MATLAB: Difference between int32 and fixed point

data typefixed-pointHDL Coder

I am trying to program a FPGA using MATLAB and the HDL coder. So for the FPGA i need to have fixed point data types. If i am using data type of int32 is it fixed point? or i should use the fi object to convert my data to fixed point.

Best Answer

HDL Coder treats all integral types (both int and uint types) as the fixed point types that they are equivalent to. There is no need to convert integral types to fixed point.

For example, an int16 type is equivalent to fixdt(1,16,0), while an uint8 is equivalent to fixdt(0,8,0).

Related Solutions

MATLAB: How to convert binary data into fixed point data

First, get the raw integer values in to MATLAB as either uint32 values or int32 values.

Second, use the reinterpretcast function

Example run this code

% Get the raw integers into a uint32 or int32
%   see MATLAB importdata and similar tools for ways
%   to accomplish this step
%


rawInteger = uint32(4275878552)
% Reinterpret data in the appropriate fixed-point type
%
vFixedPoint = reinterpretcast(rawInteger,numerictype(1,32,16))
% Compare hex values before and after
%   just to show the underlying stored integers are preserved
%
hexOrig = lower(dec2hex(rawInteger))
hexFixedPoint = vFixedPoint.hex
compareHexBeforeAfter = [hexOrig;hexFixedPoint]

see this output

rawInteger =
uint32
 4275878552
hexOrig =
  'fedcba98'
vFixedPoint = 
  -2.912711181640625e+02
        DataTypeMode: Fixed-point: binary point scaling
          Signedness: Signed
          WordLength: 32
      FractionLength: 16
hexFixedPoint =
  'fedcba98'
compareHexBeforeAfter  =
2×8 char array
  'fedcba98'
  'fedcba98'

For reference see the help

 >> help reinterpretcast
 reinterpretcast Convert fixed-point or integer data types without changing underlying data
    C = reinterpretcast(A, T) converts the input (integer or fi object) A
    to the data type specified by numerictype object T, without changing
    the underlying (stored integer) data. The result is returned in C.
    The data type of the input A must be fixed point or integer. T must be
    a numerictype object with a fully specified fixed-point data type.  The
    word length of inputs A and T must be the same.
    The reinterpretcast function differs from the MATLAB TYPECAST and CAST
    functions, in that it only operates on fi and integer types, and it
    does not allow the word length of the input to change.
    EXAMPLE:
      %%Convert from signed 8,7 to unsigned 8,0.
      a = fi([-1 pi/4], true, 8, 7)
      %   returns [-1.0000    0.7891] s8,7
      T = numerictype(false, 8, 0);
      b = reinterpretcast(a, T)
      %   returns [128   101] u8,0
      % Their binary representations are identical
      binary_rep = [bin(a);bin(b)]
      %    returns 10000000   01100101
      %            10000000   01100101

MATLAB: How to convert int32 to real number

X_real = double(X_int32);

HTH

Best Answer

Related Solutions

MATLAB: How to convert binary data into fixed point data

MATLAB: How to convert int32 to real number

Related Question