spot/buddy/examples/bddcalc
2004-07-12 14:03:53 +00:00
..
examples Initial revision 2004-06-28 14:19:59 +00:00
.cvsignore Merge BuDDy 2.3. 2004-06-28 15:22:11 +00:00
example.cal Initial revision 2004-06-28 14:19:59 +00:00
hashtbl.cxx Initial revision 2004-06-28 14:19:59 +00:00
hashtbl.h Initial revision 2004-06-28 14:19:59 +00:00
lexer.lxx Merge BuDDy 2.3. 2004-06-28 15:22:11 +00:00
Makefile.am Merge BuDDy 2.3. 2004-06-28 15:22:11 +00:00
parser.hxx Initial revision 2004-06-28 14:19:59 +00:00
parser.yxx * examples/bddcalc/parser.yxx (actionSeq, varlist): Rewrite as 2004-07-12 14:03:53 +00:00
parser_.h Merge BuDDy 2.3. 2004-06-28 15:22:11 +00:00
readme Initial revision 2004-06-28 14:19:59 +00:00
slist.h Initial revision 2004-06-28 14:19:59 +00:00

A PRIMITIVE BDD CALCULATOR
--------------------------

This is a small program that parses commands for a BDD calculator. The
input file contains the definition of the basic BDD variables (inputs)
and a sequence of actions - typically assignments. The calculator can
be used for verification of combinatorial circuits (tautology check),
such as some of the ISCAS85 circuits in the "examples" directory. These
ISCAS85 have been modified by another program before they where added to
the BuDDy examples, so please do not compare the runtime results to other
test runs.


A BDD calculator file could be something like this ("example.cal"):

  initial 100 100;

  inputs
     a b c;

  actions
     t1 = (a | b) & c;
     t2 = (a & c) | (b & c);
     t3 = t1 <> t2;
     tautology t3;  /* Verify that (a | b) & c == (a & c) | (b & c) */
   

Where the blocks are like this:

initial n c
-----------
  Initialize the BDD package using 'n' bddnodes and 'c' elements in
  the caches. This part is mandatory.


inputs id-seq
-------------
  Define all identifiers in the sequence 'id-seq' to be primary
  inputs. This also corresponds to the initial BDD variable ordering,
  starting with the first identifier in the top.

  
outputs id-seq
--------------
  Define all identifiers in the sequence 'id-seq' to be primary
  outputs. The sequence is space separated.


actions act-seq
---------------
  A list of all the calculations and more. This list is traversed and
  interpreted in the same order as written in the file. An action can
  be one of the following:

  Assignments:  "id = expression"
    Calculate the right-hand side and assign the value to the
    left-hand identifier. The expression may contain:

         +  identifiers : A previously defined identifier
	 +  true        : The constant true BDD
	 +  false       : The constant false BDD
	 +  ( ... )     : Parenteses
	 +  E1 & E2     : Conjunction of two sub-expressions
	 +  E1 | E2     : Disjunction of two sub-expressions
	 +  E1 ^ E2     : Xor
	 +  E1 => E2    : Implication
	 +  E1 <> E2    : Biimplication
	 +  ~E          : Negation
         + exist  V.E   : 
         + forall V. E  : Existential/Universal quantification of the
		          variables V in the expression E. V is a space
                          separated list of input names.


  Tautology check:  "tautology id"
    Check the variable 'id' for tautology (being equal to the constant
    true BDD).

  Size information: "size id"
    Print the number of distinct BDD nodes used to represent 'id'.

  Dump as a dot file: "dot "filename" id"
    Dump the BDD representing 'id' as commands to the graph drawing
    program Dot. The commands are written to the file 'filename'.

  Reordering: "reorder mtd"
    Do a dynamic variable reordering now using the method 'mtd'. The
    argument 'mtd' may be either: win2, win2ite or sift.
   
  Automatic reordering: "autoreorder times mtd"
    Enable automatic dynamic variable reordering using the method
    'mtd'. The argument 'mtd' may be either: win2, win2ite, sift or
    none. Use none to disable automatic reordering. The 'times'
    argument supplies the number of times that reordering may be
    done. Use for example 1 if you only want a "one-shot" reordering.

  Cache statistics: "cache"
    Print various cache statistics for the BDD kernel.