank = 15)...
Time = 30.2 s, iterations = 122000, performance = 4039.7 [i/s], average time = 0.24755 [ms/i]
Testing FFT of size 65536 (rank = 16)...
Time = 30.5 s, iterations = 44000, performance = 1443.5 [i/s], average time = 0.69277 [ms/i]

-------------------------------------------------------------------------------
scramble_fft + start_fft

Testing FFT of size 256 (rank = 8)...
Time = 30.0 s, iterations = 14553000, performance = 485089.1 [i/s], average time = 0.00206 [ms/i]
Testing FFT of size 512 (rank = 9)...
Time = 30.0 s, iterations = 7200000, performance = 239980.1 [i/s], average time = 0.00417 [ms/i]
Testing FFT of size 1024 (rank = 10)...
Time = 30.0 s, iterations = 4059000, performance = 135267.7 [i/s], average time = 0.00739 [ms/i]
Testing FFT of size 2048 (rank = 11)...
Time = 30.0 s, iterations = 2059000, performance = 68620.0 [i/s], average time = 0.01457 [ms/i]
Testing FFT of size 4096 (rank = 12)...
Time = 30.0 s, iterations = 846000, performance = 28192.7 [i/s], average time = 0.03547 [ms/i]
Testing FFT of size 8192 (rank = 13)...
Time = 30.0 s, iterations = 391000, performance = 13023.4 [i/s], average time = 0.07678 [ms/i]
Testing FFT of size 16384 (rank = 14)...
Time = 30.0 s, iterations = 190000, performance = 6328.5 [i/s], average time = 0.15802 [ms/i]
Testing FFT of size 32768 (rank = 15)...
Time = 30.1 s, iterations = 84000, performance = 2789.3 [i/s], average time = 0.35851 [ms/i]
Testing FFT of size 65536 (rank = 16)...
Time = 30.7 s, iterations = 34000, performance = 1109.1 [i/s], average time = 0.90160 [ms/i]

-------------------------------------------------------------------------------
scramble_fft + start_fft (unoptimized shufps)

Testing FFT of size 256 (rank = 8)...
Time = 30.0 s, iterations = 29129000, performance = 970959.9 [i/s], average time = 0.00103 [ms/i]
Testing FFT of size 512 (rank = 9)...
Time = 30.0 s, iterations = 15380000, performance = 512639.0 [i/s], average time = 0.00195 [ms/i]
Testing FFT of size 1024 (rank = 10)...
Time = 30.0 s, iterations = 8143000, performance = 271413.0 [i/s], average time = 0.00368 [ms/i]
Testing FFT of size 2048 (rank = 11)...
Time = 30.0 s, iterations = 3412000, performance = 113722.2 [i/s], average time = 0.00879 [ms/i]
Testing FFT of size 4096 (rank = 12)...
Time = 30.0 s, iterations = 1222000, performance = 40723.3 [i/s], average time = 0.02456 [ms/i]
Testing FFT of size 8192 (rank = 13)...
Time = 30.0 s, iterations = 535000, performance = 17830.2 [i/s], average time = 0.05608 [ms/i]
Testing FFT of size 16384 (rank = 14)...
Time = 30.0 s, iterations = 269000, performance = 8955.7 [i/s], average time = 0.11166 [ms/i]
Testing FFT of size 32768 (rank = 15)...
Time = 30.0 s, iterations = 111000, performance = 3695.9 [i/s], average time = 0.27057 [ms/i]
Testing FFT of size 65536 (rank = 16)...
Time = 30.4 s, iterations = 41000, performance = 1347.5 [i/s], average time = 0.74209 [ms/i]

-------------------------------------------------------------------------------
scramble_fft + start_fft (optimized shufps and reverse_bits)

Testing FFT of size 256 (rank = 8)...
Time = 30.0 s, iterations = 41404000, performance = 1380124.1 [i/s], average time = 0.00072 [ms/i]
Testing FFT of size 512 (rank = 9)...
Time = 30.0 s, iterations = 19758000, performance = 658567.3 [i/s], average time = 0.00152 [ms/i]
Testing FFT of size 1024 (rank = 10)...
Time = 30.0 s, iterations = 8750000, performance = 291639.3 [i/s], average time = 0.00343 [ms/i]
Testing FFT of size 2048 (rank = 11)...
Time = 30.0 s, iterations = 3486000, performance = 116179.4 [i/s], average time = 0.00861 [ms/i]
Testing FFT of size 4096 (rank = 12)...
Time = 30.0 s, iterations = 1496000, performance = 49860.4 [i/s], average time = 0.02006 [ms/i]
Testing FFT of size 8192 (rank = 13)...
Time = 30.0 s, iterations = 762000, performance = 25375.2 [i/s], average time = 0.03941 [ms/i]
Testing FFT of size 16384 (rank = 14)...
Time = 30.1 s, iterations = 381000, performance = 12670.6 [i/s], average time = 0.07892 [ms/i]
Testing FFT of size 32768 (rank = 15)...
Time = 30.1 s, iterations = 155000, performance = 5145.1 [i/s], average time = 0.19436 [ms/i]
Testing FFT of size 65536 (rank = 16)...
Time = 30.0 s, iterations = 58000, performance = 1933.3 [i/s], average time = 0.51726 [ms/i]