The La Plata Mountains as seen from above the author’s
            home.


Durango Bill’s

“C” Program to generate Ramanujan Numbers
(Source Code)



Return to the main Ramanujan page

Web page generated via Sea Monkey's Composer HTML editor
within  a Linux Cinnamon Mint 18 operating system.
(Goodbye Microsoft)


//*************************************************************************
//
//                                Ramanujans.c
//                                    by
//                                Bill Butler
//
//  This program finds Ramanujan numbers. A Ramanujan number is a number
//  formed by the sum of two cubes in 2 or more different ways. For example:
//
//  12^3 + 1^3 = 9^3 + 10^3 = 1729
//
//  There are an infinite number of other paired cubes that have a common sum.
//
//  Also, the number of pairs can be extended to higher orders.
//
//  For example, the first quadruple, Taxicab(4), occurs at:
//
//        13,322^3 + 16,630^3 =
//        10,200^3 + 18,072^3 =
//        5,436^3 + 18,948^3 =
//        2,421^3 + 19,083^3 = 6,963,472,309,248  ( 6963472309248 )
//        (Takes about 30 seconds using this program on a 3GHz processor)
//        (Also nails Taxicab(5) in less than 3 1/4 hrs.)
//
//  Imagine that the following table is an Excel speadsheet. The first row
//  across the top counts upward for J = 1, 2, 3, 4, etc. The second row gives
//  the cube for each of these numbers: 1, 8, 27, 64, etc. Similarly, the first
//  column counts upward for I = 1, 2, 3, 4, etc. and the 2nd column gives the
//  cube for each of these numbers: 1, 8, 27, 64, etc. The rest of the cells in
//  the spreadsheet are the sum of two cubes using the respective rows and
//  columns.
//
//           J    1    2    3    4    5    6    7    8    9   10   11   12   13
//           J^3  1    8   27   64  125  216  343  512  729 1000 1331 1728 2197
//  I  I^3   ------------------------------------------------------------------
//  1    1   |    2    9   28   65  126  217  344  513  730 1001 1332 1729 2198
//  2    8   |    9   16   35   72  133  224  351  520  737 1008 1339 1736 2205
//  3   27   |   28   35   54   91  152  243  370  539  756 1027 1358 1755 2224
//  4   64   |   65   72   91  128  189  280  407  576  793 1064 1395 1792 2261
//  5  125   |  126  133  152  189  250  341  468  637  854 1125 1456 1853 2322
//  6  216   |  217  224  243  280  341  432  559  728  945 1216 1547 1944 2413
//  7  343   |  344  351  370  407  468  559  686  855 1072 1343 1674 2071 2540
//  8  512   |  513  520  539  576  637  728  855 1024 1241 1512 1843 2240 2709
//  9  729   |  730  737  756  793  854  945 1072 1241 1458 1729 2060 2457 2926
// 10 1000   | 1001 1008 1027 1064 1125 1216 1343 1512 1729 2000 2331 2728 3197
// 11 1331   | 1332 1339 1358 1395 1456 1547 1674 1674 1843 2060 2331 2662 3059
// 12 1728   | 1729 1736 1755 1792 1853 1944 2071 2240 2457 2728 3059 3456 3925
// 13 2197   | 2198 2205 2224 2261 2322 2413 2540 2709 2926 3197 3528 3925 4394
//
//  The table can be extended for as many rows and columns as needed. The lower
//  left triangle of the table is a mirror image of the upper right triangle.
//  Thus, the lower left triangle can be ignored.
//
//  Within the upper right triangle, we want to look for two identical numbers.
//  After some  searching, we note that the number 1729 appears twice. It is
//  found at row 1, column 12; and again at row 9, column 10. Thus 1^3 + 12^3
//  equals 9^3 + 10^3 equals 1729. Hence, 1729 is a Ramanujan number. If we
//  infinitely increased the size of the table, we could find an infinite
//  number of other paired cube sums. Also, we could find an infinite number of
//  "triples". The bad news is that the first of these "Ramanujan triples"
//  occurs at:
//  (Row 228, Column 423), (Row 167, Column 436), and (Row 255, Column 414)
//  for 228^3 + 423^3  =  167^3 + 436^3  =  255^3 + 414^3, which all share a
//  common cell entry of 87,539,319.
//
//  If we are to extend the process of finding additional Ramanujan triples or
//  even higher order matches (e.g. Ramanujan quadruples), three major problems
//  become apparent. 1) The table will become very large. 2) The search time
//  will become prohibitive. 3) The size of the numbers in the cells will
//  run into precision problems. (Too many digits for standard computer
//  hardware.)
//
//  We note that numbers in the table are confined to specific areas. Numbers
//  from 1 to 1,000 appear only in the upper left corner. Numbers between 1,000
//  and 2,000 occur in a band that extends from upper right to lower left.
//  Numbers between 2,000 and 3,000 exist only in another band that begins off
//  the right side of the table and extends down to the left. If we are looking
//  for possible number matches, we only have to look at one band at a time.
//
//  The program sequentially generates these bands. Each new band is used for
//  the current search process. Then this old band is discarded and a new band
//  is generated. The problem then is to search all of the numbers within a
//  band to see if duplicates exist. The example above uses a band width of
//  1,000 for the search area. The program initially uses 10 billion for
//  the band width (for 4 & 5 pair combinations), but since the density of the
//  I^3 + J^3 sums decreases with larger numbers, this band width is allowed
//  to expand with time.
//
//  The program generates a little over 2.4 million numbers at a time for each
//  new band. (Stabilizes at this rate after a few dozen iterations.) These
//  numbers are inserted into a hash table, and then the hash table is searched
//  to see if duplicate entries exist. If there are enough entries for any hash
//  number, then the associated link list of results for this hash number are
//  searched for actual solutions.
//
//  The algorithm (especially the hash index portion) is very efficient and
//  appears to be at least 10 times faster than the "heap" algorithm used by
//  David Wilson who published the first Ramanujan quintuple.
//
//  If the search is continued to include the lowest known Ramanujan sextuple,
//  (= Taxicab(6)), numbers in the table would have 23 decimal digits. No
//  simple data representation (or standard computer hardware) has this much
//  precision. Thus numbers are split into two pieces. One part includes the
//  rightmost 9 decimal digits and is maintained as "int" variables. The
//  remaining leftmost 14 decimal digits are maintained as "double" variables.
//  The integer portion does double duty as it is used to construct the hash
//  indexes.
//
//  The program has been updated from an old ANSI "C" version that ran under
//  DOS on a 80486 computer. (And that was a copy of a version that ran on a
//  80386 computer, and that was a copy of the original that I wrote for a
//  32032 32-bit National Semiconductor co-processor board that I added to an
//  IBM PC-XT back in the early 80's. (Those were the days when you sold your
//  soul for 1 MB of direct RAM addressing - no PC/DOS segmentation. Anybody
//  remember what "EDLIN" was?)
//
//  This program may be used, copied, modified, etc. without any obligation by
//  any person for any not-for-profit purpose. I would appreciate that this or
//  any derivative version would include a note crediting me with the original
//  program/algorithm. (e.g. "Original algorithm by Bill Butler.")
//
//  Final note: The program runs best if you have >= 1 GB of RAM. It runs as is
//  under Windows XP when compiled by the lcc-win32 C compiler as a "console"
//  program.
//
//***************************************************************************



#include <stdheaders.h>         //  The usual stdio.h, stdlib.h, etc

            //  "ArrayLimit" controls how far you want to search. The program
            //  will search for Ramanujans up to ArrayLimit^3. (No other
            //  modifications to the code are needed.)
            //  DO NOT use anything > 30000000.
            //  At 30,000,000 for ArrayLimit, the author was able to run time
            //  trials up to 2.15E22, and successfully run a test to see if it
            //  would find the best known candidate for Taxicab(6). (Started a
            //  short distance lower than this.) However, when the program was
            //  running with "ArrayLimit" at 30,000,000 at the same time that
            //  several other programs were running, the author's computer
            //  (1 GB RAM) crashed and corrupted Norton's "Go Back" files.
            //  Norton's "System Works" had to be uninstalled and then
            //  reinstalled. (System "Commit Charge" was not known, but an
            //  overload is possible.)

#define ArrayLimit 10000000     //  Program will crash if/when search passes
                                //  ArrayLimit^3

                                //  Procedure declarations
void InitSys(void);             //  Initialize the program
void NextGroup(void);           //  Generate next band of numbers
void CheckGroup(void);          //  Search for duplicate entries
void MakeCube(int);             //  Cubes the integer and places the results
                                //  in global variables CubedHigh and CubedLow
void pausemsg(void);            //  Pauses the program at key points


                                //  Note: Each 10000000 increase in ArrayLimit
                                //  uses another 160 MB of RAM. ArrayLimit of
                                //  30000000 is good enough for Taxicab(6).
                                //  (But estimated run time is 7 years.)
double NcubedHigh[ArrayLimit+1];//  NcubedHigh[i] = leftmost 14 digits of i^3
int NcubedLow[ArrayLimit+1];    //  NcubedLow[i] = 9 rightmost digits of i^3
                                //  These arrays are used to look up the cubes
                                //  of numbers when forming the I^3 + J^3 sums
int NextJ[ArrayLimit+1];        //  Keeps track of the next "J" to try when
                                //  forming trial I^3 + J^3.

                                //  The size of the next 2 arrays matches
                                //  the 22 bit hash size. Link lists for
                                //  the data arrays below start here.
unsigned HashHd[4194304];       //  Link heads for hash table.
int ChainLen[4194304];          //  Chain length.

                                //  As trial I^3 + J^3 numbers are generated
                                //  for each new "band"/group, they are
                                //  stored in these link lists. (Purists might
                                //  want to use a structure array, but time
                                //  trials for both encoding methods showed
                                //  that simple arrays execute faster.)
double I3J3High[3000000];       //  Leftmost 14 digits of I^3 + J^3
int I3J3Low[3000000];           //  9 rightmost digits of I^3 + J^3
int Ival[3000000];              //  The "I" in I^3 + J^3
int Jval[3000000];              //  The "J" in I^3 + J^3
int Link[3000000];              //  Link lists.

double CubedHigh;               //  The MakeCube() routine is passed an integer
int CubedLow;                   //  number. It cubes this number and places the
                                //  lowest (rightmost) 9 decimal digits in
                                //  CubedLow and the remaining digits in
                                //  CubedHigh. These are then copied into the
                                //  Ncubed arrays.

char Databuff[100];             //  Dummy array for user input.

int NbrPairs;                   //  User input for minimum number of pairs
                                //  required for output display. (=2,3,4,or 5)

double InitValues[10] = {       //  Initial upper limit & increment. Will keep
    0.0, 0.0, 2000.0,           //  output display in roughly sorted order.
    20000000.0,
    10000000000.0,
    10000000000.0 };

double UpperLim;                //  For each new group, find cube sums up to
                                //  this limit. UpperLim increases by
double Increment;               //  "Increment" for each new group. In turn,
                                //  "Increment" intermittently increases as
                                //  array storage allows. "Increment" is the
                                //  same as the "band width" phrase used in
                                //  the Excel Spreadsheet example.
int NbrStored;                  //  Nbr. of items in hash table/I3J3 arrays.
unsigned Bit22Mask = 4194303;   //  Mask for 22 bit index into Hash Table

int StatusFlag;                 //  Set by user for display/no-display of
                                //  status data while program runs.




int main(void) {

  InitSys();                        //  Initialize system.

  while(1) {                        //  Do until user stops the program
    NextGroup();                    //  Form next group of cube sums.
    CheckGroup();                   //  Look for matches.
                                    //  Process a little over 2,400,000
    if (NbrStored < 2400000)        //  trial I^3 + J^3 each iter.
      Increment *= 1.05;            //  Increases by 5 % as array space allows
    UpperLim += Increment;          //  Set up for next group

    if (StatusFlag) {               //  Optional status check
      printf("Upper Limit is now %g\n", UpperLim);
      printf("Increment is now %g\n", Increment);
    }
  }
  return 0;
}



//****************************************************************************
//
//                                    InitSys
//
//  This routine initializes the system. The Ncubed[] arrays are filled with
//  cubes such that NcubedHigh[i]*E9 + NcubedLow[i] = I^3. 23+ digits of
//  precision can be handled.
//
//  Note: The program allows you to begin your search at any arbitrary
//  starting point. Thus, 2 or more computers could each be running the program
//  to simultaneously search different zones in the number field.
//
//  The NcubedLow[] array contains the integer bit version of the last 9
//  decimal digits of all the I^3's. These will be used to form a hash index.
//  The hash index system greatly speeds up the process of finding matching
//  pairs of I^3 + J^3 = K^3 + L^3, etc. (The actual hash index adds the bit
//  portions of I^3 + J^3, and uses bits 8 to 29 of the result as the hash
//  index. Note: The least significant bit is bit "0".)
//
//  The NextGroup() routine will generate a large group of trial I^3 + J^3
//  numbers. The "J's" that will be used in this routine will be >= "I" until
//  the sum of I^3 + J^3 reaches the upper limit for the current group. (Group
//  size is kept within bounds by stopping the process at "UpperLim".). This
//  routine initializes the NextJ[] array for this process.
//
//  "UpperLim" and "Increment" are initially set to 10 billion for the first
//  iteration. (Valid for 4 & 5 pairs. Lower values for pairs = 2 & 3).
//  Subsequently "UpperLim" will be increased by "Increment" to include larger
//  trial values of I^3 + J^3. The density of I^3 + J^3 numbers becomes sparser
//  as the number field expands. Therefore, "Increment" is allowed to become
//  larger with time. Thus, the number of trial I^3 + J^3 numbers that will be
//  looked at will gradually cluster at a little over 2.4 million per
//  iteration.
//
//**************************************************************************

void InitSys(void) {

  int i;
  double start, Difference, Idbl, Jcalc;

  puts("             ramanujans.c");
  puts("                 by");
  puts("             Bill Butler\n");

  puts("This program finds Ramanujan numbers.");
  puts("It will run until you manually stop the program.\n");

  puts("The number that you enter below will determine how many pairs");
  puts("are required to display the associated Ramanujan number.\n");

  puts("Enter 2 for output of 2 or more pairs (Pauses on 3 pairs)");
  puts("Enter 3 for output of 3 or more pairs (Pauses on 4 pairs)");
  puts("Enter 4 for output of 4 or more pairs (Pauses on 5 pairs)");
  puts("Enter 5 for output of 5 or more pairs (Pauses on 6 pairs)");
  puts("(For output from 2, 3, 4 use \"Pause\" key as needed)");

  gets(Databuff);
  NbrPairs = atoi(Databuff);


  start = ArrayLimit;                           //  Convert to double for calc
  printf("\nWhere do you want to start the search? (0 <= start <= %g)\n",
        0.999999 * start * start * start);
  puts("(Suggest a very small percent below true start point if > 0.)");
  gets(Databuff);
  start = atof(Databuff);

  puts("\nDo you want to display status results (1 for yes  0 for no)?");
  puts("(If yes, don't forget that real data may scroll off the screen.)");
  gets(Databuff);
  StatusFlag = atoi(Databuff);

  puts("\nInitializing the cubes table");
  puts("(Will take a few seconds.)\n");

  for (i = 1; i <= ArrayLimit; i++) {
    MakeCube(i);                                //  Calculates i^3
    NcubedHigh[i] = CubedHigh;                  //  Leftmost 14 digits
    NcubedLow[i] = CubedLow;                    //  Rightmost 9 digits

                                                //  Initialize NextJ[]
    Idbl = i;                                   //  Convert to double
    Difference = start - Idbl*Idbl*Idbl;        //  Derived from nbr = I^3+J^3
    if (Difference < 0.0)                       //  If calc will be negative
      NextJ[i] = i;                             //  then ordinary start point
    else {                                      //  Else calc which J to use
      Jcalc = pow(Difference, 1.0/3.0) + 1.0;   //  Calculated value for J.
                                                //  Adding 1.0 avoids an array
                                                //  overflow problem
      if (Jcalc > Idbl)                         //  If this is bigger than "i"
        NextJ[i] = Jcalc;                       //  use the calculated value
      else                                      //  Else
        NextJ[i] = i;                           //  use the normal value.
    }
    /*    Optional debug check
    printf("At i = %d NcubedHigh = %'.0lf * E9  NcubedLow = %'d  NextJ = %d\n",
        i, NcubedHigh[i], NcubedLow[i], NextJ[i]);
    */
  }

  UpperLim = InitValues[NbrPairs] + start;      //  Initial upper limit.
  Increment = InitValues[NbrPairs];             //  Initial increment.

  puts("Starting search\n");
}



//*************************************************************************
//
//                                NextGroup
//
//  This routine generates the next group of I^3 + J^3 sums and places these
//  sums in the hash arrays. The number of these trial "I^3 + J^3" sums that
//  are processed will average a little over 2.4 million.
//
//*************************************************************************

void NextGroup(void) {

  int Count, i, j;
  unsigned i3j3sumLow;                  //  Does double duty as hash index
  double LimNbr, i3j3sumHigh;           //  Forces "LimNbr" to a "Register"
                                        //  position. Otherwise "UpperLim"
                                        //  could be used.

  for (i = 4194303; i >= 0; i--) {      //  Clear old garbage.
    HashHd[i] = 0;
    ChainLen[i] = 0;
  }
  Count = 0;                            //  Count nbr. of I^3+J^3 in the group
                                        //  and where to put data
                                        //  Note: Usage of the "i" & "j" vars.
                                        //  in this loop matches the "I" & "J"
                                        //  row/col labels in the "Excel
                                        //  Spreadsheet" example.
  i = 1;                                //  First "I" for I^3 (Start on row 1)
  LimNbr = UpperLim;                    //  Sets up reg. variable
  do {                                  //  Do for all "i" in group (Will move
                                        //  down until I^3 + J^3 is too big)
    j = NextJ[i];                       //  First "J" for J^3 (Leftmost column
                                        //  for current row in current group)
                                        //  Do for all "J" such that
    while(1) {                          //  I^3 + J^3 is < UpperLim
      i3j3sumHigh = NcubedHigh[i] + NcubedHigh[j];    //  Forms I^3 + J^3
      i3j3sumLow = NcubedLow[i] + NcubedLow[j];
      if (i3j3sumLow >= 1000000000) {   //  If carry overflow (>= 1 billion),
        i3j3sumLow -= 1000000000;       //  decrease integer portion by 1e9
        i3j3sumHigh += 1.0;             //  and increment the "billions"
      }
                                        //  If within upper limit then
      if ((i3j3sumHigh * 1.0E9 + i3j3sumLow) < LimNbr) {
        Count++;                        //  include it in the group.
        I3J3High[Count] = i3j3sumHigh;  //  Add to the hash arrays.
        I3J3Low[Count] = i3j3sumLow;
        Ival[Count] = i;
        Jval[Count] = j;
        i3j3sumLow >>= 8;               //  Generate the
        i3j3sumLow &= Bit22Mask;        //  hash index

        Link[Count] = HashHd[i3j3sumLow];    //  Update link list
        HashHd[i3j3sumLow] = Count;
        ChainLen[i3j3sumLow]++;
        j++;                            //  Move 1 col to right in
      }                                 //  "Excel spreadsheet"
      else                              //  Repeat until too big
        break;
    }
    NextJ[i] = j;                       //  Set up for next group. Move down
    i++;                                //  1 row in "Excel spreadsheet".
  } while (j > i);                      //  Loop exits when right edge of
                                        //  band/group intercepts the sloping
                                        //  diagonal in the spreadsheet. At
                                        //  this point, "i" will equal "j".
  NbrStored = Count;
                                        //  Optional status check.
                                        //  Averages ~2,400,000+ per crack.
  if (StatusFlag) {
    printf("This iter. stored %'d trial I^3 + J^3 sums\n", NbrStored);
    puts("Must stay < 3,000,000\n");
  }
}



//************************************************************************
//
//                                CheckGroup
//
//  This routine checks the hash arrays to see if any matches exist. If the
//  quantity of numbers stored in any hash link list is >= the user defined
//  display number, an actual Ramanujan solution may exist. If true, the
//  link list chain is checked to see if at least "NbrPairs" entries are
//  identical. If true, then the information is output.
//
//  The variable "NbrPairs" (see start of program) controls how many pairs
//  have to exist for the Ramanujan number before this info is output.
//
//************************************************************************

void CheckGroup(void) {

  int HashIndex, Ilink, Jlink;
  int Count;
  int StopNbr;
                                            //  For all hash heads
  for (HashIndex = 0; HashIndex <= 4194303; HashIndex++) {
    if ((ChainLen[HashIndex] - NbrPairs) < 0)    //  Chain is too short for output.
      continue;
                                            //  Optional check to see if
                                            //  hash algorithm is efficient.
/*
    printf("Link list[%d] has %d members\n", HashIndex,    ChainLen[HashIndex]);
*/
                                            //  Check the list for this hash.
                                            //  StopNbr controls how many times
                                            //  that "Ilink" moves forward
    StopNbr = ChainLen[HashIndex] - NbrPairs;
    for (Ilink = HashHd[HashIndex]; StopNbr >= 0;
            StopNbr--, Ilink = Link[Ilink]) {
      Count = 1;                            //  Will count number of identical
                                            //  I3J3 sums
      for (Jlink = Link[Ilink]; Jlink; Jlink = Link[Jlink]) {
        if (I3J3Low[Jlink] != I3J3Low[Ilink])    //  Not a match if either is
          continue;                              //  different.
        if (I3J3High[Jlink] != I3J3High[Ilink])
          continue;
                                        //  If to here, the I3J3 sum at Jlink
                                        //  matches the I3J3 sum at Ilink
        Count++;
      }
      if (Count < NbrPairs)                 //  If not enough pair matches,
        continue;                           //  then keep looking.

      printf("\n%d pairs exist at Ramanujan Number:    %'.0lf * E+9 + %'d\n",
        Count, I3J3High[Ilink], I3J3Low[Ilink]);
      for (Jlink = Ilink; Jlink; Jlink = Link[Jlink]) {
        if (I3J3Low[Jlink] != I3J3Low[Ilink])    //  Both must match for a
          continue;                              //  paired solution
        if (I3J3High[Jlink] != I3J3High[Ilink])
          continue;
        printf("   Pair at  I = %'d   J = %'d\n", Ival[Jlink], Jval[Jlink]);
      }
      if (Count > NbrPairs)                 //  Pause if big match
        pausemsg();
    }
  }
}



//*************************************************************************
//
//                                    MakeCube
//
//  This routine cubes the passed integer number and places the lowest 9
//  decimal digits of the result in the global integer variable "CubedLow"
//  and the high 14 digits into CubedHigh (double). Subsequently these
//  results are loaded into the Ncubed arrays.
//
//  The multiplication process is similar to ordinary base 10 arithmetic
//  (and its associated "carry") except base 10,000,000 is used.
//
//  The passed number "AnInteger" is known to be <= 3E7. This is converted
//  to a "double" number which can be safely squared without any precision
//  problems. This result is then split into three seperate sections that
//  have the equivalent of 7 decimal digits each. Each of these can be safely
//  multiplied by the original "AnInteger" again without any precision error.
//
//  Finally, the end result is placed into two variables with the lowest 9
//  decimal digits stored in one variable and the high 14 digits in the other.
//  These two variables are then placed in the global variables "CubedHigh"
//  and "CubedLow". (The latter becomes an integer variable.)
//
//**************************************************************************

void MakeCube(int AnInteger) {


  double HighDigits, MidDigits, LowDigits, temp, Original;

                                        //  "AnInteger" will be <= 3E7. Thus
                                        //  there is enough precision in an
                                        //  ordinary "double" variable for
                                        //  the first multiplication.
  Original = AnInteger;                 //  Convert to "double"
  temp = Original * Original;           //  Initially this will be <= 9E14
                                        //  Split it into high, mid, low digits
  LowDigits = fmod(temp, 10000000.0);   //  These will be the low 7 digits
  MidDigits = (temp - LowDigits) / 10000000.0;    //  This is the "carry"
                                        //  Now reduce MidDigits to 7 digits
  temp = fmod(MidDigits, 10000000.0);   //  This will be the middle 7 digits
  HighDigits = (MidDigits - temp) / 10000000.0;    //  Another "carry"
  MidDigits = temp;                     //  At this point "AnInteger" has been
                                        //  been squared. The lowest 7 digits are
                                        //  in "LowDigits", the next 7 digits are
                                        //  in "MidDigits", and the remaining
                                        //  digits are in "HighDigits".
  LowDigits *= Original;                //  Mult. all three by "Original" to get
  MidDigits *= Original;                //  cube. Then will have to sort out the
  HighDigits *= Original;               //  pieces. After this multiplication, the
                                        //  true cubed result would be:
                                        //  HighDigits*E14+MidDigits*E7+LowDigits

  temp = fmod(MidDigits, 100.0);        //  Get last 2 digits from MidDigits
  MidDigits -= temp;                    //  Remove last two digits
  MidDigits /= 100.0;                   //  Shift right two decimal digits
                                        //  so it measures "Billions"
  LowDigits += 10000000.0 * temp;       //  All 9 low digits are in LowDigits
                                        //  (But still has some higher digits)

  temp = fmod(LowDigits, 1000000000.0); //  These are the final lowest 9 digits
  MidDigits += (LowDigits - temp) / 1000000000.0;    //  Add carry to Mid Digits
  MidDigits += 100000.0 * HighDigits;   //  The 14 high digits are in place

  CubedHigh = MidDigits;                //  Store final result
  CubedLow = temp;                      //  Also converts it to an integer
}




//******************************************************************
//
//                    Misc. Routines
//
//******************************************************************

void pausemsg(void) {

  puts("\nLarge number of pairs exist");
  puts("Press RETURN to continue");
  gets(Databuff);
}