predefined function/library

Discussion in 'VHDL' started by Vincent, Sep 7, 2003.

  1. Vincent

    Vincent Guest


    I tried to compile a vhdl file containing the function
    Which library contain this function ?


    Vincent, Sep 7, 2003
    1. Advertisements

  2. Vincent,

    I know the function addum (ADD Unsigned Magnitued) from the Viewlogic
    environment in the past.
    If your design includes types like VLBIT, VLBIT_1D, indeed ViewLog BIT, you
    probably need a package with the name pack1076.vhd to compile it in a VHDL
    complaint environment.

    With a google for "pack1976.vhd" I found this location
    Be sure that NOT to use VHDL-1993 but use VHDL-1987. (This package is
    already very old!).

    This function addum used vlbit_1d (not std_logic_vector). It can be that
    someone has used the same name but for type std_logic-vector.


    Egbert Molenkamp
    Egbert Molenkamp, Sep 8, 2003
    1. Advertisements

  3. Vincent

    Vincent Guest

    yes it is exactly what you mean !
    it is used with types vlbit, vlbit_1d, ...
    but it was said in the readme file it may be replaced by std_logic
    so i replaced vlbit -> std_logic
    vlbit_1d -> std_logic_vector
    v1d2int( -> to_integer(unsigned(.
    (the design is quite short, it is to design a PIC core)

    but for the function, which one may i use to remplace addum, with the
    std_logic library ?

    Furthermore, i don't know if it was for the ViewLogic library but the
    program use the following library :
    -- library synth;
    -- use synth.stdsynth.ALL;
    i deleted it, is it useful ?


    Vincent, Sep 8, 2003
  4. At that time, ~1990, most CAE tools had their own logical type. The VHDL
    standard had type BIT and the users also wanted tri-state etc.
    VLBIT is a four value logic ('X','0','1','Z'). It maybe that you can do the
    suggetsed transformation in the package pack1076.vhd but I guess a synthesis
    nowadays does not recognize that ADDUM is addition and will produce to
    much/slow logic?
    Since it is short design would indeed consider change it in std_logic.
    It is already clear that ADDUM performs an unsigned addition. So if the
    vectors are added the shortes vector is zero-exented (the function ADD2C
    performs a signed addition, in this the shortest vector in sign extended).
    If you change your design I would suggest using the IEEE package NUMERIC_STD
    (based on std_logic) (USE IEEE.NUMERIC_STD.ALL).
    variable a : unsigned (3 downto 0);
    variable b,c : unsigned (5 downto 0);

    c := a +b; performs the unsigned addition (due to the type 'unsigned').

    Ok. the is the solution HOWEVER wait a moment .. If you add two vector you
    may need an extra bit for the result! The function ADDUM generates a result
    that is 1 bit longer then the longest vector (.e.g. in contrast to the "+"
    form the package NUMERIC_STD (or STD_LOGIC_ARITH) ).
    variable a : unsigned (3 downto 0);
    variable b : unsigned (5 downto 0);
    variable c : unsigned (6 downto 0);

    c := ('0'&b) + a; -- behaviour of addum; extend the longest vector with

    However I remember that in most designs the extra bit was not needed. So
    addum 'generated' it, and the user removed the extra bit. So maybe you can
    simple perform an addition.
    Yes .. at that time. When I remember ViewLogic's synthesis tool needed it.
    But indeed you can delete it now.

    Egbert Molenkamp
    Egbert Molenkamp, Sep 8, 2003
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.