(User.programs)
Item: 145 by _rjs at hpcvbbs.UUCP
Author: [rj schwartz]
  Subj: BASE OF NATURAL LOG (e) TO N PLACES
  Date: Tue Nov 06 1990 10:49 

BASE OF NATURAL LOG (e) TO N PLACES...
 
 The program below produces a vector containing 10 digit integers which
  represent the fractional part of 'e' (.71828...).
  
 The program takes a single argument (from level 1) which is a positive
  integer divisible by 10.  The argument specifies the number of decimal
  places that are to be determined.
 
 Note - due to truncation the last few decimal places may be incorrect -
  for up to 150 digits the last 2 may be incorrect, above that and up to
  2500 digits the last 3 may be incorrect and above that and up to 35000
  digits the last 4 may be incorrect.
   
 The time to produce 30 digits (28 accurately) on the HP48SX w/ROM-D
  is 30 secs.  A similar program written some years ago for the HP-65
  took 210 secs to produce 32 digits.
 
 I've written a PC based programmable calculator program which, when run
  on a 33MHz 80386/80387, will generate 2000 digits in a little over 30
  mins.  An assembly language routine using the same algorithm and run
  on the same machine takes 1.2 secs.  I estimate that the HP48SX would
  take on the order of 15 hrs. to produce this many digits and the result
  may be incorrect - it depends on the precision of the internal floating
  point representation.  Does anyone know what the internal representation is?  
 
 The correct fractional part to 2000 decimal places (actually 1996)
  follows the program listing.
 
 Also, let me know if you're interested in obtaining a copy of the PC based
  calculator mentioned above.  The documentation is incomplete, but I'll
  finish it if there's enough interest.  A brief description of the program
  features follows:
 
              PC - Personal Computer Programmable Calculator
                      (c) 1989  RJ Schwartz   
 
 LIST OF FEATURES
 
  .   Reverse Polish Notation (RPN) with 4 register stack plus "Last x"
       register.
 
  .   15 - 16 digit accuracy, range - 1.E-308 to 1.E308, user 
       defined display precision.
 
  .   8000 data registers addressable directly or indirectly,
       4 function register arithmetic, increment (decrement) register.
   
  .   100 general purpose flag (Boolean) registers:
         1 - 19   System flags
        20 - 39   Unaffected by flag test
        40 - 59   Cleared by flag test
        60 - 79   Set by flag test
        80 - 99   Complemented by flag test
      Flag operations - set, complement, test on, test off.
       System flags indicate data entry, stack lift status,
       print enable, print to file, error conditions.
 
  .   Built in functions include trigonometrics in degree, radian,
       gradian or hyperbolic modes, e^x, ln x, 10^x, log x, abs,
       sgn, factorial, combinations and permutations, rect <-> polar
       conversions, integer, fraction, remainder, round, random
       number generator, reciprocal, square root, add, subtract,
       multiply, divide, exponentiate and more.  All functions can
       be included in programs.
 
  .   10000 step program capacity - 6 conditional tests, 100 levels of
       nested subroutines, GOTO label, GOTO prog. step, relative jump,
       decrement or increment any data register with variable length 
       jump on zero or non-zero.      
 
   .  Trace and/or time execution, single step, set breakpoint,
       variable length pause.
 
  .   40 user defined keystroke Macros (up to 40 characters each).
 
  .   Save (Load) registers, macros and/or flags to (from) disk.
 
  .   Formatted print (numeric or string data) to printer or to disk.
 
  .   Display graphs of user defined functions.
 
  .   Variable length BEEP and TONE functions.  
           
  .   Statistical functions including mean, standard deviation,
       and linear regression.
 
  .   Executes over 3000 program steps per second on an 80386 CPU @ 20 MHZ.
 
 
 HARDWARE REQUIREMENTS/RECOMENDATIONS
 
    .  Requires computer with Intel 8088 or 80x86 CPU, Microsoft
        DOS version 3.1 or higher,  IBM compatible BIOS and CGA, EGA or
        VGA graphics.
 
    .  Recommended are math coprocessor (80x87), a fast CPU and EGA or VGA
        graphics cards with color display.
 
 
 REQUEST FOR ASSISTANCE
 
 I've just purchased the HP48 after many years away from HP calculators.
  (My last was the HP-65, which, incidentally, was stolen - does anyone
  have one for sale?).  Anyway, I see from comments on this BBS,  that
  low level programming is now possible on HP48 and some of its predecessors.
  Can someone please direct me to information on the following for the HP48:
     
       1. The CPU architecture (including register sizes).
       2. The interrupt capabilities (is that what controls the clock
           and the keyboard?).
       3. The instruction set (I've seen some listings on this BBS but
           the adressing modes are foreign to me. eg. add.a  c,a).
       4. A list of ROM entry points (again I've seen partial listings
           but there seems to be much more information available).
       5. The internal structure of HP48 objects.
 
 Thanks in advance

[RJ: I'll bet Voyager is exactly what you've been waiting for!  -jkh-]