Collaboration diagram for Planning options:

Macros
#define	MUFFT_FLAG_CPU_ANY (0)
	muFFT will use any SIMD instruction set it can if supported by the CPU. More...

#define	MUFFT_FLAG_CPU_NO_SIMD ((1 << 16) - 1)
	muFFT will not use any SIMD instruction set. More...

#define	MUFFT_FLAG_CPU_NO_AVX (1 << 0)
	muFFT will not use the AVX instruction set. More...

#define	MUFFT_FLAG_CPU_NO_SSE3 (1 << 1)
	muFFT will not use the SSE3 instruction set. More...

#define	MUFFT_FLAG_CPU_NO_SSE (1 << 2)
	muFFT will not use the SSE instruction set. More...

#define	MUFFT_FLAG_FULL_R2C (1 << 16)
	The real-to-complex 1D transform will also output the redundant conjugate values X(N - k) = X(k)*. More...

#define	MUFFT_FLAG_ZERO_PAD_UPPER_HALF (1 << 17)

Detailed Description

Macro Definition Documentation

◆ MUFFT_FLAG_CPU_ANY

#define MUFFT_FLAG_CPU_ANY (0)

muFFT will use any SIMD instruction set it can if supported by the CPU.

◆ MUFFT_FLAG_CPU_NO_AVX

#define MUFFT_FLAG_CPU_NO_AVX (1 << 0)

muFFT will not use the AVX instruction set.

◆ MUFFT_FLAG_CPU_NO_SIMD

#define MUFFT_FLAG_CPU_NO_SIMD ((1 << 16) - 1)

muFFT will not use any SIMD instruction set.

◆ MUFFT_FLAG_CPU_NO_SSE

#define MUFFT_FLAG_CPU_NO_SSE (1 << 2)

muFFT will not use the SSE instruction set.

◆ MUFFT_FLAG_CPU_NO_SSE3

#define MUFFT_FLAG_CPU_NO_SSE3 (1 << 1)

muFFT will not use the SSE3 instruction set.

◆ MUFFT_FLAG_FULL_R2C

#define MUFFT_FLAG_FULL_R2C (1 << 16)

The real-to-complex 1D transform will also output the redundant conjugate values X(N - k) = X(k)*.

◆ MUFFT_FLAG_ZERO_PAD_UPPER_HALF

#define MUFFT_FLAG_ZERO_PAD_UPPER_HALF (1 << 17)

The second/upper half of the input array is assumed to be 0 and will not be read and memory for the second half of the input array does not have to be allocated. This is mostly useful when you want to do zero-padded FFTs which are very common for convolution-type operations, see 1D fast convolution. This flag is only recognized for 1D transforms.

Macros