cfft3b - compute a periodic sequence from its Fourier coefficients. The FFT operations are unnormalized, so a call of CFFT3F followed by a call of CFFT3B will multiply the input sequence by M*N*K.
SUBROUTINE CFFT3B(M, N, K, A, LDA, LD2A, WORK, LWORK) COMPLEX A(LDA,LD2A,*) INTEGER M, N, K, LDA, LD2A, LWORK REAL WORK(*) SUBROUTINE CFFT3B_64(M, N, K, A, LDA, LD2A, WORK, LWORK) COMPLEX A(LDA,LD2A,*) INTEGER*8 M, N, K, LDA, LD2A, LWORK REAL WORK(*) F95 INTERFACE SUBROUTINE FFT3B(M, N, K, A, LDA, LD2A, WORK, LWORK) COMPLEX, DIMENSION(:,:,:) :: A INTEGER :: M, N, K, LDA, LD2A, LWORK REAL, DIMENSION(:) :: WORK SUBROUTINE FFT3B_64(M, N, K, A, LDA, LD2A, WORK, LWORK) COMPLEX, DIMENSION(:,:,:) :: A INTEGER(8) :: M, N, K, LDA, LD2A, LWORK REAL, DIMENSION(:) :: WORK C INTERFACE #include <sunperf.h> void cfft3b(int m, int n, int k, complex *a, int lda, int ld2a, float *work, int lwork); void cfft3b_64(long m, long n, long k, complex *a, long lda, long ld2a, float *work, long lwork);
Oracle Solaris Studio Performance Library cfft3b(3P) NAME cfft3b - compute a periodic sequence from its Fourier coefficients. The FFT operations are unnormalized, so a call of CFFT3F followed by a call of CFFT3B will multiply the input sequence by M*N*K. SYNOPSIS SUBROUTINE CFFT3B(M, N, K, A, LDA, LD2A, WORK, LWORK) COMPLEX A(LDA,LD2A,*) INTEGER M, N, K, LDA, LD2A, LWORK REAL WORK(*) SUBROUTINE CFFT3B_64(M, N, K, A, LDA, LD2A, WORK, LWORK) COMPLEX A(LDA,LD2A,*) INTEGER*8 M, N, K, LDA, LD2A, LWORK REAL WORK(*) F95 INTERFACE SUBROUTINE FFT3B(M, N, K, A, LDA, LD2A, WORK, LWORK) COMPLEX, DIMENSION(:,:,:) :: A INTEGER :: M, N, K, LDA, LD2A, LWORK REAL, DIMENSION(:) :: WORK SUBROUTINE FFT3B_64(M, N, K, A, LDA, LD2A, WORK, LWORK) COMPLEX, DIMENSION(:,:,:) :: A INTEGER(8) :: M, N, K, LDA, LD2A, LWORK REAL, DIMENSION(:) :: WORK C INTERFACE #include <sunperf.h> void cfft3b(int m, int n, int k, complex *a, int lda, int ld2a, float *work, int lwork); void cfft3b_64(long m, long n, long k, complex *a, long lda, long ld2a, float *work, long lwork); ARGUMENTS M (input) Number of rows to be transformed. These subroutines are most efficient when M is a product of small primes. M >= 0. N (input) Number of columns to be transformed. These subroutines are most efficient when N is a product of small primes. N >= 0. K (input) Number of planes to be transformed. These subroutines are most efficient when K is a product of small primes. K >= 0. A (input/output) On entry, a three-dimensional array A(LDA,LD2A,K) that con- tains the sequences to be transformed. LDA (input) Leading dimension of the array containing the data to be transformed. LDA >= M. LD2A (input) Second dimension of the array containing the data to be transformed. LD2A >= N. WORK (input) On input, workspace WORK must have been initialized by CFFT3I. LWORK (input) The dimension of the array WORK. LWORK >= (4*(M + N + K) + 45). 7 Nov 2015 cfft3b(3P)