URL
https://opencores.org/ocsvn/yac/yac/trunk
Subversion Repositories yac
[/] [yac/] [trunk/] [c_octave/] [cordic_iterative_test.m] - Rev 12
Go to most recent revision | Compare with Previous | Blame | View Log
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%% %%%% %%%% File : cordic_iterative_test.m %%%% %%%% Project : YAC (Yet Another CORDIC Core) %%%% %%%% Creation : Feb. 2014 %%%% %%%% Limitations : %%%% %%%% Platform : Linux, Mac, Windows %%%% %%%% Target : Octave, Matlab %%%% %%%% %%%% %%%% Author(s): : Christian Haettich %%%% %%%% Email : feddischson@opencores.org %%%% %%%% %%%% %%%% %%%% %%%%% %%%%% %%%% %%%% %%%% Description %%%% %%%% Script to test/analyze the cordic C implementation %%%% %%%% and to generate stimulus data for RTL simulation. %%%% %%%% This created data is used to ensure, that the C %%%% %%%% implementation behaves the same than the VHDL %%%% %%%% implementation. %%%% %%%% %%%% <<<<<<< HEAD <<<<<<< HEAD ======= >>>>>>> Updated C and RTL model as well as the documentation %%%% Three tests are implemented: %%%% %%%% - Random test values %%%% %%%% - Linear increasing values %%%% %%%% - Limit values %%%% %%%% %%%% %%%% %%%% %%%% Please do 'mex cordic_iterative.c' to create %%%% %%%% the cordic_iterative.mex. %%%% <<<<<<< HEAD ======= >>>>>>> initial commit ======= >>>>>>> Updated C and RTL model as well as the documentation %%%% %%%% %%%%% %%%%% %%%% %%%% %%%% TODO %%%% <<<<<<< HEAD <<<<<<< HEAD %%%% The linear test is not complete %%%% ======= %%%% Some documentation and function description %%%% >>>>>>> initial commit ======= %%%% The linear test is not complete %%%% >>>>>>> Updated C and RTL model as well as the documentation %%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%% %%%% %%%% Copyright Notice %%%% %%%% %%%% %%%% This file is part of YAC - Yet Another CORDIC Core %%%% %%%% Copyright (c) 2014, Author(s), All rights reserved. %%%% %%%% %%%% %%%% YAC is free software; you can redistribute it and/or %%%% %%%% modify it under the terms of the GNU Lesser General Public %%%% %%%% License as published by the Free Software Foundation; either %%%% %%%% version 3.0 of the License, or (at your option) any later version. %%%% %%%% %%%% %%%% YAC 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 %%%% %%%% Lesser General Public License for more details. %%%% %%%% %%%% %%%% You should have received a copy of the GNU Lesser General Public %%%% %%%% License along with this library. If not, download it from %%%% %%%% http://www.gnu.org/licenses/lgpl %%%% %%%% %%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <<<<<<< HEAD <<<<<<< HEAD function cordic_iterative_test( ) ======= ======= >>>>>>> Updated C and RTL model as well as the documentation function cordic_iterative_test( ) >>>>>>> initial commit % global flags/values, they are static % through the whole script and defined below global C_FLAG_VEC_ROT C_FLAG_ATAN_3 C_MODE_CIRC C_MODE_LIN C_MODE_HYP global XY_WIDTH ANGLEWIDTH GUARDBITS RM_GAIN cordic_iterative_setup % open output file tb_fid = fopen( TB_FILE, 'w' ); % Number of tests, which are run N_TESTS = 10000; <<<<<<< HEAD % open test file tb_fid = fopen( './tb_data.txt', 'w' ); <<<<<<< HEAD <<<<<<< HEAD %tb_fid = 0; % % run test, which uses random values run_random_test( N_TESTS, tb_fid ); % % run tests, which test limits run_limit_test( tb_fid ); % % run linear value test run_linear_test( 1000, tb_fid ); % close file if tb_fid > 0 fclose( tb_fid ); end end function run_limit_test( tb_fid ) %RUN_LIMIT_TEST Test the range limit % % run_limit_test( fid ) % % This function is used to generate a test pattern % with values, which are at the range limit. % This values are then processed by the fixed-point YAC % implementation. All input and outputs are logged into % a testbench pattern file. % % The argument fid is the file-descriptor of the testbench pattern % file. % data_a = [ 0 1 0 1 -1 0 -1 1 -1 ]; data_b = [ 0 0 1 1 0 -1 -1 -1 1 ]; data_c = [ 0 0 0 0 0 0 0 0 0 ... 1 1 1 1 1 1 1 1 1 ... -1 -1 -1 -1 -1 -1 -1 -1 -1 ]; data_d = data_a * pi; data_a_div = [ 0.5 ,1 -0.5, -1, -0.5, -1 ]; data_b_div = [ 1 ,1, 1, 1, -1, -1 ]; [ ~, ~, atan_err, abs_err, it_1 ] = ccart2pol( data_a, data_b, tb_fid ); [ ~, ~, sin_err, cos_err, it_2 ] = cpol2cart( data_d, data_b, tb_fid ); [ ~, ~, x_err, y_err, it_3 ] = crot( [ data_a, data_a, data_a], ... [ data_b, data_b, data_b], ... data_c, tb_fid ); [ ~, div_err, it_4 ] = cdiv( data_a_div, data_b_div, tb_fid ); [ ~, mul_err, it_5 ] = cmul( data_a, data_b, tb_fid ); print_result_info( ... atan_err, it_1, ... abs_err, it_1, ... sin_err, it_2, ... cos_err, it_2, ... x_err, it_3, ... y_err, it_3, ... div_err, it_4, ... mul_err, it_5, ... 0, 0, ... 0, 0, ... 0, 0, ... 0, 0, ... 'Limit Value Test' ); ======= ======= %tb_fid = 0; >>>>>>> Updated C and RTL model as well as the documentation ======= >>>>>>> Removed some bugs regarding pre-rotation and negative numbers in the wb wrapper % % run test, which uses random values run_random_test( N_TESTS, tb_fid ); % % run tests, which test limits run_limit_test( tb_fid ); % % run linear value test run_linear_test( 1000, tb_fid ); % close file if tb_fid > 0 fclose( tb_fid ); end end function run_limit_test( tb_fid ) %RUN_LIMIT_TEST Test the range limit % % run_limit_test( fid ) % % This function is used to generate a test pattern % with values, which are at the range limit. % This values are then processed by the fixed-point YAC % implementation. All input and outputs are logged into % a testbench pattern file. % % The argument fid is the file-descriptor of the testbench pattern % file. % <<<<<<< HEAD >>>>>>> initial commit end <<<<<<< HEAD function run_linear_test( N_TESTS, tb_fid ) %RUN_LINEAR_TEST Generates a linear test pattern % % run_linear_test( N, fid ) % % This function is used to generate linear increasing test % values. % These values are then processed by the fixed-point YAC % implementation. All input and outputs are logged into % a testbench pattern file. In addition, the result is plotted. % % NOTE: only the hyperbolic functions are processed at the moment. % This function needs to be extended in future. % % % The argument fid is the file-descriptor of the testbench pattern % file. The argument N defines the number of values, which are processed. % % data_a_h = ones( 1, N_TESTS ); data_b_h = linspace( -1, 1, N_TESTS ) * 0.78; data_c_h = linspace( -1, 1, N_TESTS ); [ atanh_res, sqrt_res, atanh_err, sqrt_err, it_6 ] = catanh( data_a_h, data_b_h, tb_fid ); [ sinh_res, cosh_res, sinh_err, cosh_err, it_7 ] = csinhcosh( data_c_h, tb_fid ); ======= data_a = [ 0 1 0 1 -1 0 -1 1 -1 ]; data_b = [ 0 0 1 1 0 -1 -1 -1 1 ]; data_c = [ 0 0 0 0 0 0 0 0 0 ... 1 1 1 1 1 1 1 1 1 ... -1 -1 -1 -1 -1 -1 -1 -1 -1 ]; data_d = data_a * pi; data_a_div = [ 0.5 ,1 -0.5, -1, -0.5, -1 ]; data_b_div = [ 1 ,1, 1, 1, -1, -1 ]; [ ~, ~, atan_err, abs_err, it_1 ] = ccart2pol( data_a, data_b, tb_fid ); [ ~, ~, sin_err, cos_err, it_2 ] = cpol2cart( data_d, data_b, tb_fid ); [ ~, ~, x_err, y_err, it_3 ] = crot( [ data_a, data_a, data_a], ... [ data_b, data_b, data_b], ... data_c, tb_fid ); [ ~, div_err, it_4 ] = cdiv( data_a_div, data_b_div, tb_fid ); [ ~, mul_err, it_5 ] = cmul( data_a, data_b, tb_fid ); print_result( ... atan_err, it_1, ... abs_err, it_1, ... sin_err, it_2, ... cos_err, it_2, ... x_err, it_3, ... y_err, it_3, ... div_err, it_4, ... mul_err, it_5, ... 0, 0, ... 0, 0, ... 0, 0, ... 0, 0, ... 'Limit Value Test' ); >>>>>>> Updated C and RTL model as well as the documentation figure; plot( data_b_h, atanh_res ); title( 'atanh' ); figure; plot( data_b_h, atanh_err ); title( 'atanh-error' ); figure; plot( data_c_h, sinh_res, data_c_h, cosh_res ); title( 'sinh and cosh' ); figure; plot( data_c_h, sinh_err, data_c_h, cosh_err ); title( 'sinh and cosh errors' ); end <<<<<<< HEAD function run_random_test( N_TESTS, tb_fid ) %RUN_RANDOM_TEST Generates a random test pattern % % run_random_test( N, fid ) % % This function is used to generate random test % values (uniform distributed). % These values are then processed by the fixed-point YAC % implementation. All input and outputs are logged into % a testbench pattern file. ======= function run_linear_test( N_TESTS, tb_fid ) %RUN_LINEAR_TEST Generates a linear test pattern % % run_linear_test( N, fid ) % % This function is used to generate linear increasing test % values. % These values are then processed by the fixed-point YAC % implementation. All input and outputs are logged into % a testbench pattern file. In addition, the result is plotted. % % NOTE: only the hyperbolic functions are processed at the moment. % This function needs to be extended in future. >>>>>>> Updated C and RTL model as well as the documentation % % % The argument fid is the file-descriptor of the testbench pattern % file. The argument N defines the number of values, which are processed. % % <<<<<<< HEAD data_a = -1 + 2 .* rand( 1, N_TESTS ); data_b = -1 + 2 .* rand( 1, N_TESTS ); data_c = -1 + 2 .* rand( 1, N_TESTS ); data_d = -pi + 2*pi .* rand( 1, N_TESTS ); ======= function run_random_test( N_TESTS, tb_fid ) ======= data_a_h = ones( 1, N_TESTS ); data_b_h = linspace( -1, 1, N_TESTS ) * 0.78; data_c_h = linspace( -1, 1, N_TESTS ); [ atanh_res, sqrt_res, atanh_err, sqrt_err, it_6 ] = catanh( data_a_h, data_b_h, tb_fid ); [ sinh_res, cosh_res, sinh_err, cosh_err, it_7 ] = csinhcosh( data_c_h, tb_fid ); figure; plot( data_b_h, atanh_res ); title( 'atanh' ); figure; plot( data_b_h, atanh_err ); title( 'atanh-error' ); figure; plot( data_c_h, sinh_res, data_c_h, cosh_res ); title( 'sinh and cosh' ); figure; plot( data_c_h, sinh_err, data_c_h, cosh_err ); title( 'sinh and cosh errors' ); end >>>>>>> Updated C and RTL model as well as the documentation function run_random_test( N_TESTS, tb_fid ) %RUN_RANDOM_TEST Generates a random test pattern % % run_random_test( N, fid ) % % This function is used to generate random test % values (uniform distributed). % These values are then processed by the fixed-point YAC % implementation. All input and outputs are logged into % a testbench pattern file. % % % The argument fid is the file-descriptor of the testbench pattern % file. The argument N defines the number of values, which are processed. % % data_a = -1 + 2 .* rand( 1, N_TESTS ); data_b = -1 + 2 .* rand( 1, N_TESTS ); <<<<<<< HEAD >>>>>>> initial commit ======= data_c = -1 + 2 .* rand( 1, N_TESTS ); data_d = -pi + 2*pi .* rand( 1, N_TESTS ); >>>>>>> Updated C and RTL model as well as the documentation % adapat data for division data_a_div = data_a; data_b_div = data_b; swap_div = ( data_b ./ data_a ) >= 2 | ( data_b ./ data_a ) < -2 ; data_a_div( swap_div ) = data_b( swap_div ); data_b_div( swap_div ) = data_a( swap_div ); data_a_h = ones( size( data_a ) ); <<<<<<< HEAD <<<<<<< HEAD data_b_h = data_b .* 0.80694; %0.78; ======= data_b_h = data_b .* 0.78; >>>>>>> initial commit ======= data_b_h = data_b .* 0.80694; %0.78; >>>>>>> Updated C and RTL model as well as the documentation [ ~, ~, atan_err, abs_err, it_1 ] = ccart2pol( data_a, data_b, tb_fid ); <<<<<<< HEAD <<<<<<< HEAD ======= >>>>>>> Updated C and RTL model as well as the documentation [ ~, ~, sin_err, cos_err, it_2 ] = cpol2cart( data_d, data_b, tb_fid ); [ ~, ~, x_err, y_err, it_3 ] = crot( data_a, data_b, data_c, tb_fid ); [ ~, div_err, it_4 ] = cdiv( data_a_div, data_b_div, tb_fid ); [ ~, mul_err, it_5 ] = cmul( data_a, data_b, tb_fid ); [ ~, ~, atanh_err, sqrt_err, it_6 ] = catanh( data_a_h, data_b_h, tb_fid ); [ ~, ~, sinh_err, cosh_err, it_7 ] = csinhcosh( data_a, tb_fid ); print_result_header( 'Random Test Result' ); print_result2( 'atan', it_1, atan_err, data_a, data_b ); print_result2( 'abs ', it_1, abs_err, data_a, data_b ); print_result2( 'sin ', it_2, sin_err, data_a, data_b ); print_result2( 'cos ', it_2, cos_err, data_a, data_b ); print_result( atan_err, it_1, ... abs_err, it_1, ... sin_err, it_2, ... cos_err, it_2, ... x_err, it_3, ... y_err, it_3, ... div_err, it_4, ... mul_err, it_5, ... atanh_err, it_6, ... sqrt_err, it_6, ... sinh_err, it_7, ... cosh_err, it_7, ... 'Random Value Test' ); end function print_result_header( title ) printf( ' ___________________________________________________________________\n' ); fprintf( ' %s\n', title); fprintf( ' -----+-------------------+--------------------+-------------------\n' ); end function print_result2( msg, it, err, a, b ) [ verr, ierr ] = max( err ); fprintf( '%s | %.14f | (%.14f %.14f) | %.14f | %.5f \n', msg, verr, a( ierr ), b( ierr ), mean( err ), max( it ) ) end function print_result( ... atan_err, atan_it, ... abs_err, abs_it, ... sin_err, sin_it, ... cos_err, cos_it, ... x_err, x_it, ... y_err, y_it, ... div_err, div_it, ... mul_err, mul_it, ... atanh_err, atanh_it, ... sqrt_err, sqrt_it, ... sinh_err, sinh_it, ... cosh_err, cosh_it, ... title ) <<<<<<< HEAD fprintf( ' ___________________________________________________________________\n' ); fprintf( ' %s\n', title); fprintf( ' -----+-------------------+--------------------+-------------------\n' ); fprintf( ' | max error | mean error | max iterations \n' ); fprintf( ' atan | % .14f | % .14f | %.5f \n', max( atan_err ), mean( atan_err ), max( atan_it ) ); fprintf( ' abs | % .14f | % .14f | %.5f \n', max( abs_err ), mean( abs_err ), max( abs_it ) ); fprintf( ' sin | % .14f | % .14f | %.5f \n', max( sin_err ), mean( sin_err ), max( sin_it ) ); fprintf( ' cos | % .14f | % .14f | %.5f \n', max( cos_err ), mean( cos_err ), max( cos_it ) ); fprintf( ' x | % .14f | % .14f | %.5f \n', max( x_err ), mean( x_err ), max( x_it ) ); fprintf( ' y | % .14f | % .14f | %.5f \n', max( y_err ), mean( y_err ), max( y_it ) ); fprintf( ' div | % .14f | % .14f | %.5f \n', max( div_err ), mean( div_err ), max( div_it ) ); fprintf( ' mul | % .14f | % .14f | %.5f \n', max( mul_err ), mean( mul_err ), max( mul_it ) ); fprintf( ' atanh| % .14f | % .14f | %.5f \n', max( atanh_err ), mean( atanh_err ), max( atanh_it ) ); fprintf( ' sqrt | % .14f | % .14f | %.5f \n', max( sqrt_err ), mean( sqrt_err ), max( sqrt_it ) ); fprintf( ' sinh | % .14f | % .14f | %.5f \n', max( sinh_err ), mean( sinh_err ), max( sinh_it ) ); fprintf( ' cosh | % .14f | % .14f | %.5f \n', max( cosh_err ), mean( cosh_err ), max( cosh_it ) ); ======= [ ~, ~, sin_err, cos_err, it_2 ] = cpol2cart( data_a, data_b, tb_fid ); [ ~, div_err, it_3 ] = cdiv( data_a_div, data_b_div, tb_fid ); [ ~, mul_err, it_4 ] = cmul( data_a, data_b, tb_fid ); [ ~, ~, atanh_err, sqrt_err, it_5 ] = catanh( data_a_h, data_b_h, tb_fid ); [ ~, ~, sinh_err, cosh_err, it_6 ] = csinhcosh( data_a, tb_fid ); ======= >>>>>>> Updated C and RTL model as well as the documentation fprintf( ' ___________________________________________________________________\n' ); fprintf( ' %s\n', title); fprintf( ' -----+-------------------+--------------------+-------------------\n' ); fprintf( ' | max error | mean error | max iterations \n' ); fprintf( ' atan | % .14f | % .14f | %.5f \n', max( atan_err ), mean( atan_err ), max( atan_it ) ); fprintf( ' abs | % .14f | % .14f | %.5f \n', max( abs_err ), mean( abs_err ), max( abs_it ) ); fprintf( ' sin | % .14f | % .14f | %.5f \n', max( sin_err ), mean( sin_err ), max( sin_it ) ); fprintf( ' cos | % .14f | % .14f | %.5f \n', max( cos_err ), mean( cos_err ), max( cos_it ) ); fprintf( ' x | % .14f | % .14f | %.5f \n', max( x_err ), mean( x_err ), max( x_it ) ); fprintf( ' y | % .14f | % .14f | %.5f \n', max( y_err ), mean( y_err ), max( y_it ) ); fprintf( ' div | % .14f | % .14f | %.5f \n', max( div_err ), mean( div_err ), max( div_it ) ); fprintf( ' mul | % .14f | % .14f | %.5f \n', max( mul_err ), mean( mul_err ), max( mul_it ) ); fprintf( ' atanh| % .14f | % .14f | %.5f \n', max( atanh_err ), mean( atanh_err ), max( atanh_it ) ); fprintf( ' sqrt | % .14f | % .14f | %.5f \n', max( sqrt_err ), mean( sqrt_err ), max( sqrt_it ) ); fprintf( ' sinh | % .14f | % .14f | %.5f \n', max( sinh_err ), mean( sinh_err ), max( sinh_it ) ); fprintf( ' cosh | % .14f | % .14f | %.5f \n', max( cosh_err ), mean( cosh_err ), max( cosh_it ) ); end >>>>>>> initial commit end function [sinh_res, cosh_res, sinh_err, cosh_err, it ]= csinhcosh( th, fid ) global C_FLAG_VEC_ROT C_FLAG_ATAN_3 C_MODE_CIRC C_MODE_LIN C_MODE_HYP global XY_WIDTH ANGLEWIDTH GUARDBITS RM_GAIN xi = repmat( (2^(XY_WIDTH-1)-1), size( th ) ); yi = zeros( 1, length( th ) ); ai = round( th .* (2^(ANGLEWIDTH-1)-1) ); mode = C_MODE_HYP; % cordic version [ rcosh rsinh ra, it ] = cordic_iterative( ... xi, ... yi, ... ai, ... mode, ... XY_WIDTH, ... ANGLEWIDTH, ... GUARDBITS, ... RM_GAIN ); cosh_res = rcosh ./ ( 2^(XY_WIDTH-1)-1 ); sinh_res = rsinh ./ ( 2^(XY_WIDTH-1)-1 ); cosh_m = cosh( th ); sinh_m = sinh( th ); sinh_err = abs(sinh_res - sinh_m ); cosh_err = abs(cosh_res - cosh_m ); % write TB data write_tb( fid, xi, yi, ai, rcosh, rsinh, ra, mode ); end function [atan_res, abs_res, atan_err, abs_err, it ] = catanh( x, y, fid ) global C_FLAG_VEC_ROT C_FLAG_ATAN_3 C_MODE_CIRC C_MODE_LIN C_MODE_HYP global XY_WIDTH ANGLEWIDTH GUARDBITS RM_GAIN if( size( x ) ~= size( y ) ) error( 'size error' ) end ai = zeros( size( x ) ); xi = round( x * (2^(XY_WIDTH-1)-1) ); yi = round( y * (2^(XY_WIDTH-1)-1) ); mode = C_FLAG_VEC_ROT + C_MODE_HYP; % cordic version [ rx, ry, ra, it ] = cordic_iterative( xi, ... yi, ... ai, ... mode, ... XY_WIDTH, ... ANGLEWIDTH, ... GUARDBITS, ... RM_GAIN ); % matlab version m_th = atanh( y ./ x ); m_r = sqrt( x.^2 - y.^2 ); % comparison atan_res = ra ./ 2^( (ANGLEWIDTH)-1); abs_res = rx ./ ( 2^(XY_WIDTH-1) -1 ); atan_err = abs( m_th - atan_res ); abs_err = abs( m_r - abs_res ); % write TB data write_tb( fid, xi, yi, ai, rx, ry, ra, mode ); end function [mul_res, mul_err, it ] = cmul( x, y, fid ) global C_FLAG_VEC_ROT C_FLAG_ATAN_3 C_MODE_CIRC C_MODE_LIN C_MODE_HYP global XY_WIDTH ANGLEWIDTH GUARDBITS RM_GAIN if( size( x ) ~= size( y ) ) error( 'size error' ) end xi = round( x * ( 2^(XY_WIDTH-1) -1 ) ); ai = round( y * ( 2^(XY_WIDTH-1) -1 ) ); yi = zeros( size( x ) ); mode = C_MODE_LIN; % cordic version [ rx, rmul, ra, it ] = cordic_iterative( xi, ... yi, ... ai, ... mode, ... XY_WIDTH, ... ANGLEWIDTH, ... GUARDBITS, ... RM_GAIN ); mul_res = rmul ./ (2^(ANGLEWIDTH-1)-1); mul_err = abs( y.*x - mul_res ); % write TB data write_tb( fid, xi, yi, ai, rx, rmul, ra, mode ) end function [div_res, div_err, it ] = cdiv( x, y, fid ) global C_FLAG_VEC_ROT C_FLAG_ATAN_3 C_MODE_CIRC C_MODE_LIN C_MODE_HYP global XY_WIDTH ANGLEWIDTH GUARDBITS RM_GAIN if( size( x ) ~= size( y ) ) error( 'size error' ) end xi = round( x * ( 2^(XY_WIDTH-1) -1 ) ); yi = round( y * ( 2^(XY_WIDTH-1) -1 ) ); ai = zeros( size( x ) ); mode = C_FLAG_VEC_ROT + C_MODE_LIN; % cordic version [ rx, ry, rdiv, it ] = cordic_iterative( xi, ... yi, ... ai, ... mode, ... XY_WIDTH, ... ANGLEWIDTH, ... GUARDBITS, ... RM_GAIN ); div_res = rdiv ./ (2^(ANGLEWIDTH-1)-1); div_err = abs( y./x - div_res ); % write TB data write_tb( fid, xi, yi, ai, rx, ry, rdiv, mode ) end <<<<<<< HEAD <<<<<<< HEAD ======= >>>>>>> Updated C and RTL model as well as the documentation function [x_res, y_res, x_err, y_err, it ] = crot( x, y, th, fid ) % % does a multiplication with exp( th * i ) % and therefore, a rotation of the complex input value x + yi where th % defines the rotation angle % global C_FLAG_VEC_ROT C_FLAG_ATAN_3 C_MODE_CIRC C_MODE_LIN C_MODE_HYP global XY_WIDTH ANGLEWIDTH GUARDBITS RM_GAIN xi = round( x * ( 2^(XY_WIDTH-1) -1 ) ); yi = round( y * ( 2^(XY_WIDTH-1) -1 ) ); ai = round( th .* (2^(ANGLEWIDTH-1)-1) ); mode = C_MODE_CIRC; [ rx ry ra, it ] = cordic_iterative( ... xi, ... yi, ... ai, ... mode, ... XY_WIDTH, ... ANGLEWIDTH, ... GUARDBITS, ... RM_GAIN ); tmp = ( x + 1i * y ) .* exp( i * th ); x_res = rx ./ ( 2^(XY_WIDTH-1)-1 ); y_res = ry ./ ( 2^(XY_WIDTH-1)-1 ); y_err = abs(x_res - real(tmp) ); x_err = abs(y_res - imag(tmp) ); % write TB data write_tb( fid, xi, yi, ai, rx, ry, ra, mode ) end function [sin_res, cos_res, sin_err, cos_err, it ]= cpol2cart( th, r, fid ) % % does the Matlab equivalent pol2cart % <<<<<<< HEAD ======= function [sin_res, cos_res, sin_err, cos_err, it ]= cpol2cart( th, r, fid ) >>>>>>> initial commit ======= >>>>>>> Updated C and RTL model as well as the documentation global C_FLAG_VEC_ROT C_FLAG_ATAN_3 C_MODE_CIRC C_MODE_LIN C_MODE_HYP global XY_WIDTH ANGLEWIDTH GUARDBITS RM_GAIN xi = r .* (2^(XY_WIDTH-1)-1); yi = zeros( 1, length( th ) ); ai = round( th .* (2^(ANGLEWIDTH-1)-1) ); <<<<<<< HEAD <<<<<<< HEAD mode = C_MODE_CIRC; ======= mode = C_MODE_CIRC; % cordic version >>>>>>> initial commit ======= mode = C_MODE_CIRC; >>>>>>> Updated C and RTL model as well as the documentation [ rcos rsin ra, it ] = cordic_iterative( ... xi, ... yi, ... ai, ... mode, ... XY_WIDTH, ... ANGLEWIDTH, ... GUARDBITS, ... RM_GAIN ); cos_res = rcos ./ ( 2^(XY_WIDTH-1)-1 ); sin_res = rsin ./ ( 2^(XY_WIDTH-1)-1 ); [ cos_m, sin_m ] = pol2cart( th, r ); sin_err = abs(sin_res - sin_m ); cos_err = abs(cos_res - cos_m ); % write TB data write_tb( fid, xi, yi, ai, rcos, rsin, ra, mode ) end <<<<<<< HEAD <<<<<<< HEAD ======= >>>>>>> initial commit ======= >>>>>>> Updated C and RTL model as well as the documentation function [atan_res, abs_res, atan_err, abs_err, it ] = ccart2pol( x, y, fid ) global C_FLAG_VEC_ROT C_FLAG_ATAN_3 C_MODE_CIRC C_MODE_LIN C_MODE_HYP global XY_WIDTH ANGLEWIDTH GUARDBITS RM_GAIN if( size( x ) ~= size( y ) ) error( 'size error' ) end ai = zeros( size( x ) ); xi = round( x * (2^(XY_WIDTH-1)-1) ); yi = round( y * (2^(XY_WIDTH-1)-1) ); mode = C_FLAG_VEC_ROT + C_MODE_CIRC; % cordic version [ rx, ry, ra, it ] = cordic_iterative( xi, ... yi, ... ai, ... mode, ... XY_WIDTH, ... ANGLEWIDTH, ... GUARDBITS, ... RM_GAIN ); <<<<<<< HEAD <<<<<<< HEAD ======= >>>>>>> Updated C and RTL model as well as the documentation % matlab version: m_th = atan2( y, x ); m_r = sqrt( x.^2 + y.^2 ); <<<<<<< HEAD ======= % matlab version [m_th, m_r ] = cart2pol( x, y ); >>>>>>> initial commit ======= >>>>>>> Updated C and RTL model as well as the documentation % comparison atan_res = ra ./ 2^( (ANGLEWIDTH)-1); abs_res = rx ./ ( 2^(XY_WIDTH-1) -1 ); atan_err = abs( m_th - atan_res ); abs_err = abs( m_r - abs_res ); <<<<<<< HEAD <<<<<<< HEAD % TODO: ATAN oder ATAN2 atan( 0 / x ) != atan2( 0, x )!!!! ======= >>>>>>> initial commit ======= % TODO: ATAN oder ATAN2 atan( 0 / x ) != atan2( 0, x )!!!! >>>>>>> Updated C and RTL model as well as the documentation % write TB data write_tb( fid, xi, yi, ai, rx, ry, ra, mode ) end
Go to most recent revision | Compare with Previous | Blame | View Log