PAL CARD CATALOG ENTRY

SHORT DESCRIPTION

36-bit machine arithmetic on a 32-bit machine


MOVEMENT WITHIN THE PAL CARD CATALOG
Move to top-level taxonomy
Move to keyword list

ASSET PROFILE

UNIT NAME
Emulation of Machine Arithmetic
VERSION
1.0
REVIEW CODE
AR; C1 1.0 A
INET ADDRESS
Not documented in PAL database
AUTHOR
SYSCON Corp
3990 Sherman Street
San Diego, CA 92110
John Reddan
(619) 296-0085
Contact: Lt. Colonel Falgiano
ESD/SCW
Hanscom AFB, MA 01731
RIGHTS
PUBLIC DOMAIN
COPYRIGHT
(c) 1985, SYSCON Corp
DATE CREATED
June, 1985
DATE RELEASED
June, 1985
DATE LAST UPDATED
June, 1985
LOCATION
ASR
AdaNET
ENVIRONMENT
VAX/VMS TeleSoft
LIMITATIONS
Not documented in PAL database
CERTIFICATION
Ada System Certifier_1 1.0
Date/Time of Processing: Monday    30 May       1994 11:28:40Am
Overall Assessment of System: OK
Classification of System: A
Basis of Classification --
Syntax Errors                               PASS
Completeness                                PASS
Independence from External Libraries        PASS
Independence from a Specific Ada Compiler   PASS

Number of ...
Files               1
Library Units      10
Lines            6521
Statements       2696
Comments         1490

FILE LISTING

Directory Display


languages/ada/asr/math/m36:
  File Name                 Size
  ---------                 ----
  m36.zip                 84,081


Totals
  ==============  ==============
    1 Files               84,081

ABSTRACT

        The purpose of this package is to emulate 36 bit machine  host
arithmetic  on  a  32  bit  host  machine.   This package will provide
support for 36 bit integer, real, and double precision real numbers in
the form of the standard predefined arithmetic operations.  The ranges
of the supported types are as follows:

   Integer
      range of -2**35 to 2**35-1
   Real
      range of 10**-38 to 10**38 and 0
      mantissa => 27 bit binary fraction
      exponent => -128 to 127
   Double Precision Real
      range of 10**-38 to 10**38 and 0
      mantissa => 63 bit binary fraction
      exponent => -128 to 127

        Any  errors which occur during use of the arithmetic exception
declaration in the package specification can be changed to a rename of
the  predefined  exception  "NUMERIC_ERROR"  for  programs  needing to
handle arithmetic errors in a general fashion.   Conversion  functions
are  provided  to  assist in programming mixed operand (32 and 36 bit)
arithmetic, and to facilitate IO.  These functions should  be  renamed
if they will be used extensively so that the impact of the readability
of a program's arithmetic expressions is  minimized.   The  underlying
arithmetic will be performed in twos complement arithmetic.


REVISION HISTORY

DATE           VERSION AUTHOR                  HISTORY 
6/85           1.0     SYSCON                  Initial Release


RELEASE NOTICE

This prologue must be included in all copies of this software.

This software is copyright by the author.

This software is released to the Ada community.
This software is released to the Public Domain (note:
  software released to the Public Domain is not subject
  to copyright protection).
Restrictions on use or distribution:  NONE


DISCLAIMER

	This software and its documentation are provided "AS IS" and
without any expressed or implied warranties whatsoever.  No warranties
as to performance, merchantability, or fitness for a particular
purpose exist.
	The user is advised to test the software thoroughly before
relying on it.  The user must assume the entire risk and liability of
using this software.  In no event shall any person or organization of
people be held responsible for any direct, indirect, consequential or
inconsequential damages or lost profits.