| Server IP : 68.183.124.220 / Your IP : 216.73.217.137 Web Server : Apache/2.4.18 (Ubuntu) System : Linux Sandbox-A 4.4.0-210-generic #242-Ubuntu SMP Fri Apr 16 09:57:56 UTC 2021 x86_64 User : gavin ( 1000) PHP Version : 7.0.33-0ubuntu0.16.04.16 Disable Function : pcntl_alarm,pcntl_fork,pcntl_waitpid,pcntl_wait,pcntl_wifexited,pcntl_wifstopped,pcntl_wifsignaled,pcntl_wifcontinued,pcntl_wexitstatus,pcntl_wtermsig,pcntl_wstopsig,pcntl_signal,pcntl_signal_dispatch,pcntl_get_last_error,pcntl_strerror,pcntl_sigprocmask,pcntl_sigwaitinfo,pcntl_sigtimedwait,pcntl_exec,pcntl_getpriority,pcntl_setpriority, MySQL : OFF | cURL : ON | WGET : ON | Perl : ON | Python : ON | Sudo : ON | Pkexec : ON Directory : /usr/src/Python-3.6.8/Modules/_decimal/libmpdec/literature/ |
Upload File : |
(* Copyright (c) 2011 Stefan Krah. All rights reserved. *)
The Six Step Transform:
=======================
In libmpdec, the six-step transform is the Matrix Fourier Transform (See
matrix-transform.txt) in disguise. It is called six-step transform after
a variant that appears in [1]. The algorithm requires that the input
array can be viewed as an R*C matrix.
Algorithm six-step (forward transform):
---------------------------------------
1a) Transpose the matrix.
1b) Apply a length R FNT to each row.
1c) Transpose the matrix.
2) Multiply each matrix element (addressed by j*C+m) by r**(j*m).
3) Apply a length C FNT to each row.
4) Transpose the matrix.
Note that steps 1a) - 1c) are exactly equivalent to step 1) of the Matrix
Fourier Transform. For large R, it is faster to transpose twice and do
a transform on the rows than to perform a column transpose directly.
Algorithm six-step (inverse transform):
---------------------------------------
0) View the matrix as a C*R matrix.
1) Transpose the matrix, producing an R*C matrix.
2) Apply a length C FNT to each row.
3) Multiply each matrix element (addressed by i*C+n) by r**(i*n).
4a) Transpose the matrix.
4b) Apply a length R FNT to each row.
4c) Transpose the matrix.
Again, steps 4a) - 4c) are equivalent to step 4) of the Matrix Fourier
Transform.
--
[1] David H. Bailey: FFTs in External or Hierarchical Memory
http://crd.lbl.gov/~dhbailey/dhbpapers/