Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: FFTW 3.3.8: FFTW MPI Transposes Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82:
Chris@82:

Chris@82: Next: , Previous: , Up: Distributed-memory FFTW with MPI   [Contents][Index]

Chris@82:
Chris@82:
Chris@82: Chris@82:

6.7 FFTW MPI Transposes

Chris@82: Chris@82: Chris@82:

The FFTW’s MPI Fourier transforms rely on one or more global Chris@82: transposition step for their communications. For example, the Chris@82: multidimensional transforms work by transforming along some Chris@82: dimensions, then transposing to make the first dimension local and Chris@82: transforming that, then transposing back. Because global Chris@82: transposition of a block-distributed matrix has many other potential Chris@82: uses besides FFTs, FFTW’s transpose routines can be called directly, Chris@82: as documented in this section. Chris@82:

Chris@82: Chris@82: Chris@82: Chris@82: Chris@82:
• Basic distributed-transpose interface:   Chris@82:
• Advanced distributed-transpose interface:   Chris@82:
• An improved replacement for MPI_Alltoall:   Chris@82:
Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: Chris@82: