4096 point FFT test result

• Raspberry Pi3: ARM Cortex A53, 1.2 GHz, 32 bit Raspbian 9 beta (gcc-6.2)
• Odroid-C2: ARM Cortex A53, 1.5 GHz, 64 bit Ubuntu 17.04 alpha (gcc-6.3)
• Intel i5-3337u: Ivy Bridge, 1.8 GHz (turbo 2,7 GHz), 64 bit Ubuntu 16.10 (gcc-6.2)

Architecture measurement: C-fast time in milliseconds at FFT_REPEAT=10 000 (gcc-6.3; 1 core test)

C-fast_...	double	float	integer
Raspberry Pi1 0.7 GHz ARM11	5.884 ms	4.712 ms	5.344 ms
BeagleBone Black 1 GHz ARM Cortex A8	4.009 ms	3.509 ms	1.759 ms
Odroid-C1 1.5 GHz ARM Cortex A5	1.472 ms	0.787 ms	1.268 ms
Raspberry Pi3 1.2 GHz ARM Cortex A53	0.915 ms	0.671 ms	1.053 ms
Odroid-C2@64 bit 1.5 GHz ARM Cortex A53	0.738 ms	0.496 ms	0.432 ms
Intel Celeron J1900	0.371 ms	0.342 ms	0.324 ms
Intel Core laptop i5-3337u	0.158 ms	0.107 ms	0.113 ms

C relative speed compared with C-fast_double; floatemu if not FPU in the processor (e.g. some 32 bit microcontroller)

Language	Raspberry Pi3	Odroid-C2@64 bit	i5-3337u
C-fast_double	1 (0.915 ms)	1 (0.726 ms)	1 (0.180 ms)
C-fast_float	0.82	0.74	0.69
C-fast_integer	1.12	0.58	0.79
C-fast_longlong_64bit	-	1.46	1.38
C-fast_float-nocomplex-floatemu	4.77	4.21	5.87
C++
C++-fast_double	1.03 (clang 6)	1.13	1.02

Other compiled language speed compared with C-fast_double

Language	Raspberry Pi3	Odroid-C2@64 bit	i5-3337u
C# (.NET 6)	1.26
Rust (rustc-1.14)		1.34	1.35
Go (gccgo-6.3)	1.45	1.30	1.49
C# (mono-5.18)	2.78
Java (Oracle)	7.94		1.38
C# (mono-4.x)	8.32	5.73	5.22
F# (fsharpc 4.0)	8.31		4.93

Script language speed compared with C-fast_double

Language	Raspberry Pi3	Odroid-C2@64 bit	i5-3337u
LUA (LUAjit)	8.01	8.49	5.4
Python (Pypy 5.4)	27.5		4.91
JavaScript (nodejs 4.x)	25	23.7	17.3
PHP7	129	135	88.2
Ruby	481	395	287
AWK	625	541	499
PERL	5139	5377	5710
Octave	9326		7972