Using fprintf to output hex values
Steve Heidorn
I'm having trouble using fprintf with the %x format specifier when
printing negative numbers. For instance,
fprintf('%x\n',11) = b
but
fprintf('%x\n',-11) = -1.1000000e+00
In other words, the input argument of '-11' was passed through unchanged.
We tried this in C, and the output was as expected so long as the input
argument was of type integer. However, C also messed up when the input
argument was of type float. Either way, I seem to be stuck since Matlab
has no notion of integer typing. Matlab support says fprintf is working
as designed. Is there a way around this?
Steve Heidorn Phone: 703-367-5377
Combat Systems Architecture Fax: 703-367-5067
Loral Federal Systems - Manassas Internet: hei...@lfs.loral.com
Tom L. Davis
(Steve Heidorn) wrote:
> Subject: Using fprintf to print hex values
>
> I'm having trouble using fprintf with the %x format specifier when
> printing negative numbers. For instance,
>
> fprintf('%x\n',11) = b
>
> but
>
> fprintf('%x\n',-11) = -1.1000000e+00
>
> In other words, the input argument of '-11' was passed through unchanged.
> We tried this in C, and the output was as expected so long as the input
> argument was of type integer. However, C also messed up when the input
> argument was of type float. Either way, I seem to be stuck since Matlab
> has no notion of integer typing. Matlab support says fprintf is working
> as designed. Is there a way around this?
The problem is double precision floating point can represent integers upto
2^52 or so. The x format assumes unsigned 32 bit integers. The following
code converts negative integers to 32 bit integers:
x = -11;
y = mod( floor(x), pow2(32) );
fprintf( 1, '%08x\n', y )
Let me know if you find a better, faster way. Here's TMW's suggestion for
the mod function:
function z = mod(x, y)
%MOD Remainder after division.
% If X and Y are matrices of the same size, or one of them is a scalar,
% then both MOD(X,Y) and the built-in function REM(X,Y) compute the
% remainder after division of the elements of X by the elements of Y.
% The two functions agree where X and Y have the same sign, but
% differ by Y where X and Y have opposite signs.
% Specificially,
% MOD(X, Y) = X - Y .* FLOOR(X ./ Y)
% REM(X, Y) = X - Y .* FIX(X ./ Y)
% This implies that
% SIGN(MOD(X,Y)) = SIGN(Y)
% SIGN(REM(X,Y)) = SIGN(X)
% Note that MOD(X,Y) is a periodic function of X with period Y,
% but that REM(X,Y) is not periodic when X changes sign.
% And, there is one special case: Y = 0, MOD(X, 0) = X, REM(X, 0) = NaN.
%
% Examples:
% mod(0.1000,pi) = 0.1000
% mod(-0.1000,pi) = 3.0416
% rem(0.1000,pi) = 0.1000
% rem(-0.1000,pi) = -0.1000
%
% See also REM, FLOOR, FIX.
% Copyright (c) 1984-94 by The MathWorks, Inc.
z = x - y .* floor(x ./ (y + (y == 0)));