Huge numbers

This unit uses an array of bytes to represent a LARGE number. The number is binairy-stored in the array, with the Least Significant Byte (LSB) first and the Most Significant Byte (MSB) last, like all Intel-integer types.

Arithmetic is not 10-based or 2-based, but 256-based, so that each byte represents one (1) digit.

The HugeInttype numbers are Signed Numbers.

When Compiled with the R+ directive, ADD and MUL wil generate an "Arithmetic Overflow Error" (RunError(215)) when needed. Otherwise use the "HugeIntCarry" variable.

Use the "HugeIntDiv0" variable to check on division by zero.

Use {$DEFINE HugeInt_xx } or "Conditional defines" from the "Compiler options" for sizing, where xx is 64, 32 or 16, otherwhise HugeIntSize will be set to 8 bytes.

unit HugeInts;

interface

const

{$IFDEF HugeInt_64 }

HugeIntSize = 64;

{$ELSE}{$IFDEF HugeInt_32 }

HugeIntSize = 32;

{$ELSE}{$IFDEF HugeInt_16 }

HugeIntSize = 16;

{$ELSE}

HugeIntSize = 8;

{$ENDIF}{$ENDIF}{$ENDIF}

HugeIntMSB = HugeIntSize-1;

type

HugeInt = array[0..HugeIntMSB] of Byte;

const

HugeIntCarry: Boolean = False;

HugeIntDiv0: Boolean = False;

procedure HugeInt_Min(var a: HugeInt); { a := -a }

procedure HugeInt_Inc(var a: HugeInt); { a := a + 1 }

procedure HugeInt_Dec(var a: HugeInt); { a := a - 1 }

procedure HugeInt_Add(a, b: HugeInt; var R: HugeInt); { R := a + b }

procedure HugeInt_Sub(a, b: HugeInt; var R: HugeInt); { R := a - b }

procedure HugeInt_Mul(a, b: HugeInt; var R: HugeInt); { R := a * b }

procedure HugeInt_Div(a, b: HugeInt; var R: HugeInt); { R := a div b }

procedure HugeInt_Mod(a, b: HugeInt; var R: HugeInt); { R := a mod b }

function HugeInt_IsNeg(a: HugeInt): Boolean;

function HugeInt_Zero(a: HugeInt): Boolean;

function HugeInt_Odd(a: HugeInt): Boolean;

function HugeInt_Comp(a, b: HugeInt): Integer; {-1:a< 0; 1:a>

}

procedure HugeInt_Copy(Src: HugeInt; var Dest: HugeInt);{ Dest := Src }

procedure String2HugeInt(AString: string; var a: HugeInt);

procedure Integer2HugeInt(AInteger: Integer; var a: HugeInt);

procedure HugeInt2String(a: HugeInt; var S: string);

implementation

procedure HugeInt_Copy(Src: HugeInt; var Dest: HugeInt);

{ Dest := Src }

begin

Move(Src, Dest, SizeOf(HugeInt));

end;{ HugeInt_Copy }

function HugeInt_IsNeg(a: HugeInt): Boolean;

begin

HugeInt_IsNeg := a[HugeIntMSB] and $80 > 0;

end;{ HugeInt_IsNeg }

function HugeInt_Zero(a: HugeInt): Boolean;

var i: Integer;

begin

HugeInt_Zero := False;

for i := 0 to HugeIntMSB do

if a[i] <> 0 then Exit;

HugeInt_Zero := True;

end;{ HugeInt_Zero }

function HugeInt_Odd(a: HugeInt): Boolean;

begin

HugeInt_Odd := a[0] and 1 > 0;

end;{ HugeInt_Odd }

function HugeInt_HCD(a: HugeInt): Integer;

var i: Integer;

begin

i := HugeIntMSB;

while (i > 0) and (a[i] = 0) do Dec(i);

HugeInt_HCD := i;

end;{ HugeInt_HCD }

procedure HugeInt_SHL(var a: HugeInt; Digits: Integer);

{ Shift "a" "Digits", digits (bytes) to the left,

"Digits" bytes will 'fall off' on the MSB side

Fill the LSB side with 0's }

var t: Integer;

b: HugeInt;

begin

if Digits > HugeIntMSB then

FillChar(a, SizeOf(HugeInt), 0)

else if Digits > 0 then

begin

Move(a[0], a[Digits], HugeIntSize-Digits);

FillChar(a[0], Digits, 0);

end;{ else if }

end;{ HugeInt_SHL }

procedure HugeInt_SHR(var a: HugeInt; Digits: Integer);

var t: Integer;

begin

if Digits > HugeIntMSB then

FillChar(a, SizeOf(HugeInt), 0)

else if Digits > 0 then

begin

Move(a[Digits], a[0], HugeIntSize-Digits);

FillChar(a[HugeIntSize-Digits], Digits, 0);

end;{ else if }

end;{ HugeInt_SHR }

procedure HugeInt_Inc(var a: HugeInt);

{ a := a + 1 }

var

i: Integer;

h: Word;

begin

i := 0; h := 1;

repeat

h := h + a[i];

a[i] := Lo(h);

h := Hi(h);

Inc(i);

until (i > HugeIntMSB) or (h = 0);

HugeIntCarry := h > 0;

{$IFOPT R+ }

if HugeIntCarry then RunError(215);

{$ENDIF}

end;{ HugeInt_Inc }

procedure HugeInt_Dec(var a: HugeInt);

{ a := a - 1 }

var Minus_1: HugeInt;

begin

{ this is the easy way out }

FillChar(Minus_1, SizeOf(HugeInt), $FF); { -1 }

HugeInt_Add(a, Minus_1, a);

end;{ HugeInt_Dec }

procedure HugeInt_Min(var a: HugeInt);

{ a := -a }

var i: Integer;

begin

for i := 0 to HugeIntMSB do

a[i] := not a[i];

HugeInt_Inc(a);

end;{ HugeInt_Min }

function HugeInt_Comp(a, b: HugeInt): Integer;

{ a = b: ==0; a > b: ==1; a < b: ==-1 }

var

A_IsNeg, B_IsNeg: Boolean;

i: Integer;

begin

A_IsNeg := HugeInt_IsNeg(a);

B_IsNeg := HugeInt_IsNeg(b);

if A_IsNeg xor B_IsNeg then

if A_IsNeg then HugeInt_Comp := -1

else HugeInt_Comp := 1

else

begin

if A_IsNeg then HugeInt_Min(a);

if B_IsNeg then HugeInt_Min(b);

i := HugeIntMSB;

while (i > 0) and (a[i] = b[i]) do Dec(i);

if A_IsNeg then { both negative! }

if a[i] > b[i] then HugeInt_Comp := -1

else if a[i] < b[i] then HugeInt_Comp := 1

else HugeInt_Comp := 0

else { both positive }

if a[i] > b[i] then HugeInt_Comp := 1

else if a[i] < b[i] then HugeInt_Comp := -1

else HugeInt_Comp := 0;

end;{ else }

end;{ HugeInt_Comp }

procedure HugeInt_Add(a, b: HugeInt; var R: HugeInt);

{ R := a + b }

var

i: Integer;

h: Word;

begin

h := 0;

for i := 0 to HugeIntMSB do

begin

h := h + a[i] + b[i];

R[i] := Lo(h);

h := Hi(h);

end;{ for }

HugeIntCarry := h > 0;

{$IFOPT R+ }

if HugeIntCarry then RunError(215);

{$ENDIF}

end;{ HugeInt_Add }

procedure HugeInt_Sub(a, b: HugeInt; var R: HugeInt);

{ R := a - b }

var

i: Integer;

h: Word;

begin

HugeInt_Min(b);

HugeInt_Add(a, b, R);

end;{ HugeInt_Sub }

procedure HugeInt_Mul(a, b: HugeInt; var R: HugeInt);

{ R := a * b }

var

i, j, k: Integer;

A_end, B_end: Integer;

A_IsNeg, B_IsNeg: Boolean;

h: Word;

begin

A_IsNeg := HugeInt_IsNeg(a);

B_IsNeg := HugeInt_IsNeg(b);

if A_IsNeg then HugeInt_Min(a);

if B_IsNeg then HugeInt_Min(b);

A_End := HugeInt_HCD(a);

B_End := HugeInt_HCD(b);

FillChar(R, SizeOf(R), 0);

HugeIntCarry := False;

for i := 0 to A_end do

begin

h := 0;

for j:= 0 to B_end do

if (i + j) < HugeIntSize then

begin

h := h + R[i+j] + a[i] * b[j];

R[i+j] := Lo(h);

h := Hi(h);

end;{ if }

k := i + B_End + 1;

while (k < HugeIntSize) and (h > 0) do

begin

h := h + R[k];

R[k] := Lo(h);

h := Hi(h);

Inc(k);

end;{ while }

HugeIntCarry := h > 0;

{$IFOPT R+}

if HugeIntCarry then RunError(215);

{$ENDIF}

end;{ for }

{ if all's well... }

if A_IsNeg xor B_IsNeg then HugeInt_Min(R);

end;{ HugeInt_Mul }

procedure HugeInt_DivMod(var a: HugeInt; b: HugeInt; var R: HugeInt);

{ R := a div b a := a mod b }

var

MaxShifts, s, q: Integer;

d, e: HugeInt;

A_IsNeg, B_IsNeg: Boolean;

begin

if HugeInt_Zero(b) then

begin

HugeIntDiv0 := True;

Exit;

end{ if }

else HugeIntDiv0 := False;

A_IsNeg := HugeInt_IsNeg(a);

B_IsNeg := HugeInt_IsNeg(b);

if A_IsNeg then HugeInt_Min(a);

if B_IsNeg then HugeInt_Min(b);

if HugeInt_Comp(a, b) < 0 then

{ a<b; no need to divide }

FillChar(R, SizeOf(R), 0)

else

begin

FillChar(R, SizeOf(R), 0);

repeat

Move(b, d, SizeOf(HugeInt));

{ first work out the number of shifts }

MaxShifts := HugeInt_HCD(a) - HugeInt_HCD(b);

s := 0;

while (s <= MaxShifts) and (HugeInt_Comp(a, d) >= 0) do

begin

Inc(s);

HugeInt_SHL(d, 1);

end;{ while }

Dec(s);

{ Make a new copy of b }

Move(b, d, SizeOf(HugeInt));

{ Shift d as needed }

HugeInt_ShL(d, S);

{ Use e = -d for addition, faster then subtracting d }

Move(d, e, SizeOf(HugeInt));

HugeInt_Min(e);

Q := 0;

{ while a >= d do a := a+-d and keep trek of # in Q}

while HugeInt_Comp(a, d) >= 0 do

begin

HugeInt_Add(a, e, a);

Inc(Q);

end;{ while }

{ OOps!, one too many subtractions; correct }

if HugeInt_IsNeg(a) then

begin

HugeInt_Add(a, d, a);

Dec(Q);

end;{ if }

HugeInt_SHL(R, 1);

R[0] := Q;

until HugeInt_Comp(a, b) < 0;

if A_IsNeg xor B_IsNeg then HugeInt_Min(R);

end;{ else }

end;{ HugeInt_Div }

procedure HugeInt_DivMod100(var a: HugeInt; var R: Integer);

{ This works on positive numbers only

256-Based division: R := a mod 100; a:= a div 100; }

var

Q: HugeInt;

S: Integer;

begin

R := 0; FillChar(Q, SizeOf(Q), 0);

S := HugeInt_HCD(a);

repeat

r := 256*R + a[S];

HugeInt_SHL(Q, 1);

Q[0] := R div 100;

R := R mod 100;

Dec(S);

until S < 0;

Move(Q, a, SizeOf(Q));

end;{ HugeInt_DivMod100 }

procedure HugeInt_Div(a, b: HugeInt; var R: HugeInt);

begin

HugeInt_DivMod(a, b, R);

end;{ HugeInt_Div }

procedure HugeInt_Mod(a, b: HugeInt; var R: HugeInt);

begin

HugeInt_DivMod(a, b, R);

Move(a, R, SizeOf(HugeInt));

end;{ HugeInt_Mod }

procedure HugeInt2String(a: HugeInt; var S: string);

function Str100(i: Integer): string;

begin

Str100 := Chr(i div 10 + Ord('0')) + Chr(i mod 10 + Ord('0'));

end;{ Str100 }

var

R: Integer;

Is_Neg: Boolean;

begin

S := '';

Is_Neg := HugeInt_IsNeg(a);

if Is_Neg then HugeInt_Min(a);

repeat

HugeInt_DivMod100(a, R);

Insert(Str100(R), S, 1);

until HugeInt_Zero(a) or (Length(S) = 254);

while (Length(S) > 1) and (S[1] = '0') do Delete(S, 1, 1);

if Is_Neg then Insert('-', S, 1);

end;{ HugeInt2String }

procedure String_DivMod256(var S: string; var R: Integer);

{ This works on Positive numbers Only

10(00)-based division: R := S mod 256; S := S div 256 }

var Q: string;

begin

FillChar(Q, SizeOf(Q), 0);

R := 0;

while S <> '' do

begin

R := 10*R + Ord(S[1]) - Ord('0'); Delete(S, 1, 1);

Q := Q + Chr(R div 256 + Ord('0'));

R := R mod 256;

end;{ while }

while (Q <> '') and (Q[1] = '0') do Delete(Q, 1, 1);

S := Q;

end;{ String_DivMod256 }

procedure String2HugeInt(AString: string; var a: HugeInt);

var

i, h: Integer;

Is_Neg: Boolean;

begin

if AString = '' then AString := '0';

Is_Neg := AString[1] = '-';

if Is_Neg then Delete(Astring, 1, 1);

i := 0;

while (AString <> '') and (i <= HugeIntMSB) do

begin

String_DivMod256(AString, h);

a[i] := h;

Inc(i);

end;{ while }

if Is_Neg then HugeInt_Min(a);

end;{ String2HugeInt }

procedure Integer2HugeInt(AInteger: Integer; var a: HugeInt);

var Is_Neg: Boolean;

begin

Is_Neg := AInteger < 0;

if Is_Neg then AInteger := -AInteger;

FillChar(a, SizeOf(HugeInt), 0);

Move(AInteger, a, SizeOf(Integer));

if Is_Neg then HugeInt_Min(a);

end;{ Integer2HugeInt }

end.

{ This code came from Lloyd's help file! }