4-point:
    338 decicycles in           av_tx (fft),134211225 runs,   6503 skips
    701 decicycles in   av_fft_permute+calc,134203127 runs,  14601 skips    52% slower
    503 decicycles in         fftwf_execute,134207977 runs,   9751 skips    33% slower

8-point:
    648 decicycles in           av_tx (fft),134210263 runs,   7465 skips
    841 decicycles in   av_fft_permute+calc,134205381 runs,  12347 skips    23% slower
    693 decicycles in         fftwf_execute,134205401 runs,  12327 skips     6% slower

16-point:
    826 decicycles in           av_tx (fft),134206886 runs,  10842 skips
   1102 decicycles in   av_fft_permute+calc,134200530 runs,  17198 skips    25% slower
   1179 decicycles in         fftwf_execute,134202284 runs,  15444 skips    30% slower

32-point:
   1079 decicycles in           av_tx (fft),134206314 runs,  11414 skips
   1650 decicycles in   av_fft_permute+calc,134200531 runs,  17197 skips    35% slower
   1290 decicycles in         fftwf_execute,134201876 runs,  15852 skips    16% slower

64-point:
   2147 decicycles in           av_tx (fft),67092009 runs,  16855 skips
   2818 decicycles in   av_fft_permute+calc,67086191 runs,  22673 skips     24% slower
   1922 decicycles in         fftwf_execute,67093709 runs,  15155 skips     11% faster

128-point:
   4166 decicycles in           av_tx (fft),67081298 runs,  27566 skips
   5441 decicycles in   av_fft_permute+calc,67064991 runs,  43873 skips     23% slower
   3196 decicycles in         fftwf_execute,67080914 runs,  27950 skips     30% faster

256-point:
   8576 decicycles in           av_tx (fft),33509344 runs,  45088 skips
  11331 decicycles in   av_fft_permute+calc,33504672 runs,  49760 skips     24% slower
   6292 decicycles in         fftwf_execute,33519321 runs,  35111 skips     36% faster

512-point:
  18061 decicycles in           av_tx (fft),16734541 runs,  42675 skips
  23105 decicycles in   av_fft_permute+calc,16727657 runs,  49559 skips     22% slower
  12523 decicycles in         fftwf_execute,16743874 runs,  33342 skips     44% faster

1024-point:
  39627 decicycles in           av_tx (fft),16705169 runs,  72047 skips
  49626 decicycles in   av_fft_permute+calc,16704245 runs,  72971 skips     20% slower
  26777 decicycles in         fftwf_execute,16726971 runs,  50245 skips     48% faster

2048-point:
  94013 decicycles in           av_tx (fft),16777120 runs,     96 skips
 131233 decicycles in   av_fft_permute+calc,16777102 runs,    114 skips     29% slower
  73309 decicycles in         fftwf_execute,16771817 runs,   5399 skips     28% faster

4096-point:
 230427 decicycles in           av_tx (fft), 8388537 runs,     71 skips
 442121 decicycles in   av_fft_permute+calc, 8388468 runs,    140 skips     48% slower
 172721 decicycles in         fftwf_execute, 8388531 runs,     77 skips     33% faster

8192-point:
 540304 decicycles in           av_tx (fft), 2097111 runs,     41 skips     
 881453 decicycles in   av_fft_permute+calc, 2097082 runs,     70 skips     39% slower
 468983 decicycles in         fftwf_execute, 2097115 runs,     37 skips     15% faster

16384-point:
1267442 decicycles in           av_tx (fft), 1048541 runs,     35 skips
1951607 decicycles in   av_fft_permute+calc, 1048576 runs,      0 skips     35% slower
1122339 decicycles in         fftwf_execute, 1048535 runs,     41 skips     12% faster

32768-point:
3005730 decicycles in           av_tx (fft),  262144 runs,      0 skips
4678102 decicycles in   av_fft_permute+calc,  262144 runs,      0 skips     35% slower
2728375 decicycles in         fftwf_execute,  262144 runs,      0 skips     10% faster

65536-point;
6848365 decicycles in           av_tx (fft),  131072 runs,      0 skips
10018454 decicycles in   av_fft_permute+calc,  131072 runs,      0 skips    32% slower
7197566 decicycles in         fftwf_execute,  131072 runs,      0 skips      5% slower!!!

131072-point:
14336848 decicycles in           av_tx (fft),  131072 runs,      0 skips
52757378 decicycles in   av_fft_permute+calc,  131072 runs,      0 skips    73% slower!
15795066 decicycles in         fftwf_execute,  131072 runs,      0 skips    10% SLOWER!!!