Cipolla's algorithm
From Rosetta Code
Cipolla's algorithm is a draft programming task. It is not yet considered ready to be promoted as a complete task, for reasons that should be found in its talk page.
Cipolla's algorithm
Solve x² ≡ n (mod p)
In computational number theory, Cipolla's algorithm is a technique for solving an equation of the form x² ≡ n (mod p), where p is an odd prime and x ,n ∊ Fp = {0, 1, ... p-1}.
To apply the algorithm we need the Legendre symbol, and arithmetic in Fp².
Legendre symbol
- The Legendre symbol ( a | p) denotes the value of a ^ ((p-1)/2) (mod p)
- (a | p) ≡ 1 if a is a square (mod p)
- (a | p) ≡ -1 if a is not a square (mod p)
- (a | p) ≡ 0 is a ≡ 0
Arithmetic in Fp²
Let ω a symbol such as ω² is a member of Fp and not a square, x and y members of Fp. The set Fp² is defined as {x + ω y }. The subset { x + 0 ω} of Fp² is Fp. Fp² is somewhat equivalent to the field of complex number, with ω analoguous to i, and i² = -1 . Remembering that all operations are modulo p, addition, multiplication and exponentiation in Fp² are defined as :
- (x1 + ω y1) + (x2 + ω y2) := (x1 + x2 + ω (y1 + y2))
- (x1 + ω y1) * (x2 + ω y2) := (x1*x2 + y1*y2*ω²) + ω (x1*y2 + x2*y1)
- (0 + ω) * (0 + ω) := (ω² + 0 ω) ≡ ω² in Fp
- (x1 + ω y1) ^ n := (x + ω y) * (x + ω y) * ... ( n times) (1)
Algorithm pseudo-code
- Input : p an odd prime, and n ≠ 0 in Fp
- Step 0. Check that n is indeed a square : (n | p) must be ≡ 1
- Step 1. Find, by trial and error, an a > 0 such as (a² - n) is not a square : (a²-n | p) must be ≡ -1.
- Step 2. Let ω² = a² - n. Compute, in Fp2 : (a + ω) ^ ((p + 1)/2) (mod p)
To compute this step, use a pair of numbers, initially [a,1], and use repeated "multiplication" which is defined such that [c,d] times [e,f] is (mod p) [ c*c + ω²*f*f, d*e + c*f ].
- Step 3. Check that the result is ≡ x + 0 * ω in Fp2, that is x in Fp.
- Step 4. Output the two positive solutions, x and p - x (mod p).
- Step 5. Check that x * x ≡ n (mod p)
Example from Wikipedia
n := 10 , p := 13
Legendre(10,13) → 1 // 10 is indeed a square
a := 2 // try
ω² := a*a - 10 // ≡ 7 ≡ -6
Legendre (ω² , 13) → -1 // ok - not square
(2 + ω) ^ 7 → 6 + 0 ω // by modular exponentiation (1)
// 6 and (13 - 6) = 7 are solutions
(6 * 6) % 13 → 10 // = n . Checked.
Task
Implement the above.
Find solutions (if any) for
- n = 10 p = 13
- n = 56 p = 101
- n = 8218 p = 10007
- n = 8219 p = 10007
- n = 331575 p = 1000003
Extra credit
- n 665165880 p 1000000007
- n 881398088036 p 1000000000039
- n = 34035243914635549601583369544560650254325084643201 p = 10^50 + 151
See also:
Contents
C#[edit]
using System;
using System.Numerics;
namespace CipollaAlgorithm {
class Program {
static readonly BigInteger BIG = BigInteger.Pow(10, 50) + 151;
private static Tuple<BigInteger, BigInteger, bool> C(string ns, string ps) {
BigInteger n = BigInteger.Parse(ns);
BigInteger p = ps.Length > 0 ? BigInteger.Parse(ps) : BIG;
// Legendre symbol. Returns 1, 0, or p-1
BigInteger ls(BigInteger a0) => BigInteger.ModPow(a0, (p - 1) / 2, p);
// Step 0: validate arguments
if (ls(n) != 1) {
return new Tuple<BigInteger, BigInteger, bool>(0, 0, false);
}
// Step 1: Find a, omega2
BigInteger a = 0;
BigInteger omega2;
while (true) {
omega2 = (a * a + p - n) % p;
if (ls(omega2) == p - 1) {
break;
}
a += 1;
}
// Multiplication in Fp2
BigInteger finalOmega = omega2;
Tuple<BigInteger, BigInteger> mul(Tuple<BigInteger, BigInteger> aa, Tuple<BigInteger, BigInteger> bb) {
return new Tuple<BigInteger, BigInteger>(
(aa.Item1 * bb.Item1 + aa.Item2 * bb.Item2 * finalOmega) % p,
(aa.Item1 * bb.Item2 + bb.Item1 * aa.Item2) % p
);
}
// Step 2: Compute power
Tuple<BigInteger, BigInteger> r = new Tuple<BigInteger, BigInteger>(1, 0);
Tuple<BigInteger, BigInteger> s = new Tuple<BigInteger, BigInteger>(a, 1);
BigInteger nn = ((p + 1) >> 1) % p;
while (nn > 0) {
if ((nn & 1) == 1) {
r = mul(r, s);
}
s = mul(s, s);
nn >>= 1;
}
// Step 3: Check x in Fp
if (r.Item2 != 0) {
return new Tuple<BigInteger, BigInteger, bool>(0, 0, false);
}
// Step 5: Check x * x = n
if (r.Item1 * r.Item1 % p != n) {
return new Tuple<BigInteger, BigInteger, bool>(0, 0, false);
}
// Step 4: Solutions
return new Tuple<BigInteger, BigInteger, bool>(r.Item1, p - r.Item1, true);
}
static void Main(string[] args) {
Console.WriteLine(C("10", "13"));
Console.WriteLine(C("56", "101"));
Console.WriteLine(C("8218", "10007"));
Console.WriteLine(C("8219", "10007"));
Console.WriteLine(C("331575", "1000003"));
Console.WriteLine(C("665165880", "1000000007"));
Console.WriteLine(C("881398088036", "1000000000039"));
Console.WriteLine(C("34035243914635549601583369544560650254325084643201", ""));
}
}
}
- Output:
(6, 7, True) (37, 64, True) (9872, 135, True) (0, 0, False) (855842, 144161, True) (475131702, 524868305, True) (791399408049, 208600591990, True) (82563118828090362261378993957450213573687113690751, 17436881171909637738621006042549786426312886309400, True)
D[edit]
import std.bigint;
import std.stdio;
import std.typecons;
enum BIGZERO = BigInt(0);
/// https://en.wikipedia.org/wiki/Modular_exponentiation#Right-to-left_binary_method
BigInt modPow(BigInt b, BigInt e, BigInt n) {
if (n == 1) return BIGZERO;
BigInt result = 1;
b = b % n;
while (e > 0) {
if (e % 2 == 1) {
result = (result * b) % n;
}
e >>= 1;
b = (b*b) % n;
}
return result;
}
alias Point = Tuple!(BigInt, "x", BigInt, "y");
alias Triple = Tuple!(BigInt, "x", BigInt, "y", bool, "b");
Triple c(string ns, string ps) {
auto n = BigInt(ns);
BigInt p;
if (ps.length > 0) {
p = BigInt(ps);
} else {
p = BigInt(10)^^50 + 151;
}
// Legendre symbol, returns 1, 0 or p - 1
auto ls = (BigInt a) => modPow(a, (p-1)/2, p);
// Step 0, validate arguments
if (ls(n) != 1) return Triple(BIGZERO, BIGZERO, false);
// Step 1, find a, omega2
auto a = BIGZERO;
BigInt omega2;
while (true) {
omega2 = (a * a + p - n) % p;
if (ls(omega2) == p-1) break;
a++;
}
// multiplication in Fp2
auto mul = (Point aa, Point bb) => Point(
(aa.x * bb.x + aa.y * bb.y * omega2) % p,
(aa.x * bb.y + bb.x * aa.y) % p
);
// Step 2, compute power
auto r = Point(BigInt(1), BIGZERO);
auto s = Point(a, BigInt(1));
auto nn = ((p+1) >> 1) % p;
while (nn > 0) {
if ((nn & 1) == 1) r = mul(r, s);
s = mul(s, s);
nn >>= 1;
}
// Step 3, check x in Fp
if (r.y != 0) return Triple(BIGZERO, BIGZERO, false);
// Step 5, check x * x = n
if (r.x*r.x%p!=n) return Triple(BIGZERO, BIGZERO, false);
// Step 4, solutions
return Triple(r.x, p-r.x, true);
}
void main() {
writeln(c("10", "13"));
writeln(c("56", "101"));
writeln(c("8218", "10007"));
writeln(c("8219", "10007"));
writeln(c("331575", "1000003"));
writeln(c("665165880", "1000000007"));
writeln(c("881398088036", "1000000000039"));
writeln(c("34035243914635549601583369544560650254325084643201", ""));
}
- Output:
Tuple!(BigInt, "x", BigInt, "y", bool, "b")(6, 7, true) Tuple!(BigInt, "x", BigInt, "y", bool, "b")(37, 64, true) Tuple!(BigInt, "x", BigInt, "y", bool, "b")(9872, 135, true) Tuple!(BigInt, "x", BigInt, "y", bool, "b")(0, 0, false) Tuple!(BigInt, "x", BigInt, "y", bool, "b")(855842, 144161, true) Tuple!(BigInt, "x", BigInt, "y", bool, "b")(475131702, 524868305, true) Tuple!(BigInt, "x", BigInt, "y", bool, "b")(791399408049, 208600591990, true) Tuple!(BigInt, "x", BigInt, "y", bool, "b")(82563118828090362261378993957450213573687113690751, 17436881171909637738621006042549786426312886309400, true)
EchoLisp[edit]
(lib 'struct)
(lib 'types)
(lib 'bigint)
;; test equality mod p
(define-syntax-rule (mod= a b p)
(zero? (% (- a b) p)))
(define (Legendre a p)
(powmod a (/ (1- p) 2) p))
;; Arithmetic in Fp²
(struct Fp² ( x y ))
;; a + b
(define (Fp²-add Fp²:a Fp²:b p ω2)
(Fp² (% (+ a.x b.x) p) (% (+ a.y b.y) p)))
;; a * b
(define (Fp²-mul Fp²:a Fp²:b p ω2)
(Fp² (% (+ (* a.x b.x) (* ω2 a.y b.y)) p) (% (+ (* a.x b.y) (* a.y b.x)) p)))
;; a * a
(define (Fp²-square Fp²:a p ω2)
(Fp² (% (+ (* a.x a.x) (* ω2 a.y a.y)) p) (% (* 2 a.x a.y) p)))
;; a ^ n
(define (Fp²-pow Fp²:a n p ω2)
(cond
((= 0 n) (Fp² 1 0))
((= 1 n) (Fp² a.x a.y))
((= 2 n) (Fp²-mul a a p ω2))
((even? n) (Fp²-square (Fp²-pow a (/ n 2) p ω2) p ω2))
(else (Fp²-mul a (Fp²-pow a (1- n) p ω2) p ω2))))
;; x^2 ≡ n (mod p) ?
(define (Cipolla n p)
;; check n is a square
(unless (= 1 (Legendre n p)) (error "not a square (mod p)" (list n p)))
;; iterate until suitable 'a' found
(define a
(for ((t (in-range 2 p))) ;; t = tentative a
#:break (= (1- p) (Legendre (- (* t t) n) p)) => t
))
(define ω2 (- (* a a) n))
;; (writeln 'a-> a 'ω2-> ω2 'ω-> 'ω)
;; (Fp² a 1) = a + ω
(define r (Fp²-pow (Fp² a 1) (/ (1+ p) 2) p ω2))
;; (writeln 'r r)
(define x (Fp²-x r))
(assert (zero? (Fp²-y r))) ;; hope that ω has vanished
(assert (mod= n (* x x) p)) ;; checking the result
(printf "Roots of %d are (%d,%d) (mod %d)" n x (% (- p x) p) p))
- Output:
(Cipolla 10 13) Roots of 10 are (6,7) (mod 13) (% (* 6 6) 13) → 10 ;; checking (Cipolla 56 101) Roots of 56 are (37,64) (mod 101) (Cipolla 8218 10007) Roots of 8218 are (9872,135) (mod 10007) Cipolla 8219 10007) ❌ error: not a square (mod p) (8219 10007) (Cipolla 331575 1000003) Roots of 331575 are (855842,144161) (mod 1000003) (% ( * 855842 855842) 1000003) → 331575
FreeBASIC[edit]
LongInt version[edit]
Had a close look at the EchoLisp code for step 2. Used the FreeBASIC code from the Miller-Rabin task for prime testing.
' version 08-04-2017
' compile with: fbc -s console
' maximum for p is 17 digits to be on the save side
' TRUE/FALSE are built-in constants since FreeBASIC 1.04
' But we have to define them for older versions.
#Ifndef TRUE
#Define FALSE 0
#Define TRUE Not FALSE
#EndIf
Type fp2
x As LongInt
y As LongInt
End Type
Function mul_mod(a As ULongInt, b As ULongInt, modulus As ULongInt) As ULongInt
' returns a * b mod modulus
Dim As ULongInt x, y = a mod modulus
While b > 0
If (b And 1) = 1 Then
x = (x + y) Mod modulus
End If
y = (y Shl 1) Mod modulus
b = b Shr 1
Wend
Return x
End Function
Function pow_mod(b As ULongInt, power As ULongInt, modulus As ULongInt) As ULongInt
' returns b ^ power mod modulus
Dim As ULongInt x = 1
While power > 0
If (power And 1) = 1 Then
' x = (x * b) Mod modulus
x = mul_mod(x, b, modulus)
End If
' b = (b * b) Mod modulus
b = mul_mod(b, b, modulus)
power = power Shr 1
Wend
Return x
End Function
Function Isprime(n As ULongInt, k As Long) As Long
' miller-rabin prime test
If n > 9223372036854775808ull Then ' limit 2^63, pow_mod/mul_mod can't handle bigger numbers
Print "number is to big, program will end"
Sleep
End
End If
' 2 is a prime, if n is smaller then 2 or n is even then n = composite
If n = 2 Then Return TRUE
If (n < 2) OrElse ((n And 1) = 0) Then Return FALSE
Dim As ULongInt a, x, n_one = n - 1, d = n_one
Dim As UInteger s
While (d And 1) = 0
d = d Shr 1
s = s + 1
Wend
While k > 0
k = k - 1
a = Int(Rnd * (n -2)) +2 ' 2 <= a < n
x = pow_mod(a, d, n)
If (x = 1) Or (x = n_one) Then Continue While
For r As Integer = 1 To s -1
x = pow_mod(x, 2, n)
If x = 1 Then Return FALSE
If x = n_one Then Continue While
Next
If x <> n_one Then Return FALSE
Wend
Return TRUE
End Function
Function legendre_symbol (a As LongInt, p As LongInt) As LongInt
Dim As LongInt x = pow_mod(a, ((p -1) \ 2), p)
If p -1 = x Then
Return x - p
Else
Return x
End If
End Function
Function fp2mul(a As fp2, b As fp2, p As LongInt, w2 As LongInt) As fp2
Dim As fp2 answer
Dim As ULongInt tmp1, tmp2
' needs to be broken down in smaller steps to avoid overflow
' answer.x = (a.x * b.x + a.y * b.y * w2) Mod p
' answer.y = (a.x * b.y + a.y * b.x) Mod p
tmp1 = mul_mod(a.x, b.x, p)
tmp2 = mul_mod(a.y, b.y, p)
tmp2 = mul_mod(tmp2, w2, p)
answer.x = (tmp1 + tmp2) Mod p
tmp1 = mul_mod(a.x, b.y, p)
tmp2 = mul_mod(a.y, b.x, p)
answer.y = (tmp1 + tmp2) Mod p
Return answer
End Function
Function fp2square(a As fp2, p As LongInt, w2 As LongInt) As fp2
Return fp2mul(a, a, p, w2)
End Function
Function fp2pow(a As fp2, n As LongInt, p As LongInt, w2 As LongInt) As fp2
If n = 0 Then Return Type (1, 0)
If n = 1 Then Return a
If n = 2 Then Return fp2square(a, p, w2)
If (n And 1) = 0 Then
Return fp2square(fp2pow(a, n \ 2, p, w2), p , w2)
Else
Return fp2mul(a, fp2pow(a, n -1, p, w2), p, w2)
End If
End Function
' ------=< MAIN >=------
Data 10, 13, 56, 101, 8218, 10007,8219, 10007
Data 331575, 1000003, 665165880, 1000000007
Data 881398088036, 1000000000039
Randomize Timer
Dim As LongInt n, p, a, w2
Dim As LongInt i, x1, x2
Dim As fp2 answer
For i = 1 To 7
Read n, p
Print "Find solution for n =";n ; " and p =";p
If p = 2 OrElse Isprime(p,15) = FALSE Then
Print "No solution, p is not a odd prime"
Continue For
End If
' p is checked and is a odd prime
If legendre_symbol(n, p) <> 1 Then
Print n; " is not a square in F";Str(p)
Continue For
End If
Do
Do
a = Rnd * (p -2) +2
w2 = a * a - n
Loop Until legendre_symbol(w2, p) = -1
answer = Type(a, 1)
answer = fp2pow(answer, (p +1) \ 2, p, w2)
If answer.y <> 0 Then Continue Do
x1 = answer.x : x2 = p - x1
If mul_mod(x1, x1, p) = n AndAlso mul_mod(x2, x2, p) = n Then
Print "Solution found: x1 ="; x1; ", "; "x2 ="; x2
Exit Do
End If
Loop ' loop until solution is found
Next
' empty keyboard buffer
While Inkey <> "" : Wend
Print : Print "hit any key to end program"
Sleep
End
- Output:
Find solution for n = 10 and p = 13 Solution found: x1 = 7, x2 = 6 Find solution for n = 56 and p = 101 Solution found: x1 = 37, x2 = 64 Find solution for n = 8218 and p = 10007 Solution found: x1 = 9872, x2 = 135 Find solution for n = 8219 and p = 10007 8219 is not a square in F10007 Find solution for n = 331575 and p = 1000003 Solution found: x1 = 144161, x2 = 855842 Find solution for n = 665165880 and p = 1000000007 Solution found: x1 = 475131702, x2 = 524868305 Find solution for n = 881398088036 and p = 1000000000039 Solution found: x1 = 791399408049, x2 = 208600591990
GMP version[edit]
' version 12-04-2017
' compile with: fbc -s console
#Include Once "gmp.bi"
Type fp2
x As Mpz_ptr
y As Mpz_ptr
End Type
Data "10", "13"
Data "56", "101"
Data "8218", "10007"
Data "8219", "10007"
Data "331575", "1000003"
Data "665165880", "1000000007"
Data "881398088036", "1000000000039"
Data "34035243914635549601583369544560650254325084643201" ', 10^50 + 151
Function fp2mul(a As fp2, b As fp2, p As Mpz_ptr, w2 As Mpz_ptr) As fp2
Dim As fp2 r
r.x = Allocate(Len(__Mpz_struct)) : Mpz_init(r.x)
r.y = Allocate(Len(__Mpz_struct)) : Mpz_init(r.y)
Mpz_mul (r.x, a.y, b.y)
Mpz_mul (r.x, r.x, w2)
Mpz_addmul(r.x, a.x, b.x)
Mpz_mod (r.x, r.x, p)
Mpz_mul (r.y, a.x, b.y)
Mpz_addmul(r.y, a.y, b.x)
Mpz_mod (r.y, r.y, p)
Return r
End Function
Function fp2square(a As fp2, p As Mpz_ptr, w2 As Mpz_ptr) As fp2
Return fp2mul(a, a, p, w2)
End Function
Function fp2pow(a As fp2, n As Mpz_ptr, p As Mpz_ptr, w2 As Mpz_ptr) As fp2
If Mpz_cmp_ui(n, 0) = 0 Then
Mpz_set_ui(a.x, 1)
Mpz_set_ui(a.y, 0)
Return a
End If
If Mpz_cmp_ui(n, 1) = 0 Then Return a
If Mpz_cmp_ui(n, 2) = 0 Then Return fp2square(a, p, w2)
If Mpz_tstbit(n, 0) = 0 Then
Mpz_fdiv_q_2exp(n, n, 1) ' even
Return fp2square(fp2pow(a, n, p, w2), p, w2)
Else
Mpz_sub_ui(n, n, 1) ' odd
Return fp2mul(a, fp2pow(a, n, p, w2), p, w2)
End If
End Function
' ------=< MAIN >=------
Dim As Long i
Dim As ZString Ptr zstr
Dim As String n_str, p_str
Dim As Mpz_ptr a, n, p, p2, w2, x1, x2
a = Allocate(Len(__Mpz_struct)) : Mpz_init(a)
n = Allocate(Len(__Mpz_struct)) : Mpz_init(n)
p = Allocate(Len(__Mpz_struct)) : Mpz_init(p)
p2 = Allocate(Len(__Mpz_struct)) : Mpz_init(p2)
w2 = Allocate(Len(__Mpz_struct)) : Mpz_init(w2)
x1 = Allocate(Len(__Mpz_struct)) : Mpz_init(x1)
x2 = Allocate(Len(__Mpz_struct)) : Mpz_init(x2)
Dim As fp2 answer
answer.x = Allocate(Len(__Mpz_struct)) : Mpz_init(answer.x)
answer.y = Allocate(Len(__Mpz_struct)) : Mpz_init(answer.y)
For i = 1 To 8
Read n_str
Mpz_set_str(n, n_str, 10)
If i < 8 Then
Read p_str
Mpz_set_str(p, p_str, 10)
Else
p_str = "10^50 + 151" ' set up last n
Mpz_set_str(p, "1" + String(50, "0"), 10)
Mpz_add_ui(p, p, 151)
End If
Print "Find solution for n = "; n_str; " and p = "; p_str
If Mpz_tstbit(p, 0) = 0 OrElse Mpz_probab_prime_p(p, 20) = 0 Then
Print p_str; "is not a odd prime"
Continue For
End If
' p is checked and is a odd prime
' legendre symbol needs to be 1
If Mpz_legendre(n, p) <> 1 Then
Print n_str; " is not a square in F"; p_str
Continue For
End If
Mpz_set_ui(a, 1)
Do
Do
Do
Mpz_add_ui(a, a, 1)
Mpz_mul(w2, a, a)
Mpz_sub(w2, w2, n)
Loop Until Mpz_legendre(w2, p) = -1
Mpz_set(answer.x, a)
Mpz_set_ui(answer.y, 1)
Mpz_add_ui(p2, p, 1) ' p2 = p + 1
Mpz_fdiv_q_2exp(p2, p2, 1) ' p2 = p2 \ 2 (p2 shr 1)
answer = fp2pow(answer, p2, p, w2)
Loop Until Mpz_cmp_ui(answer.y, 0) = 0
Mpz_set(x1, answer.x)
Mpz_sub(x2, p, x1)
Mpz_powm_ui(a, x1, 2, p)
Mpz_powm_ui(p2, x2, 2, p)
If Mpz_cmp(a, n) = 0 AndAlso Mpz_cmp(p2, n) = 0 Then Exit Do
Loop
zstr = Mpz_get_str(0, 10, x1)
Print "Solution found: x1 = "; *zstr;
zstr = Mpz_get_str(0, 10, x2)
Print ", x2 = "; *zstr
Next
Mpz_clear(x1) : Mpz_clear(p2) : Mpz_clear(p) : Mpz_clear(a) : Mpz_clear(n)
Mpz_clear(x2) : Mpz_clear(w2) : Mpz_clear(answer.x) : Mpz_clear(answer.y)
' empty keyboard buffer
While Inkey <> "" : Wend
Print : Print "hit any key to end program"
Sleep
End
- Output:
Find solution for n = 10 and p = 13 Solution found: x1 = 6, x2 = 7 Find solution for n = 56 and p = 101 Solution found: x1 = 37, x2 = 64 Find solution for n = 8218 and p = 10007 Solution found: x1 = 9872, x2 = 135 Find solution for n = 8219 and p = 10007 8219 is not a square in F10007 Find solution for n = 331575 and p = 1000003 Solution found: x1 = 855842, x2 = 144161 Find solution for n = 665165880 and p = 1000000007 Solution found: x1 = 524868305, x2 = 475131702 Find solution for n = 881398088036 and p = 1000000000039 Solution found: x1 = 208600591990, x2 = 791399408049 Find solution for n = 34035243914635549601583369544560650254325084643201 and p = 10^50 + 151 Solution found: x1 = 17436881171909637738621006042549786426312886309400, x2 = 82563118828090362261378993957450213573687113690751
Go[edit]
int[edit]
Implementation following the pseudocode in the task description.
package main
import "fmt"
func c(n, p int) (R1, R2 int, ok bool) {
// a^e mod p
powModP := func(a, e int) int {
s := 1
for ; e > 0; e-- {
s = s * a % p
}
return s
}
// Legendre symbol, returns 1, 0, or -1 mod p -- that's 1, 0, or p-1.
ls := func(a int) int {
return powModP(a, (p-1)/2)
}
// Step 0, validate arguments
if ls(n) != 1 {
return
}
// Step 1, find a, ω2
var a, ω2 int
for a = 0; ; a++ {
// integer % in Go uses T-division, add p to keep the result positive
ω2 = (a*a + p - n) % p
if ls(ω2) == p-1 {
break
}
}
// muliplication in fp2
type point struct{ x, y int }
mul := func(a, b point) point {
return point{(a.x*b.x + a.y*b.y*ω2) % p, (a.x*b.y + b.x*a.y) % p}
}
// Step2, compute power
r := point{1, 0}
s := point{a, 1}
for n := (p + 1) >> 1 % p; n > 0; n >>= 1 {
if n&1 == 1 {
r = mul(r, s)
}
s = mul(s, s)
}
// Step3, check x in Fp
if r.y != 0 {
return
}
// Step5, check x*x=n
if r.x*r.x%p != n {
return
}
// Step4, solutions
return r.x, p - r.x, true
}
func main() {
fmt.Println(c(10, 13))
fmt.Println(c(56, 101))
fmt.Println(c(8218, 10007))
fmt.Println(c(8219, 10007))
fmt.Println(c(331575, 1000003))
}
- Output:
6 7 true 37 64 true 9872 135 true 0 0 false 855842 144161 true
big.Int[edit]
Extra credit:
package main
import (
"fmt"
"math/big"
)
func c(n, p big.Int) (R1, R2 big.Int, ok bool) {
if big.Jacobi(&n, &p) != 1 {
return
}
var one, a, ω2 big.Int
one.SetInt64(1)
for ; ; a.Add(&a, &one) {
// big.Int Mod uses Euclidean division, result is always >= 0
ω2.Mod(ω2.Sub(ω2.Mul(&a, &a), &n), &p)
if big.Jacobi(&ω2, &p) == -1 {
break
}
}
type point struct{ x, y big.Int }
mul := func(a, b point) (z point) {
var w big.Int
z.x.Mod(z.x.Add(z.x.Mul(&a.x, &b.x), w.Mul(w.Mul(&a.y, &a.y), &ω2)), &p)
z.y.Mod(z.y.Add(z.y.Mul(&a.x, &b.y), w.Mul(&b.x, &a.y)), &p)
return
}
var r, s point
r.x.SetInt64(1)
s.x.Set(&a)
s.y.SetInt64(1)
var e big.Int
for e.Rsh(e.Add(&p, &one), 1); len(e.Bits()) > 0; e.Rsh(&e, 1) {
if e.Bit(0) == 1 {
r = mul(r, s)
}
s = mul(s, s)
}
R2.Sub(&p, &r.x)
return r.x, R2, true
}
func main() {
var n, p big.Int
n.SetInt64(665165880)
p.SetInt64(1000000007)
R1, R2, ok := c(n, p)
fmt.Println(&R1, &R2, ok)
n.SetInt64(881398088036)
p.SetInt64(1000000000039)
R1, R2, ok = c(n, p)
fmt.Println(&R1, &R2, ok)
n.SetString("34035243914635549601583369544560650254325084643201", 10)
p.SetString("100000000000000000000000000000000000000000000000151", 10)
R1, R2, ok = c(n, p)
fmt.Println(&R1)
fmt.Println(&R2)
}
- Output:
475131702 524868305 true 791399408049 208600591990 true 82563118828090362261378993957450213573687113690751 17436881171909637738621006042549786426312886309400
J[edit]
Based on the echolisp implementation:
leg=: dyad define
x (y&|)@^ (y-1)%2
)
mul2=: conjunction define
m| (*&{. + n**&{:), (+/ .* |.)
)
pow2=: conjunction define
:
if. 0=y do. 1 0
elseif. 1=y do. x
elseif. 2=y do. x (m mul2 n) x
elseif. 0=2|y do. (m mul2 n)~ x (m pow2 n) y%2
elseif. do. x (m mul2 n) x (m pow2 n) y-1
end.
)
cipolla=: dyad define
assert. 1=1 p: y [ 'y must be prime'
assert. 1= x leg y [ 'x must be square mod y'
a=.1
whilst. (0 ~:{: r) do. a=. a+1
while. 1>: leg&y@(x -~ *:) a do. a=.a+1 end.
w2=. y|(*:a) - x
r=. (a,1) (y pow2 w2) (y+1)%2
end.
if. x =&(y&|) *:{.r do.
y|(,-){.r
else.
smoutput 'got ',":~.y|(,-){.r
assert. 'not a valid square root'
end.
)
Task examples:
10 cipolla 13
6 7
56 cipolla 101
37 64
8218 cipolla 10007
9872 135
8219 cipolla 10007
|assertion failure: cipolla
| 1=x leg y['x must be square mod y'
331575 cipolla 1000003
855842 144161
665165880x cipolla 1000000007x
524868305 475131702
881398088036x cipolla 1000000000039x
208600591990 791399408049
34035243914635549601583369544560650254325084643201x cipolla (10^50x) + 151
17436881171909637738621006042549786426312886309400 82563118828090362261378993957450213573687113690751
Java[edit]
import java.math.BigInteger;
import java.util.function.BiFunction;
import java.util.function.Function;
public class CipollasAlgorithm {
private static final BigInteger BIG = BigInteger.TEN.pow(50).add(BigInteger.valueOf(151));
private static final BigInteger BIG_TWO = BigInteger.valueOf(2);
private static class Point {
BigInteger x;
BigInteger y;
Point(BigInteger x, BigInteger y) {
this.x = x;
this.y = y;
}
@Override
public String toString() {
return String.format("(%s, %s)", this.x, this.y);
}
}
private static class Triple {
BigInteger x;
BigInteger y;
boolean b;
Triple(BigInteger x, BigInteger y, boolean b) {
this.x = x;
this.y = y;
this.b = b;
}
@Override
public String toString() {
return String.format("(%s, %s, %s)", this.x, this.y, this.b);
}
}
private static Triple c(String ns, String ps) {
BigInteger n = new BigInteger(ns);
BigInteger p = !ps.isEmpty() ? new BigInteger(ps) : BIG;
// Legendre symbol, returns 1, 0 or p - 1
Function<BigInteger, BigInteger> ls = (BigInteger a)
-> a.modPow(p.subtract(BigInteger.ONE).divide(BIG_TWO), p);
// Step 0, validate arguments
if (!ls.apply(n).equals(BigInteger.ONE)) {
return new Triple(BigInteger.ZERO, BigInteger.ZERO, false);
}
// Step 1, find a, omega2
BigInteger a = BigInteger.ZERO;
BigInteger omega2;
while (true) {
omega2 = a.multiply(a).add(p).subtract(n).mod(p);
if (ls.apply(omega2).equals(p.subtract(BigInteger.ONE))) {
break;
}
a = a.add(BigInteger.ONE);
}
// multiplication in Fp2
BigInteger finalOmega = omega2;
BiFunction<Point, Point, Point> mul = (Point aa, Point bb) -> new Point(
aa.x.multiply(bb.x).add(aa.y.multiply(bb.y).multiply(finalOmega)).mod(p),
aa.x.multiply(bb.y).add(bb.x.multiply(aa.y)).mod(p)
);
// Step 2, compute power
Point r = new Point(BigInteger.ONE, BigInteger.ZERO);
Point s = new Point(a, BigInteger.ONE);
BigInteger nn = p.add(BigInteger.ONE).shiftRight(1).mod(p);
while (nn.compareTo(BigInteger.ZERO) > 0) {
if (nn.and(BigInteger.ONE).equals(BigInteger.ONE)) {
r = mul.apply(r, s);
}
s = mul.apply(s, s);
nn = nn.shiftRight(1);
}
// Step 3, check x in Fp
if (!r.y.equals(BigInteger.ZERO)) {
return new Triple(BigInteger.ZERO, BigInteger.ZERO, false);
}
// Step 5, check x * x = n
if (!r.x.multiply(r.x).mod(p).equals(n)) {
return new Triple(BigInteger.ZERO, BigInteger.ZERO, false);
}
// Step 4, solutions
return new Triple(r.x, p.subtract(r.x), true);
}
public static void main(String[] args) {
System.out.println(c("10", "13"));
System.out.println(c("56", "101"));
System.out.println(c("8218", "10007"));
System.out.println(c("8219", "10007"));
System.out.println(c("331575", "1000003"));
System.out.println(c("665165880", "1000000007"));
System.out.println(c("881398088036", "1000000000039"));
System.out.println(c("34035243914635549601583369544560650254325084643201", ""));
}
}
- Output:
(6, 7, true) (37, 64, true) (9872, 135, true) (0, 0, false) (855842, 144161, true) (475131702, 524868305, true) (791399408049, 208600591990, true) (82563118828090362261378993957450213573687113690751, 17436881171909637738621006042549786426312886309400, true)
Kotlin[edit]
// version 1.2.0
import java.math.BigInteger
class Point(val x: BigInteger, val y: BigInteger)
val bigZero = BigInteger.ZERO
val bigOne = BigInteger.ONE
val bigTwo = BigInteger.valueOf(2L)
val bigBig = BigInteger.TEN.pow(50) + BigInteger.valueOf(151L)
fun c(ns: String, ps: String): Triple<BigInteger, BigInteger, Boolean> {
val n = BigInteger(ns)
val p = if (!ps.isEmpty()) BigInteger(ps) else bigBig
// Legendre symbol, returns 1, 0 or p - 1
fun ls(a: BigInteger) = a.modPow((p - bigOne) / bigTwo, p)
// Step 0, validate arguments
if (ls(n) != bigOne) return Triple(bigZero, bigZero, false)
// Step 1, find a, omega2
var a = bigZero
var omega2: BigInteger
while (true) {
omega2 = (a * a + p - n) % p
if (ls(omega2) == p - bigOne) break
a++
}
// multiplication in Fp2
fun mul(aa: Point, bb: Point) =
Point(
(aa.x * bb.x + aa.y * bb.y * omega2) % p,
(aa.x * bb.y + bb.x * aa.y) % p
)
// Step 2, compute power
var r = Point(bigOne, bigZero)
var s = Point(a, bigOne)
var nn = ((p + bigOne) shr 1) % p
while (nn > bigZero) {
if ((nn and bigOne) == bigOne) r = mul(r, s)
s = mul(s, s)
nn = nn shr 1
}
// Step 3, check x in Fp
if (r.y != bigZero) return Triple(bigZero, bigZero, false)
// Step 5, check x * x = n
if (r.x * r.x % p != n) return Triple(bigZero, bigZero, false)
// Step 4, solutions
return Triple(r.x, p - r.x, true)
}
fun main(args: Array<String>) {
println(c("10", "13"))
println(c("56", "101"))
println(c("8218", "10007"))
println(c("8219", "10007"))
println(c("331575", "1000003"))
println(c("665165880", "1000000007"))
println(c("881398088036", "1000000000039"))
println(c("34035243914635549601583369544560650254325084643201", ""))
}
- Output:
(6, 7, true) (37, 64, true) (9872, 135, true) (0, 0, false) (855842, 144161, true) (475131702, 524868305, true) (791399408049, 208600591990, true) (82563118828090362261378993957450213573687113690751, 17436881171909637738621006042549786426312886309400, true)
Perl[edit]
use bigint;
use ntheory qw(is_prime);
sub Legendre {
my($n,$p) = @_;
return -1 unless $p != 2 && is_prime($p);
my $x = ($n->as_int())->bmodpow(int(($p-1)/2), $p); # $n coerced to BigInt
if ($x==0) { return 0 }
elsif ($x==1) { return 1 }
else { return -1 }
}
sub Cipolla {
my($n, $p) = @_;
return undef if Legendre($n,$p) != 1;
my $w2;
my $a = 0;
$a++ until Legendre(($w2 = ($a**2 - $n) % $p), $p) < 0;
my %r = ( x=> 1, y=> 0 );
my %s = ( x=> $a, y=> 1 );
my $i = $p + 1;
while (1 <= ($i >>= 1)) {
%r = ( x => (($r{x} * $s{x} + $r{y} * $s{y} * $w2) % $p),
y => (($r{x} * $s{y} + $s{x} * $r{y}) % $p)
) if $i % 2;
%s = ( x => (($s{x} * $s{x} + $s{y} * $s{y} * $w2) % $p),
y => (($s{x} * $s{y} + $s{x} * $s{y}) % $p)
)
}
$r{y} ? undef : $r{x}
}
my @tests = (
(10, 13),
(56, 101),
(8218, 10007),
(8219, 10007),
(331575, 1000003),
(665165880, 1000000007),
(881398088036, 1000000000039),
);
while (@tests) {
$n = shift @tests;
$p = shift @tests;
my $r = Cipolla($n, $p);
$r ? printf "Roots of %d are (%d, %d) mod %d\n", $n, $r, $p-$r, $p
: print "No solution for ($n, $p)\n"
}
- Output:
Roots of 10 are (6, 7) mod 13 Roots of 56 are (37, 64) mod 101 Roots of 8218 are (9872, 135) mod 10007 No solution for (8219, 10007) Roots of 331575 are (855842, 144161) mod 1000003 Roots of 665165880 are (475131702, 524868305) mod 1000000007 Roots of 881398088036 are (791399408049, 208600591990) mod 1000000000039
Perl 6[edit]
# Legendre operator (𝑛│𝑝)
sub infix:<│> (Int \𝑛, Int \𝑝 where 𝑝.is-prime && (𝑝 != 2)) {
given 𝑛.expmod( (𝑝-1) div 2, 𝑝 ) {
when 0 { 0 }
when 1 { 1 }
default { -1 }
}
}
# a coordinate in a Field of p elements
class Fp {
has Int $.x;
has Int $.y;
}
sub cipolla ( Int \𝑛, Int \𝑝 ) {
note "Invalid parameters ({𝑛}, {𝑝})"
and return Nil if (𝑛│𝑝) != 1;
my $ω2;
my $a = 0;
loop {
last if ($ω2 = ($a² - 𝑛) % 𝑝)│𝑝 < 0;
$a++;
}
# define a local multiply operator for Field coordinates
multi sub infix:<*> ( Fp $a, Fp $b ){
Fp.new: :x(($a.x * $b.x + $a.y * $b.y * $ω2) % 𝑝),
:y(($a.x * $b.y + $b.x * $a.y) % 𝑝)
}
my $r = Fp.new: :x(1), :y(0);
my $s = Fp.new: :x($a), :y(1);
for (𝑝+1) +> 1, * +> 1 ... 1 {
$r *= $s if $_ % 2;
$s *= $s;
}
return Nil if $r.y;
$r.x;
}
my @tests = (
(10, 13),
(56, 101),
(8218, 10007),
(8219, 10007),
(331575, 1000003),
(665165880, 1000000007),
(881398088036, 1000000000039),
(34035243914635549601583369544560650254325084643201,
100000000000000000000000000000000000000000000000151)
);
for @tests -> ($n, $p) {
my $r = cipolla($n, $p);
say $r ?? "Roots of $n are ($r, {$p-$r}) mod $p"
!! "No solution for ($n, $p)"
}
- Output:
Roots of 10 are (6, 7) mod 13 Roots of 56 are (37, 64) mod 101 Roots of 8218 are (9872, 135) mod 10007 Invalid parameters (8219, 10007) No solution for (8219, 10007) Roots of 331575 are (855842, 144161) mod 1000003 Roots of 665165880 are (475131702, 524868305) mod 1000000007 Roots of 881398088036 are (791399408049, 208600591990) mod 1000000000039 Roots of 34035243914635549601583369544560650254325084643201 are (82563118828090362261378993957450213573687113690751, 17436881171909637738621006042549786426312886309400) mod 100000000000000000000000000000000000000000000000151
Phix[edit]
include bigatom.e
function legendre(bigatom a, p)
-- Legendre symbol, returns 1, 0 or p - 1
return ba_mod_exp(a,ba_idiv(ba_sub(p,1),2),p)
end function
function mul_point(sequence a, b, bigatom omega2, p)
object {xa,ya} = a, {xb,yb} = b
return {ba_mod(ba_add(ba_mul(xa,xb),ba_mul(ba_mul(ya,yb),omega2)),p),
ba_mod(ba_add(ba_mul(xa,yb),ba_mul(xb,ya)),p)}
end function
function cipolla(object no, po)
bigatom n = ba_new(no),
p = ba_new(po)
-- Step 0, validate arguments
if legendre(n,p)!=BA_ONE then return {"0","0","false"} end if
-- Step 1, find a, omega2
integer a = 0
bigatom omega2
while true do
omega2 = ba_mod(ba_add(a*a,ba_sub(p,n)),p)
if ba_compare(legendre(omega2,p),ba_sub(p,BA_ONE))=0 then exit end if
a += 1
end while
-- Step 2, compute power
sequence r = {1,0},
s = {a,1}
bigatom nn = ba_mod(ba_idiv(ba_add(p,1),2),p)
while ba_compare(nn,BA_ZERO)>0 do
if ba_mod(nn,2)=BA_ONE then
r = mul_point(r,s,omega2,p)
end if
s = mul_point(s,s,omega2,p)
nn = ba_idiv(nn,2)
end while
-- Step 3, check x in Fp
if ba_compare(r[2],BA_ZERO)!=0 then return {"0","0","false"} end if
-- Step 5, check x * x = n
if ba_compare(ba_mod_exp(r[1],2,p),n)!=0 then return {"0","0","false"} end if
-- Step 4, solutions
return {ba_sprint(r[1]), ba_sprint(ba_sub(p,r[1])), "true"}
end function
constant tests = {{10,13},
{56,101},
{8218,10007},
{8219,10007},
{331575,1000003},
{665165880,1000000007},
{"881398088036","1000000000039"},
{"34035243914635549601583369544560650254325084643201",
ba_sprint(ba_add(ba_power(10,50),151))}}
for i=1 to length(tests) do
object {n,p} = tests[i]
?{n,p,cipolla(n,p)}
end for
Obviously were you to use that in anger, you would probably rip out a few ba_sprint() and return false rather than "false", etc.
- Output:
{10,13,{"6","7","true"}}
{56,101,{"37","64","true"}}
{8218,10007,{"9872","135","true"}}
{8219,10007,{"0","0","false"}}
{331575,1000003,{"855842","144161","true"}}
{665165880,1000000007,{"475131702","524868305","true"}}
{"881398088036","1000000000039",{"791399408049","208600591990","true"}}
{"34035243914635549601583369544560650254325084643201","100000000000000000000000000000000000000000000000151",
{"82563118828090362261378993957450213573687113690751","17436881171909637738621006042549786426312886309400","true"}}
PicoLisp[edit]
# from @lib/rsa.l
(de **Mod (X Y N)
(let M 1
(loop
(when (bit? 1 Y)
(setq M (% (* M X) N)) )
(T (=0 (setq Y (>> 1 Y)))
M )
(setq X (% (* X X) N)) ) ) )
(de legendre (N P)
(**Mod N (/ (dec P) 2) P) )
(de mul ("A" B P W2)
(let (A (copy "A") B (copy B))
(set
"A"
(%
(+
(* (car A) (car B))
(* (cadr A) (cadr B) W2) )
P )
(cdr "A")
(%
(+
(* (car A) (cadr B))
(* (car B) (cadr A)) )
P ) ) ) )
(de ci (N P)
(and
(=1 (legendre N P))
(let
(A 0
W2 0
R NIL
S NIL )
(loop
(setq W2
(% (- (+ (* A A) P) N) P) )
(T (= (dec P) (legendre W2 P)))
(inc 'A) )
(setq R (list 1 0) S (list A 1))
(for
(N
(% (>> 1 (inc P)) P)
(> N 0)
(>> 1 N) )
(and (bit? 1 N) (mul R S P W2))
(mul S S P W2) )
(=0 (cadr R))
(=
N
(% (* (car R) (car R)) P) )
(list (car R) (- P (car R))) ) ) )
(println (ci 10 13))
(println (ci 56 101))
(println (ci 8218 10007))
(println (ci 8219 10007))
(println (ci 331575 1000003))
(println (ci 665165880 1000000007))
(println (ci 881398088036 1000000000039))
(println (ci 34035243914635549601583369544560650254325084643201 (+ (** 10 50) 151)))
- Output:
(6 7) (37 64) (9872 135) NIL (855842 144161) (475131702 524868305) (791399408049 208600591990) (82563118828090362261378993957450213573687113690751 17436881171909637738621006042549786426312886309400)
Python[edit]
#Converts n to base b as a list of integers between 0 and b-1
#Most-significant digit on the left
def convertToBase(n, b):
if(n < 2):
return [n];
temp = n;
ans = [];
while(temp != 0):
ans = [temp % b]+ ans;
temp /= b;
return ans;
#Takes integer n and odd prime p
#Returns both square roots of n modulo p as a pair (a,b)
#Returns () if no root
def cipolla(n,p):
n %= p
if(n == 0 or n == 1):
return (n,-n%p)
phi = p - 1
if(pow(n, phi/2, p) != 1):
return ()
if(p%4 == 3):
ans = pow(n,(p+1)/4,p)
return (ans,-ans%p)
aa = 0
for i in xrange(1,p):
temp = pow((i*i-n)%p,phi/2,p)
if(temp == phi):
aa = i
break;
exponent = convertToBase((p+1)/2,2)
def cipollaMult((a,b),(c,d),w,p):
return ((a*c+b*d*w)%p,(a*d+b*c)%p)
x1 = (aa,1)
x2 = cipollaMult(x1,x1,aa*aa-n,p)
for i in xrange(1,len(exponent)):
if(exponent[i] == 0):
x2 = cipollaMult(x2,x1,aa*aa-n,p)
x1 = cipollaMult(x1,x1,aa*aa-n,p)
else:
x1 = cipollaMult(x1,x2,aa*aa-n,p)
x2 = cipollaMult(x2,x2,aa*aa-n,p)
return (x1[0],-x1[0]%p)
print "Roots of 2 mod 7: " +str(cipolla(2,7))
print "Roots of 8218 mod 10007: " +str(cipolla(8218,10007))
print "Roots of 56 mod 101: " +str(cipolla(56,101))
print "Roots of 1 mod 11: " +str(cipolla(1,11))
print "Roots of 8219 mod 10007: " +str(cipolla(8219,10007))
- Output:
Roots of 2 mod 7: (4, 3) Roots of 8218 mod 10007: (9872, 135) Roots of 56 mod 101: (37, 64) Roots of 1 mod 11: (1, 10) Roots of 8219 mod 10007: ()
Racket[edit]
#lang racket
(require math/number-theory)
;; math/number-theory allows us to parameterize a "current-modulus"
;; which obviates the need for p to be passed around constantly
(define (Cipolla n p) (with-modulus p (mod-Cipolla n)))
(define (mod-Legendre a)
(modexpt a (/ (sub1 (current-modulus)) 2)))
;; Arithmetic in Fp²
(struct Fp² (x y))
(define-syntax-rule (Fp²-destruct* (a a.x a.y) ...)
(begin (match-define (Fp² a.x a.y) a) ...) )
;; a + b
(define (Fp²-add a b ω2)
(Fp²-destruct* (a a.x a.y) (b b.x b.y))
(Fp² (mod+ a.x b.x) (mod+ a.y b.y)))
;; a * b
(define (Fp²-mul a b ω2)
(Fp²-destruct* (a a.x a.y) (b b.x b.y))
(Fp² (mod+ (* a.x b.x) (* ω2 a.y b.y)) (mod+ (* a.x b.y) (* a.y b.x))))
;; a * a
(define (Fp²-square a ω2)
(Fp²-destruct* (a a.x a.y))
(Fp² (mod+ (sqr a.x) (* ω2 (sqr a.y))) (mod* 2 a.x a.y)))
;; a ^ n
(define (Fp²-pow a n ω2)
(Fp²-destruct* (a a.x a.y))
(cond
((= 0 n) (Fp² 1 0))
((= 1 n) a)
((= 2 n) (Fp²-mul a a ω2))
((even? n) (Fp²-square (Fp²-pow a (/ n 2) ω2) ω2))
(else (Fp²-mul a (Fp²-pow a (sub1 n) ω2) ω2))))
;; x^2 ≡ n (mod p) ?
(define (mod-Cipolla n)
;; check n is a square
(unless (= 1 (mod-Legendre n)) (error 'Cipolla "~a not a square (mod ~a)" n (current-modulus)))
;; iterate until suitable 'a' found
(define a (for/first ((t (in-range 2 (current-modulus))) ;; t = tentative a
#:when (= (sub1 (current-modulus))
(mod-Legendre (- (* t t) n))))
t))
(define ω2 (- (* a a) n))
;; (Fp² a 1) = a + ω
(define r (Fp²-pow (Fp² a 1) (/ (add1 (current-modulus)) 2) ω2))
(define x (Fp²-x r))
(unless (zero? (Fp²-y r)) (error 'Cipolla "ω has not vanished")) ;; hope that ω has vanished
(unless (mod= n (* x x)) (error 'Cipolla "result check failed")) ;; checking the result
(values x (mod- (current-modulus) x)))
(define (report-Cipolla n p)
(with-handlers ((exn:fail? (λ (x) (eprintf "Caught error: ~s~%" (exn-message x)))))
(define-values (r1 r2) (Cipolla n p))
(printf "Roots of ~a are (~a,~a) (mod ~a)~%" n r1 r2 p)))
(module+ test
(report-Cipolla 10 13)
(report-Cipolla 56 101)
(report-Cipolla 8218 10007)
(report-Cipolla 8219 10007)
(report-Cipolla 331575 1000003)
(report-Cipolla 665165880 1000000007)
(report-Cipolla 881398088036 1000000000039)
(report-Cipolla 34035243914635549601583369544560650254325084643201
100000000000000000000000000000000000000000000000151))
- Output:
Roots of 10 are (6,7) (mod 13) Roots of 56 are (37,64) (mod 101) Roots of 8218 are (9872,135) (mod 10007) Caught error: "Cipolla: 8219 not a square (mod 10007)" Roots of 331575 are (855842,144161) (mod 1000003) Roots of 665165880 are (524868305,475131702) (mod 1000000007) Roots of 881398088036 are (208600591990,791399408049) (mod 1000000000039) Roots of 34035243914635549601583369544560650254325084643201 are (17436881171909637738621006042549786426312886309400,82563118828090362261378993957450213573687113690751) (mod 100000000000000000000000000000000000000000000000151)
Sage[edit]
def eulerCriterion(a, p):
return -1 if pow(a, int((p-1)/2), p) == p-1 else 1
def cipollaMult(x1, y1, x2, y2, u, p):
return ((x1*x2 + y1*y2*u) % p), ((x1*y2 + x2*y1) % p)
def cipollaAlgorithm(n, p):
a = Mod(n, p)
out = []
if eulerCriterion(a, p) == -1:
print "❌ " + str(a) + " is not a quadratic residue modulo " + str(p)
return False
if not is_prime(p):
conglst = [] #congruence list
crtlst = []
factors = []
for k in list(factor(p)):
factors.append(int(k[0]))
for f in factors:
conglst.append(cipollaAlgorithm(a, f))
for i in Permutations([0, 1] * len(factors), len(factors)).list():
for j in range(len(factors)):
crtlst.append(int(conglst[ j ][ i[j] ]))
out.append(crt(crtlst, factors))
crtlst = []
return sorted(out)
if pow(p, 1, 4) == 3:
temp = pow(a, int((p+1)/4), p)
return [temp, p - temp]
t = randrange(2, p)
u = pow(t**2 - a, 1, p)
while (eulerCriterion(u, p) == 1):
t = randrange(2, p)
u = pow(t**2 - a, 1, p)
x0, y0 = t, 1
x, y = t, 1
for i in range(int((p + 1) / 2) - 1):
x, y = cipollaMult(x, y, x0, y0, u, p)
out.extend([x, p - x])
return sorted(out)
- Output:
sage: cipollaAlgorithm(10, 13) [6, 7] sage: cipollaAlgorithm(56, 101) [37, 64] sage: cipollaAlgorithm(8218, 10007) [135, 9872] sage: cipollaAlgorithm(331575, 1000003) [144161, 855842] sage: cipollaAlgorithm(8219, 10007) ❌ 8219 is not a quadratic residue modulo 10007 False
Scala[edit]
Imperative solution[edit]
object CipollasAlgorithm extends App {
private val BIG = BigInt(10).pow(50) + BigInt(151)
println(c("10", "13"))
println(c("56", "101"))
println(c("8218", "10007"))
println(c("8219", "10007"))
println(c("331575", "1000003"))
println(c("665165880", "1000000007"))
println(c("881398088036", "1000000000039"))
println(c("34035243914635549601583369544560650254325084643201", ""))
private def c(ns: String, ps: String): Triple = {
val (n, p) = (BigInt(ns), if (ps.isEmpty) BIG else BigInt(ps))
// Legendre symbol, returns 1, 0 or p - 1
def ls(a: BigInt) = a.modPow((p - BigInt(1)) / BigInt(2), p)
// multiplication in Fp2
def mul(aa: Point, bb: Point, omega2: BigInt) =
new Point((aa.x * bb.x + aa.y * bb.y * omega2) % p, (aa.x * bb.y + (bb.x * aa.y)) % p)
// Step 0, validate arguments
if (ls(n) != BigInt(1)) new Triple(0, 0, false)
else {
// Step 1, find a, omega2
var (a, flag, omega2) = (BigInt(0), true, BigInt(0))
while (flag) {
omega2 = (a * a + p - n) % p
if (ls(omega2) == p - BigInt(1)) flag = false else a = a + BigInt(1)
}
// Step 2, compute power
var (nn, r, s) = ((p + BigInt(1) >> 1) % p, new Point(BigInt(1), 0), new Point(a, BigInt(1)))
while (nn > 0) {
if ((nn & BigInt(1)) == BigInt(1)) r = mul(r, s, omega2)
s = mul(s, s, omega2)
nn = nn >> 1
}
// Step 3, check x in Fp
if (r.y != 0) new Triple(0, 0, false)
else // Step 5, check x * x = n
if ((r.x * r.x) % p != n) new Triple(0, 0, false)
else new Triple(r.x, p - r.x, true) // Step 4, solutions
}
}
private class Point(val x: BigInt, val y: BigInt)
private class Triple(val x: BigInt, val y: BigInt, val b: Boolean) {
override def toString: String = f"($x%s, $y%s, $b%s)"
}
}
- Output:
Sidef[edit]
func cipolla(n, p) {
legendre(n, p) == 1 || return nil
var (a = 0, ω2 = 0)
loop {
ω2 = ((a*a - n) % p)
if (legendre(ω2, p) == -1) {
break
}
++a
}
struct point { x, y }
func mul(a, b) {
point((a.x*b.x + a.y*b.y*ω2) % p, (a.x*b.y + b.x*a.y) % p)
}
var r = point(1, 0)
var s = point(a, 1)
for (var n = ((p+1) >> 1); n > 0; n >>= 1) {
r = mul(r, s) if n.is_odd
s = mul(s, s)
}
r.y == 0 ? r.x : nil
}
var tests = [
[10, 13],
[56, 101],
[8218, 10007],
[8219, 10007],
[331575, 1000003],
[665165880, 1000000007],
[881398088036 1000000000039],
[34035243914635549601583369544560650254325084643201, 10**50 + 151],
]
for n,p in tests {
var r = cipolla(n, p)
if (defined(r)) {
say "Roots of #{n} are (#{r} #{p-r}) mod #{p}"
} else {
say "No solution for (#{n}, #{p})"
}
}
- Output:
Roots of 10 are (6 7) mod 13 Roots of 56 are (37 64) mod 101 Roots of 8218 are (9872 135) mod 10007 No solution for (8219, 10007) Roots of 331575 are (855842 144161) mod 1000003 Roots of 665165880 are (475131702 524868305) mod 1000000007 Roots of 881398088036 are (791399408049 208600591990) mod 1000000000039 Roots of 34035243914635549601583369544560650254325084643201 are (82563118828090362261378993957450213573687113690751 17436881171909637738621006042549786426312886309400) mod 100000000000000000000000000000000000000000000000151
zkl[edit]
Uses lib GMP (GNU MP Bignum Library).
var [const] BN=Import("zklBigNum"); //libGMP
fcn modEq(a,b,p) { (a-b)%p==0 }
fcn Legendre(a,p){ a.powm((p - 1)/2,p) }
class Fp2{ // Arithmetic in Fp^2
fcn init(_x,_y){ var [const] x=BN(_x), y=BN(_y) } // two big ints
//fcn add(b,p){ self((x + b.x)%p,(y + b.y)%p) } // a + b
fcn mul(b,p,w2){ self(( x*b.x + y*b.y*w2 )%p, (x*b.y + y*b.x) %p) } // a * b
fcn square(p,w2){ mul(self,p,w2) } // a * a == self.mul(self,p,w2)
fcn pow(n,p,w2){ // a ^ n
if (n==0) self(1,0);
else if(n==1) self;
else if(n==2) square(p,w2);
else if(n.isEven) pow(n/2,p,w2).square(p,w2);
else mul(pow(n-1,p,w2),p,w2)
}
}
fcn Cipolla(n,p){ n=BN(n); // x^2 == n (mod p) ?
if(Legendre(n,p)!=1) // check n is a square
throw(Exception.AssertionError("not a square (mod p)"+vm.arglist));
// iterate until suitable 'a' found (the first one found)
a:=[BN(2)..p].filter1('wrap(a){ Legendre(a*a-n,p)==(p-1) });
w2:=a*a - n;
r:=Fp2(a,1).pow((p + 1)/2,p,w2); // (Fp2 a 1) = a + w2
x:=r.x;
_assert_(r.y==0,"r.y==0 : "+r.y); // hope that w has vanished
_assert_(modEq(n,x*x,p),"modEq(n,x*x,p)"); // checking the result
println("Roots of %d are (%d,%d) (mod %d)".fmt(n,x,(p-x)%p,p));
return(x,(p-x)%p);
}
foreach n,p in (T(
T(10,13),T(56,101),T(8218,10007),T(8219,10007),T(331575,1000003),
T(665165880,1000000007),T(881398088036,1000000000039),
T("34035243914635549601583369544560650254325084643201",
BN(10).pow(50) + 151) )){
try{ Cipolla(n,p) }catch{ println(__exception) }
}
- Output:
Roots of 10 are (6,7) (mod 13) Roots of 56 are (37,64) (mod 101) Roots of 8218 are (9872,135) (mod 10007) AssertionError(not a square (mod p)L(8219,10007)) Roots of 331575 are (855842,144161) (mod 1000003) Roots of 665165880 are (524868305,475131702) (mod 1000000007) Roots of 881398088036 are (208600591990,791399408049) (mod 1000000000039) Roots of 34035243914635549601583369544560650254325084643201 are (17436881171909637738621006042549786426312886309400,82563118828090362261378993957450213573687113690751) (mod 100000000000000000000000000000000000000000000000151)
