148 lines
		
	
	
		
			4.7 KiB
		
	
	
	
		
			OCaml
		
	
	
	
	
	
		
		
			
		
	
	
			148 lines
		
	
	
		
			4.7 KiB
		
	
	
	
		
			OCaml
		
	
	
	
	
	
| 
								 | 
							
								(*
							 | 
						||
| 
								 | 
							
								 * Copyright (c) 1997-1999 Massachusetts Institute of Technology
							 | 
						||
| 
								 | 
							
								 * Copyright (c) 2003, 2007-14 Matteo Frigo
							 | 
						||
| 
								 | 
							
								 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
							 | 
						||
| 
								 | 
							
								 *
							 | 
						||
| 
								 | 
							
								 * This program is free software; you can redistribute it and/or modify
							 | 
						||
| 
								 | 
							
								 * it under the terms of the GNU General Public License as published by
							 | 
						||
| 
								 | 
							
								 * the Free Software Foundation; either version 2 of the License, or
							 | 
						||
| 
								 | 
							
								 * (at your option) any later version.
							 | 
						||
| 
								 | 
							
								 *
							 | 
						||
| 
								 | 
							
								 * This program is distributed in the hope that it will be useful,
							 | 
						||
| 
								 | 
							
								 * but WITHOUT ANY WARRANTY; without even the implied warranty of
							 | 
						||
| 
								 | 
							
								 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
							 | 
						||
| 
								 | 
							
								 * GNU General Public License for more details.
							 | 
						||
| 
								 | 
							
								 *
							 | 
						||
| 
								 | 
							
								 * You should have received a copy of the GNU General Public License
							 | 
						||
| 
								 | 
							
								 * along with this program; if not, write to the Free Software
							 | 
						||
| 
								 | 
							
								 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
							 | 
						||
| 
								 | 
							
								 *
							 | 
						||
| 
								 | 
							
								 *)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								(* abstraction layer for complex operations *)
							 | 
						||
| 
								 | 
							
								open Littlesimp
							 | 
						||
| 
								 | 
							
								open Expr
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								(* type of complex expressions *)
							 | 
						||
| 
								 | 
							
								type expr = CE of Expr.expr * Expr.expr
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let two = CE (makeNum Number.two, makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let one = CE (makeNum Number.one, makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let i = CE (makeNum Number.zero, makeNum Number.one)
							 | 
						||
| 
								 | 
							
								let zero = CE (makeNum Number.zero, makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let make (r, i) = CE (r, i)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let uminus (CE (a, b)) =  CE (makeUminus a, makeUminus b)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let inverse_int n = CE (makeNum (Number.div Number.one (Number.of_int n)),
							 | 
						||
| 
								 | 
							
											makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let inverse_int_sqrt n = 
							 | 
						||
| 
								 | 
							
								  CE (makeNum (Number.div Number.one (Number.sqrt (Number.of_int n))),
							 | 
						||
| 
								 | 
							
								      makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let int_sqrt n = 
							 | 
						||
| 
								 | 
							
								  CE (makeNum (Number.sqrt (Number.of_int n)),
							 | 
						||
| 
								 | 
							
								      makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let nan x = CE (NaN x, makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let half = inverse_int 2
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let times3x3 (CE (a, b)) (CE (c, d)) = 
							 | 
						||
| 
								 | 
							
								  CE (makePlus [makeTimes (c, makePlus [a; makeUminus (b)]);
							 | 
						||
| 
								 | 
							
									        makeTimes (b, makePlus [c; makeUminus (d)])],
							 | 
						||
| 
								 | 
							
								      makePlus [makeTimes (a, makePlus [c; d]);
							 | 
						||
| 
								 | 
							
									        makeUminus(makeTimes (c, makePlus [a; makeUminus (b)]))])
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let times (CE (a, b)) (CE (c, d)) = 
							 | 
						||
| 
								 | 
							
								  if not !Magic.threemult then
							 | 
						||
| 
								 | 
							
								    CE (makePlus [makeTimes (a, c); makeUminus (makeTimes (b, d))],
							 | 
						||
| 
								 | 
							
								        makePlus [makeTimes (a, d); makeTimes (b, c)])
							 | 
						||
| 
								 | 
							
								  else if is_constant c && is_constant d then
							 | 
						||
| 
								 | 
							
								    times3x3 (CE (a, b)) (CE (c, d))
							 | 
						||
| 
								 | 
							
								  else (* hope a and b are constant expressions *)
							 | 
						||
| 
								 | 
							
								    times3x3 (CE (c, d)) (CE (a, b))
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let ctimes (CE (a, _)) (CE (c, _)) = 
							 | 
						||
| 
								 | 
							
								  CE (CTimes (a, c), makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let ctimesj (CE (a, _)) (CE (c, _)) = 
							 | 
						||
| 
								 | 
							
								  CE (CTimesJ (a, c), makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								      
							 | 
						||
| 
								 | 
							
								(* complex exponential (of root of unity); returns exp(2*pi*i/n * m) *)
							 | 
						||
| 
								 | 
							
								let exp n i =
							 | 
						||
| 
								 | 
							
								  let (c, s) = Number.cexp n i
							 | 
						||
| 
								 | 
							
								  in CE (makeNum c, makeNum s)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								(* various trig functions evaluated at (2*pi*i/n * m) *)
							 | 
						||
| 
								 | 
							
								let sec n m =
							 | 
						||
| 
								 | 
							
								  let (c, s) = Number.cexp n m
							 | 
						||
| 
								 | 
							
								  in CE (makeNum (Number.div Number.one c), makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let csc n m =
							 | 
						||
| 
								 | 
							
								  let (c, s) = Number.cexp n m
							 | 
						||
| 
								 | 
							
								  in CE (makeNum (Number.div Number.one s), makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let tan n m =
							 | 
						||
| 
								 | 
							
								  let (c, s) = Number.cexp n m
							 | 
						||
| 
								 | 
							
								  in CE (makeNum (Number.div s c), makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let cot n m =
							 | 
						||
| 
								 | 
							
								  let (c, s) = Number.cexp n m
							 | 
						||
| 
								 | 
							
								  in CE (makeNum (Number.div c s), makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								    
							 | 
						||
| 
								 | 
							
								(* complex sum *)
							 | 
						||
| 
								 | 
							
								let plus a =
							 | 
						||
| 
								 | 
							
								  let rec unzip_complex = function
							 | 
						||
| 
								 | 
							
								      [] -> ([], [])
							 | 
						||
| 
								 | 
							
								    | ((CE (a, b)) :: s) ->
							 | 
						||
| 
								 | 
							
								        let (r,i) = unzip_complex s
							 | 
						||
| 
								 | 
							
									in
							 | 
						||
| 
								 | 
							
									(a::r), (b::i) in
							 | 
						||
| 
								 | 
							
								  let (c, d) = unzip_complex a in
							 | 
						||
| 
								 | 
							
								  CE (makePlus c, makePlus d)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								(* extract real/imaginary *)
							 | 
						||
| 
								 | 
							
								let real (CE (a, b)) = CE (a, makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let imag (CE (a, b)) = CE (b, makeNum Number.zero)
							 | 
						||
| 
								 | 
							
								let iimag (CE (a, b)) = CE (makeNum Number.zero, b)
							 | 
						||
| 
								 | 
							
								let conj (CE (a, b)) = CE (a, makeUminus b)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								    
							 | 
						||
| 
								 | 
							
								(* abstraction of sum_{i=0}^{n-1} *)
							 | 
						||
| 
								 | 
							
								let sigma a b f = plus (List.map f (Util.interval a b))
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								(* store and assignment operations *)
							 | 
						||
| 
								 | 
							
								let store_real v (CE (a, b)) = Expr.Store (v, a)
							 | 
						||
| 
								 | 
							
								let store_imag v (CE (a, b)) = Expr.Store (v, b)
							 | 
						||
| 
								 | 
							
								let store (vr, vi) x = (store_real vr x, store_imag vi x)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let assign_real v (CE (a, b)) = Expr.Assign (v, a)
							 | 
						||
| 
								 | 
							
								let assign_imag v (CE (a, b)) = Expr.Assign (v, b)
							 | 
						||
| 
								 | 
							
								let assign (vr, vi) x = (assign_real vr x, assign_imag vi x)
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								(************************
							 | 
						||
| 
								 | 
							
								   shortcuts 
							 | 
						||
| 
								 | 
							
								 ************************)
							 | 
						||
| 
								 | 
							
								let (@*) = times
							 | 
						||
| 
								 | 
							
								let (@+) a b = plus [a; b]
							 | 
						||
| 
								 | 
							
								let (@-) a b = plus [a; uminus b]
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								(* type of complex signals *)
							 | 
						||
| 
								 | 
							
								type signal = int -> expr
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								(* make a finite signal infinite *)
							 | 
						||
| 
								 | 
							
								let infinite n signal i = if ((0 <= i) && (i < n)) then signal i else zero
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let hermitian n a =
							 | 
						||
| 
								 | 
							
								  Util.array n (fun i ->
							 | 
						||
| 
								 | 
							
								    if (i = 0) then real (a 0)
							 | 
						||
| 
								 | 
							
								    else if (i < n - i)  then (a i)
							 | 
						||
| 
								 | 
							
								    else if (i > n - i)  then conj (a (n - i))
							 | 
						||
| 
								 | 
							
								    else real (a i))
							 | 
						||
| 
								 | 
							
								
							 | 
						||
| 
								 | 
							
								let antihermitian n a =
							 | 
						||
| 
								 | 
							
								  Util.array n (fun i ->
							 | 
						||
| 
								 | 
							
								    if (i = 0) then iimag (a 0)
							 | 
						||
| 
								 | 
							
								    else if (i < n - i)  then (a i)
							 | 
						||
| 
								 | 
							
								    else if (i > n - i)  then uminus (conj (a (n - i)))
							 | 
						||
| 
								 | 
							
								    else iimag (a i))
							 |