A.cpp

#include <iostream>
#include <cstring>
#include <cmath>
#include <string>
#include <random>

namespace UPmath
{
#define FFT_MULTIPLICATION
#define SIZE_OF_PIECE 4
#define BITS_OF_DWORD 32
    typedef unsigned int uint32;
    static_assert(sizeof(uint32) == SIZE_OF_PIECE, "sizeof(uint32) must be 4 bytes!");

    extern std::random_device gRandomDevice;

    class BigInteger
    {
    public:
        uint32 capacity = 0, size = 0;//in DWORD (NOT byte)
        bool negativity = false;//*this <= 0 when `negativity` is true and >= 0 when it is false

        BigInteger();
        BigInteger(bool isHexadecimal, const std::string&);//decimal or hexdecimal
        explicit BigInteger(const std::string&);//init as hexadecimal only if the string starts with `0x` or `0X`
        BigInteger(int);
        BigInteger(long long);
        BigInteger(unsigned long long);
        BigInteger(const void* dataPtr, size_t dataSize);//for cryptography
        BigInteger(const BigInteger& source);
        const BigInteger& operator=(const BigInteger& source);
        BigInteger(BigInteger&& source);
        const BigInteger& operator=(BigInteger&& source);
        ~BigInteger() { delete[] _valPtr; }

        int intValue() const;//abs modular 0x80000000
        double doubleValue() const;
        void saveBytesToBuffer(char* dst, size_t writingSize) const;
        const std::string toString() const;
        char* toHexadecimalCharArray() const;
        bool* convertAbsToBinaryArray(size_t* out_digitsSize) const;//from lower digit to higher. eg: 4 -> 001
        size_t getBitLength() const;

        void setLowerBitsToRandom(uint32 bitLength);
        bool testBit(uint32 bitPos) const;
        int compareAbsoluteValueTo(const BigInteger&) const;
        int compareTo(const BigInteger&) const;
        bool operator==(const BigInteger& rhs) const { return compareTo(rhs) == 0; }
        bool operator!=(const BigInteger& rhs) const { return compareTo(rhs) != 0; }
        bool operator<(const BigInteger& rhs) const { return compareTo(rhs) < 0; }
        bool operator>(const BigInteger& rhs) const { return compareTo(rhs) > 0; }
        bool operator<=(const BigInteger& rhs) const { return compareTo(rhs) <= 0; }
        bool operator>=(const BigInteger& rhs) const { return compareTo(rhs) >= 0; }
        BigInteger operator-() const { BigInteger t(*this); t.negativity = !t.negativity; return t; }
        const BigInteger& operator+=(const BigInteger& rhs);
        const BigInteger& operator-=(const BigInteger& rhs);
        const BigInteger& operator>>=(const uint32 shift);
        const BigInteger& operator<<=(const uint32 shift);
        friend BigInteger operator*(const BigInteger& lhs, const BigInteger& rhs);
        BigInteger absDivideAndSetThisToRemainder(const BigInteger& rhs);
        const BigInteger& operator%=(const BigInteger& rhs);
        const BigInteger& operator/=(const BigInteger& rhs);
        BigInteger sqrt(bool ceilling = false) const;
        BigInteger pow(uint32 positive_exponent) const;
        BigInteger modInverse(const BigInteger& m) const;//return BigInteger(0) when the inverse doesnot exist
        BigInteger modPow(const BigInteger& exponent, const BigInteger& m) const;
        struct _EEAstruct;
        static BigInteger gcd(const BigInteger& a, const BigInteger& b);
        static int JacobiSymbol(const BigInteger& upperArgument, const BigInteger& lowerArgument);
        bool isPrime(int confidenceFactor = -1) const;
        bool weakerBailliePSWPrimeTest() const;
        BigInteger nextProbablePrime() const;

        static const BigInteger kBigInteger_0;
        static const BigInteger kBigInteger_1;
    private:
        uint32* _valPtr = nullptr;
        class _SharedMemBigInteger;
        template <uint32 FractionalBitLen> friend class FixedPointNum;
        void _setSize(uint32 maxPossibleSize);
        void _absSubtract(const BigInteger& rhs);//only if abs(rhs) <= abs(*this)
        static BigInteger _absMultiply(const BigInteger& lhs, const BigInteger& rhs);
        void _seperateToHigherAndLowerHalfs(BigInteger* dst) const;
#ifdef FFT_MULTIPLICATION
#define FFT_MAX_N 1024u
#define FFT_WORD_BITLEN 16
#define _IS_LITTLE_ENDIAN() (true)
    public:
        inline void _FFTgetBeginAndEndOfVal(unsigned short* &begin, unsigned short* &end) const
        {
            begin = reinterpret_cast<unsigned short*>(_valPtr);
            end = reinterpret_cast<unsigned short*>(_valPtr + size);
        }
    private:
        static bool _isRootsOfUnityInitialized;
        static void _FFTMultiply(BigInteger& dst, const BigInteger& lhs, const BigInteger& rhs, void* buffer = nullptr);
#endif
        BigInteger _trivialModPow(bool* binaryArrayOfExponent, size_t bitsOfExponent, const BigInteger& m) const;
        BigInteger _montgomeryModPow(bool* binaryArrayOfExponent, size_t bitsOfExponent, const BigInteger& m) const;
        static _EEAstruct _extendedEuclid(BigInteger* a, BigInteger* b);
        static BigInteger* _euclidGcd(BigInteger* a, BigInteger* b);
        bool _MillerRabinPrimalityTest(const BigInteger& a, const BigInteger& thisMinus1) const;//logically wrong when *this <= 2
        static int _JacobiSymbolImpl(BigInteger* numerator, BigInteger* denominator);
        std::pair<BigInteger, BigInteger> _getModLucasSequence(const BigInteger& k, const BigInteger& D, const BigInteger& P, const BigInteger& Q) const;
        bool _LucasPseudoprimeTest() const;//logically wrong when *this <= 2

        static unsigned char _requiredCapacityList[129];
        static uint32 _calcMinimumCapacity(uint32 size_);
        inline static uint32 _getMinimumCapacity(uint32 size_)
        {
            if (size_ < sizeof(_requiredCapacityList)) return _requiredCapacityList[size_];
            return _calcMinimumCapacity(size_);
        }

    };

    struct BigInteger::_EEAstruct { BigInteger *d, *x, *y; };

    class BigInteger::_SharedMemBigInteger : public BigInteger {
    public:
        _SharedMemBigInteger(const BigInteger& src) { _valPtr = src._valPtr; negativity = src.negativity; capacity = src.capacity; size = src.size;	}
        _SharedMemBigInteger(uint32 size_, uint32* val);
        ~_SharedMemBigInteger() { _valPtr = nullptr; }
    };

    template <class T> T operator+(const T& lhs, const T& rhs) { return T(lhs) += rhs; }
    template <class T> T operator-(const T& lhs, const T& rhs) { return T(lhs) -= rhs; }
    template <class T> T operator%(const T& lhs, const T& rhs) { return T(lhs) %= rhs; }
    template <class T> T operator/(const T& lhs, const T& rhs) { return T(lhs) /= rhs; }
    template <class T> T operator>>(const T& lhs, const uint32 shift) { return T(lhs) >>= shift; }
    template <class T> T operator<<(const T& lhs, const uint32 shift) { return T(lhs) <<= shift; }

    inline std::ostream& operator<<(std::ostream& os, const BigInteger& src)
    {
        os << src.toString();
        return os;
    }
    inline std::istream& operator>>(std::istream& is, BigInteger& dst)
    {
        std::string str;
        is >> str;
        dst = BigInteger(str);
        return is;
    }


}//namespace UPMath

namespace UPmath
{
    std::random_device gRandomDevice;

    const BigInteger BigInteger::kBigInteger_0 = BigInteger(0);
    const BigInteger BigInteger::kBigInteger_1 = BigInteger(1);

    BigInteger::BigInteger() { }

    BigInteger::BigInteger(int num)
    {
        if (num) {
            if (num < 0) { negativity = true; num = -num; }
            (_valPtr = new uint32[capacity = size = 1])[0] = num;
        }
    }

    BigInteger::BigInteger(long long num) : BigInteger((unsigned long long)(num < 0 ? -num : num)) { negativity = (num < 0); }
    BigInteger::BigInteger(unsigned long long num)
    {
        static_assert(sizeof(long long) / sizeof(uint32) == 2, "Compile error in BigInteger(long long)");
        _valPtr = new uint32[capacity = 2];
        if (num) _valPtr[0] = static_cast<uint32>(num), ++size;
        num >>= BITS_OF_DWORD;
        if (num) _valPtr[1] = static_cast<uint32>(num), ++size;
    }

    BigInteger::BigInteger(bool isHexadecimal, const std::string& str)
    {
        int i = 0, strLen = (int)str.length(), t;
        bool neg_ = false;
        if (str[i] == '-') { neg_ = true; ++i; }
        else if (str[i] == '+') ++i;
        if (!isHexadecimal)
        {
            while (i < strLen && (t = (int)str[i] - '0') >= 0 && t < 10) {
                BigInteger tupi = *this <<= 1;
                *this <<= 2; *this += tupi; *this += BigInteger(t);
                ++i;
            }
        }
        else
        {
            (_valPtr = new uint32[capacity = 1])[0] = 0;
            if (strLen >= 2 && ( ('\x5F' & str[i + 1]) == 'X')) i += 2;
            while (i < strLen && (t = (int)str[i] - '0') >= 0) {
                if (t > 9) {
                    t = (int)(str[i] & '\x5F') - 'A' + 10;
                    if ((t < 10) | (t >= 16)) break;
                }
                *this <<= 4;
                this->_valPtr[0] |= t;
                if (size == 0 && t) ++size;
                ++i;
            }
        }
        negativity = neg_;
    }

    BigInteger::BigInteger(const BigInteger& source)
    {
        negativity = source.negativity;
        size = source.size;
        capacity = _getMinimumCapacity(size);
        _valPtr = new uint32[capacity];
        uint32 *_temptr = _valPtr, *_srcptr = source._valPtr;
        for (uint32* _comptr = _temptr + size; _temptr < _comptr; ++_temptr, ++_srcptr)	*_temptr = *_srcptr;
    }

    const BigInteger& BigInteger::operator=(const BigInteger& source)
    {
        if (this == &source) return *this;
        if (source.size > this->capacity) {
            delete[] this->_valPtr;
            this->_valPtr = new uint32[this->capacity = _getMinimumCapacity(source.size)];
        }
        negativity = source.negativity;
        size = source.size;
        uint32 *_temptr = _valPtr, *_srcptr = source._valPtr;
        for (uint32* _comptr = _temptr + size; _temptr < _comptr; ++_temptr, ++_srcptr)	*_temptr = *_srcptr;
        return *this;
    }

    const BigInteger& BigInteger::operator=(BigInteger&& source)
    {
        delete[] _valPtr;
        _valPtr = source._valPtr;
        negativity = source.negativity;
        capacity = source.capacity;
        size = source.size;
        source._valPtr = nullptr;
        source.capacity = source.size = 0;
        return *this;
    }

    BigInteger::BigInteger(BigInteger&& source)
    {
        _valPtr = source._valPtr;
        negativity = source.negativity;
        capacity = source.capacity;
        size = source.size;
        source._valPtr = nullptr;
        source.capacity = source.size = 0;
    }

    BigInteger::BigInteger(const std::string& str) :
            BigInteger((str.length() >= 2 && str[0] == '0' && (('\x5F' & str[1]) == 'X')), str) { }

    void BigInteger::_setSize(uint32 maxPossibleSize)
    {
        if ((size = maxPossibleSize))
            for (uint32* ptr = _valPtr + size; --ptr >= _valPtr; --size)
                if (*ptr) break;
    }

    unsigned char BigInteger::_requiredCapacityList[129] = 	{
            0, 1, 2, 4, 4, 8, 8, 8, 8, 16, 16, 16, 16, 16, 16, 16, 16, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32,
            64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
            128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128,
            128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128 };

    uint32 BigInteger::_calcMinimumCapacity(uint32 size_)
    {
        uint32 minCap = (size_ < sizeof(_requiredCapacityList)) ? 1 : 256;
        while (size_ > minCap) minCap <<= 1;
        return minCap;
    }

    int BigInteger::compareAbsoluteValueTo(const BigInteger& rhs) const
    {
        if (size != rhs.size)
            return (size > rhs.size) ? 1 : -1;
        for (uint32 i = size; i > 0; ) {
            --i;
            if (_valPtr[i] != rhs._valPtr[i])
                return (_valPtr[i] > rhs._valPtr[i]) ? 1 : -1;
        }
        return 0;
    }

    int BigInteger::compareTo(const BigInteger& rhs) const
    {
        if (negativity != rhs.negativity) {
            if (0 == (size | rhs.size)) return 0;
            return negativity ? -1 : 1;
        }
        int result = compareAbsoluteValueTo(rhs);
        return negativity ? -result : result;
    }

    //only if abs(rhs) <= abs(*this)
    void BigInteger::_absSubtract(const BigInteger& rhs)
    {
        bool borrowFlag = false;
        uint32* ptr = _valPtr;
        for (uint32 ti = 0; ti < rhs.size; ++ti, ++ptr) {
            uint32 temp = *ptr;
            *ptr -= rhs._valPtr[ti];
            *ptr -= borrowFlag;
            borrowFlag = (*ptr > temp) | ((*ptr == temp) & borrowFlag);
        }
        if (borrowFlag) while (true) if ((*(ptr++))--) break;
        _setSize(size);
    }

    const BigInteger& BigInteger::operator+=(const BigInteger& rhs)
    {
        if (this == &rhs) *this <<= 1;
        else if (negativity == rhs.negativity)
        {
            uint32 commonPiecesCount = (size > rhs.size) ? rhs.size : size;
            uint32 *l_ptr, *r_ptr, *c_ptr;
            if (commonPiecesCount != rhs.size)
            {//rhs is longer
                if (capacity < rhs.size) {
                    uint32* newptr = new uint32[capacity = _getMinimumCapacity(rhs.size)];
                    for (c_ptr = (l_ptr = _valPtr) + size; l_ptr < c_ptr; ++newptr, ++l_ptr) *newptr = *l_ptr;
                    delete[] _valPtr;
                    _valPtr = newptr - size;
                }
                l_ptr = _valPtr + size;
                r_ptr = rhs._valPtr + size;
                c_ptr = rhs._valPtr + (size = rhs.size);
                while (r_ptr < c_ptr) *(l_ptr++) = *(r_ptr++);
            }
            l_ptr = _valPtr;
            r_ptr = rhs._valPtr;
            bool carryFlag = false;
            for (c_ptr = rhs._valPtr + commonPiecesCount; r_ptr < c_ptr; ++l_ptr, ++r_ptr)
            {
                *l_ptr += *r_ptr;
                *l_ptr += carryFlag;
                carryFlag = (*l_ptr < *r_ptr) | ((*l_ptr == *r_ptr) & carryFlag);
            }
            if (carryFlag) {
                for (c_ptr = _valPtr + size; l_ptr < c_ptr; ++l_ptr) if (++*l_ptr) break;
                if (l_ptr == c_ptr) {
                    if (capacity > size)
                        _valPtr[size++] = 1;
                    else {
                        l_ptr = new uint32[capacity <<= 1];
                        c_ptr = l_ptr + size;
                        while (l_ptr < c_ptr) *(l_ptr++) = *(_valPtr++);
                        *l_ptr = 1;
                        delete[] (_valPtr -= size);
                        _valPtr = l_ptr - size++;
                    }
                }
            }
        }
        else if (compareAbsoluteValueTo(rhs) >= 0)
            this->_absSubtract(rhs);
        else
        {
            BigInteger absGreaterOne(rhs);
            absGreaterOne._absSubtract(*this);
            *this = std::move(absGreaterOne);
        }
        return *this;
    }

    const BigInteger& BigInteger::operator-=(const BigInteger& rhs)
    {
        if (this == &rhs) { this->size = 0; return *this; }
        _SharedMemBigInteger signModified_rhs(rhs);
        signModified_rhs.negativity = !rhs.negativity;
        *this += signModified_rhs;
        return *this;
    }

    const BigInteger& BigInteger::operator>>=(const uint32 shift)
    {
        uint32 a = shift / BITS_OF_DWORD;
        if (a >= size) { this->size = 0; return *this; }
        else if (a) {
            uint32 i = 0;
            size -= a;
            do {
                _valPtr[i] = _valPtr[i + a];
                ++i;
            } while (i < size);
            for (a += size; i < a; ++i) _valPtr[i] = 0;
        }
        a = shift & (BITS_OF_DWORD - 1);
        if (a && size)
        {
            uint32 pm1 = size - 1;
            for (uint32 i = 0; i < pm1; ++i)
                _valPtr[i] = (_valPtr[i] >> a) | ((_valPtr[i + 1] & ((1 << a) - 1)) << (32 - a));
            if (0 == (_valPtr[pm1] >>= a)) --size;
        }
        return *this;
    }

    const BigInteger& BigInteger::operator<<=(const uint32 shift)
    {
        if (size)
        {
            uint32 a = shift / BITS_OF_DWORD, b = shift & (BITS_OF_DWORD - 1);
            bool newMemAlloced = false;
            if (b) ++a;
            if (a) {
                uint32 *souptr, *tarptr, *comptr;
                if (capacity < (size += a)) {
                    newMemAlloced = true;
                    capacity = _getMinimumCapacity(size);
                    tarptr = new uint32[capacity];
                    comptr = tarptr + size;
                    tarptr += capacity;
                    while (tarptr > comptr) *(--tarptr) = 0;
                }
                else
                    tarptr = _valPtr + size;
                souptr = _valPtr + (size - a);
                while (souptr > _valPtr) *(--tarptr) = *(--souptr);
                for (comptr = tarptr - a; tarptr > comptr; ) *(--tarptr) = 0;
                if (newMemAlloced) {
                    delete[] _valPtr;
                    _valPtr = tarptr;
                }
            }
            if (b) *this >>= (BITS_OF_DWORD - b);
        }
        return *this;
    }

    BigInteger operator*(const BigInteger& lhs, const BigInteger& rhs)
    {//multipication
        BigInteger ret = BigInteger::_absMultiply(lhs, rhs);
        ret.negativity = (lhs.negativity != rhs.negativity);
        return ret;
    }

    void BigInteger::_seperateToHigherAndLowerHalfs(BigInteger* dst) const
    {
        uint32 halfCapacity = capacity >> 1;
        dst[0]._valPtr = _valPtr + halfCapacity;
        dst[0].size = (size < halfCapacity) ? 0 : size - halfCapacity;
        dst[0].capacity = _getMinimumCapacity(dst[0].size);
        dst[1]._valPtr = _valPtr;
        dst[1]._setSize((size < halfCapacity) ? size : halfCapacity);
        dst[1].capacity = _getMinimumCapacity(dst[1].size);
    }

    BigInteger BigInteger::_absMultiply(const BigInteger& lhs, const BigInteger& rhs)
    {
        if ((!lhs.size) | (!rhs.size)) return BigInteger();
#ifdef FFT_MULTIPLICATION
        uint32 sizeSum = lhs.size + rhs.size;
        if (sizeSum == 2) return BigInteger((unsigned long long)lhs._valPtr[0] * rhs._valPtr[0]);
        if (sizeSum <= FFT_MAX_N / (BITS_OF_DWORD / FFT_WORD_BITLEN)) {
            BigInteger ret;
            _FFTMultiply(ret, lhs, rhs);
            return ret;
        }
#endif
        if (lhs.capacity != rhs.capacity)
        {
            const BigInteger& longer = (lhs.capacity >= rhs.capacity) ? lhs : rhs;
            const BigInteger& shorter = (lhs.capacity >= rhs.capacity) ? rhs : lhs;
            unsigned char _sharedMemBigIntegers[2 * sizeof(BigInteger)];
            BigInteger* sharedMemeoryBigIntegers = reinterpret_cast<BigInteger*>(_sharedMemBigIntegers);
            longer._seperateToHigherAndLowerHalfs(sharedMemeoryBigIntegers);//initialize read only variables
            BigInteger ret = _absMultiply(sharedMemeoryBigIntegers[0], shorter);
            ret <<= (longer.capacity << 4);
            ret += _absMultiply(sharedMemeoryBigIntegers[1], shorter);
            return ret;
        }
        if (lhs.capacity > 1)
        {
            unsigned char _sharedMemBigIntegers[4 * sizeof(BigInteger)];
            BigInteger *lhsParts = reinterpret_cast<BigInteger*>(_sharedMemBigIntegers), *rhsParts = 2 + lhsParts;//read only variables
            lhs._seperateToHigherAndLowerHalfs(lhsParts);
            rhs._seperateToHigherAndLowerHalfs(rhsParts);//initialize read only variables
            BigInteger higher_part = _absMultiply(lhsParts[0], rhsParts[0]);
            BigInteger lower_part = _absMultiply(lhsParts[1], rhsParts[1]);
            BigInteger mid_part = higher_part + lower_part;
            mid_part += ((lhsParts[1] - lhsParts[0]) * (rhsParts[0] - rhsParts[1]));
            higher_part <<= (lhs.capacity * BITS_OF_DWORD);
            mid_part <<= (lhs.capacity * (BITS_OF_DWORD / 2));
            higher_part += (mid_part += lower_part);
            return higher_part;
        }
        return BigInteger((unsigned long long)*lhs._valPtr * *rhs._valPtr);
    }

    //Set *this to remainder and return the quotient.
    //It is ok to execute this->absDivideAndSetThisToRemainder(*this)
    BigInteger BigInteger::absDivideAndSetThisToRemainder(const BigInteger& rhs)
    {
        int pDiff = this->size - rhs.size;
        if (0 == rhs.size) throw std::logic_error("Divide by zero");
        if (pDiff < 0) return BigInteger();
        int pow = pDiff * BITS_OF_DWORD;
        int lhsLeftmostBitLen = 0, rhsLeftmostBitLen = 0;
        for (uint32 lhsLeftmost = _valPtr[size - 1], rhsLeftmost = rhs._valPtr[rhs.size - 1]; lhsLeftmost | rhsLeftmost; ) {
            if (lhsLeftmost) ++lhsLeftmostBitLen;
            if (rhsLeftmost) ++rhsLeftmostBitLen;
            lhsLeftmost >>= 1; rhsLeftmost >>= 1;
        }
        pow += 1 + lhsLeftmostBitLen - rhsLeftmostBitLen;
        if (pow <= 0) return BigInteger();
        BigInteger newrhs(rhs);
        newrhs <<= --pow;
        int leftMostBit = pow;
        bool fillOne = !(newrhs.compareAbsoluteValueTo(*this) > 0);
        if (fillOne)
            this->_absSubtract(newrhs);
        else if (--leftMostBit < 0)
            return BigInteger();
        bool* binaryList = new bool[pow + 1];
        binaryList[pow] = fillOne;
        while (--pow >= 0) {
            newrhs >>= 1;
            fillOne = binaryList[pow] = !(newrhs.compareAbsoluteValueTo(*this) > 0);
            if (fillOne) this->_absSubtract(newrhs);
        }
        BigInteger ret;
        ret.size = 1 + ((uint32)leftMostBit / BITS_OF_DWORD);
        ret._valPtr = new uint32[ret.capacity = _getMinimumCapacity(ret.size)];
        for (uint32 piece = 0; pow < leftMostBit; ++piece) {
            ret._valPtr[piece] = 0;
            int biMax = leftMostBit - pow;
            if (biMax > BITS_OF_DWORD) biMax = BITS_OF_DWORD;
            for (int bi = 0; bi < biMax; ++bi)
                ret._valPtr[piece] |= ((uint32)binaryList[++pow] << bi);
        }
        delete[] binaryList;
        return ret;
    }

    const BigInteger& BigInteger::operator/=(const BigInteger& rhs)
    {
        bool neg_ = (this->negativity != rhs.negativity);
        *this = this->absDivideAndSetThisToRemainder(rhs);
        this->negativity = neg_;
        return *this;
    }

    const BigInteger& BigInteger::operator%=(const BigInteger& rhs)
    {
        this->absDivideAndSetThisToRemainder(rhs);
        return *this;
    }


    int BigInteger::intValue() const
    {
        if (size == 0) return 0;
        int ret = _valPtr[0] & 0x7FFFFFFF;
        return negativity ? -ret : ret;
    }

    double BigInteger::doubleValue() const
    {
        double ret = 0.0;
        for (uint32 i = 0; i < size; ++i) {
            ret *= (double)0x100000000;
            ret += _valPtr[i];
        }
        return negativity ? -ret : ret;
    }

    const std::string BigInteger::toString() const
    {
        if (size == 0) return std::string("0");
        uint32 t = (size * 10) + 2;
        char* buffer = new char[t];
        char* c = buffer + t;
        *(--c) = '\0';
        const BigInteger ten(10);
        BigInteger* newthis = new BigInteger(*this);
        do {
            BigInteger* temptr = new BigInteger(newthis->absDivideAndSetThisToRemainder(ten));
            *(--c) = *(newthis->_valPtr) + '0';
            delete newthis;
            newthis = temptr;
        } while (newthis->size);
        delete newthis;
        if (negativity) *(--c) = '-';
        std::string ret(c);
        delete[] buffer;
        return ret;
    }

    char* BigInteger::toHexadecimalCharArray() const
    {
        char* c_str = new char[(size << 3) + 2];
        char* c = c_str;
        if (size)
        {
            if (negativity) *(c++) = '-';
            uint32 *v_ptr;
            for (v_ptr = _valPtr + size; v_ptr > _valPtr; ) {
                uint32 piece = *(--v_ptr), shifting = BITS_OF_DWORD;
                do {
                    shifting -= 4;
                    uint32 letter = (piece >> shifting) & 0xF;
                    *(c++) = (letter > 9) ? ('A' - 10 + letter) : ('0' + letter);
                } while (shifting);
            }
        }
        else *(c++) = '0';
        *c = '\0';
        return c_str;
    }

    bool* BigInteger::convertAbsToBinaryArray(size_t* out_digitsSize) const
    {
        if (!this->size) {
            *out_digitsSize = 0;
            return nullptr;
        }
        size_t digitsSize = this->size * BITS_OF_DWORD;
        bool* ret = new bool[digitsSize];
        for (uint32 i = 0, pieceIndex = 0; pieceIndex < this->size; ++pieceIndex, i += BITS_OF_DWORD)
        {
            uint32 piece = this->_valPtr[pieceIndex];
            ret[i + 0] = (piece & (1 << 0)) != 0; ret[i + 1] = (piece & (1 << 1)) != 0;	ret[i + 2] = (piece & (1 << 2)) != 0; ret[i + 3] = (piece & (1 << 3)) != 0;
            ret[i + 4] = (piece & (1 << 4)) != 0; ret[i + 5] = (piece & (1 << 5)) != 0;	ret[i + 6] = (piece & (1 << 6)) != 0; ret[i + 7] = (piece & (1 << 7)) != 0;
            ret[i + 8] = (piece & (1 << 8)) != 0; ret[i + 9] = (piece & (1 << 9)) != 0; ret[i + 10] = (piece & (1 << 10)) != 0; ret[i + 11] = (piece & (1 << 11)) != 0;
            ret[i + 12] = (piece & (1 << 12)) != 0;	ret[i + 13] = (piece & (1 << 13)) != 0;	ret[i + 14] = (piece & (1 << 14)) != 0;	ret[i + 15] = (piece & (1 << 15)) != 0;
            static_assert(BITS_OF_DWORD == 32, "convertAbsToBinaryArray");
            ret[i + 16] = (piece & (1 << 16)) != 0;	ret[i + 17] = (piece & (1 << 17)) != 0;	ret[i + 18] = (piece & (1 << 18)) != 0;	ret[i + 19] = (piece & (1 << 19)) != 0;
            ret[i + 20] = (piece & (1 << 20)) != 0;	ret[i + 21] = (piece & (1 << 21)) != 0;	ret[i + 22] = (piece & (1 << 22)) != 0;	ret[i + 23] = (piece & (1 << 23)) != 0;
            ret[i + 24] = (piece & (1 << 24)) != 0;	ret[i + 25] = (piece & (1 << 25)) != 0;	ret[i + 26] = (piece & (1 << 26)) != 0;	ret[i + 27] = (piece & (1 << 27)) != 0;
            ret[i + 28] = (piece & (1 << 28)) != 0;	ret[i + 29] = (piece & (1 << 29)) != 0;	ret[i + 30] = (piece & (1 << 30)) != 0;	ret[i + 31] = (piece & (1 << 31)) != 0;
        }
        if (out_digitsSize) {
            do { --digitsSize; } while (!ret[digitsSize]);
            *out_digitsSize = ++digitsSize;
        }
        return ret;
    }

    size_t BigInteger::getBitLength() const
    {
        if (size == 0) return 0;
        size_t lowerBound = size * BITS_OF_DWORD - BITS_OF_DWORD;
        uint32 x = _valPtr[size - 1];
        while (x) { ++lowerBound; x >>= 1; }
        return lowerBound;
    }

    BigInteger BigInteger::sqrt(bool ceilling) const
    {//x = (prev_x + *this/prev_x) / 2
        if (this->negativity && this->size != 0) throw std::logic_error("Cannot evaluate square root of a negative Biginteger.");
        BigInteger x(1), prev_x;
        do {
            prev_x = x;
            x += *this / prev_x;
            x >>= 1;
        } while (x != prev_x);
        if (ceilling) {
            BigInteger xSquared = x * x;
            if (xSquared < *this) x += kBigInteger_1;
        }
        return x;
    }

    BigInteger BigInteger::pow(uint32 positive_exponent) const
    {
        if (0 == positive_exponent) return BigInteger(1);
        BigInteger t = this->pow(positive_exponent >> 1);
        BigInteger tSquared = _absMultiply(t, t);
        if (positive_exponent & 1) {
            BigInteger ret = _absMultiply(*this, tSquared);
            ret.negativity = this->negativity;
            return ret;
        }
        return tSquared;
    }

    //`m` must be a odd number and larger than BigInteger(2) to apply montgomeryModPow method
    BigInteger BigInteger::modPow(const BigInteger& exponent, const BigInteger& m) const
    {
        if (0 == exponent.size) return BigInteger(1);
        if ((0 == m.size) | m.negativity) throw std::logic_error("modular for `modPow` function must be positive.");
        if (1 == (m._valPtr[0] | m.size)) return BigInteger(0);
        bool montgomeryModPow = m._valPtr[0] & 1;
        size_t digitsSizeOfExponent;
        bool* binaryArrayOfExponent = exponent.convertAbsToBinaryArray(&digitsSizeOfExponent);
        BigInteger base = exponent.negativity ? this->modInverse(m) : *this;
        if (base.compareAbsoluteValueTo(m) > 0) base.absDivideAndSetThisToRemainder(m);
        if (base.negativity) base += m;
        BigInteger result = !montgomeryModPow ?
                            base._trivialModPow(binaryArrayOfExponent, digitsSizeOfExponent, m) :
                            base._montgomeryModPow(binaryArrayOfExponent, digitsSizeOfExponent, m);
        delete[] binaryArrayOfExponent;
        return result;
    }

    BigInteger BigInteger::_trivialModPow(bool* binaryArrayOfExponent, size_t bitsOfExponent, const BigInteger& m) const
    {
        BigInteger result(1), newResult;
        BigInteger tSquared(*this), thisPow_i;
        for (size_t i = 0; i < bitsOfExponent; ++i) {
            thisPow_i = std::move(tSquared);
            if (binaryArrayOfExponent[i]) {
                newResult = _absMultiply(result, thisPow_i);
                newResult.absDivideAndSetThisToRemainder(m);
                result = std::move(newResult);
            }
            tSquared = _absMultiply(thisPow_i, thisPow_i);
            tSquared.absDivideAndSetThisToRemainder(m);
        }
        return result;
    }

    //http://www.hackersdelight.org/MontgomeryMultiplication.pdf
    BigInteger BigInteger::_montgomeryModPow(bool* binaryArrayOfExponent, size_t bitsOfExponent, const BigInteger& m) const
    {
        BigInteger r(1);
        r <<= m.size * BITS_OF_DWORD;
        BigInteger r_tmp(r), m_tmp(m);
        _EEAstruct eeaStruct = _extendedEuclid(&r_tmp, &m_tmp);
        BigInteger r_slash = std::move(*eeaStruct.x);
        BigInteger m_slash = std::move(*eeaStruct.y); m_slash.negativity = !m_slash.negativity;
        delete eeaStruct.d; delete eeaStruct.x; delete eeaStruct.y;

        BigInteger result_mf(r), newResult;
        BigInteger thisPow_2i(*this * r), thisPow_i;
        result_mf.absDivideAndSetThisToRemainder(m);
        thisPow_2i.absDivideAndSetThisToRemainder(m);
        for (size_t i = 0; i < bitsOfExponent; ++i) {
            thisPow_i = std::move(thisPow_2i);
            if (binaryArrayOfExponent[i]) {
                newResult = result_mf * thisPow_i;//t = a_ * b_
                BigInteger tm_slash = newResult * m_slash;//tm'
                if (tm_slash.size > m.size) tm_slash.size = m.size;//tm' mod r
                newResult += tm_slash * m;// (t + (tm' mod r)m)
                newResult >>= BITS_OF_DWORD * m.size;//u = (t + (tm' mod r)m) / r
                if (newResult.compareAbsoluteValueTo(m) >= 0) newResult._absSubtract(m);//if (u >= m) return u - m;
                result_mf = std::move(newResult);
            }
            thisPow_2i = thisPow_i * thisPow_i;
            BigInteger tm_slash = thisPow_2i * m_slash;
            if (tm_slash.size > m.size) tm_slash.size = m.size;
            thisPow_2i += tm_slash * m;
            thisPow_2i >>= BITS_OF_DWORD * m.size;
            if (thisPow_2i.compareAbsoluteValueTo(m) >= 0) thisPow_2i._absSubtract(m);
        }

        BigInteger result = result_mf * r.modInverse(m);
        result.absDivideAndSetThisToRemainder(m);
        if (result < 0) result += m;
        return result;
    }

    //return BigInteger(0) when the inverse doesnot exist
    BigInteger BigInteger::modInverse(const BigInteger& m) const
    {
        if ((0 == m.size) | m.negativity) throw std::logic_error("modular for `modInverse` function must be positive.");
        if (gcd(*this, m) != kBigInteger_1) return BigInteger(0);
        BigInteger a_(*this), m_(m);
        _EEAstruct eeaStruct = _extendedEuclid(&a_, &m_);
        BigInteger ddd = *this * *eeaStruct.x + m * *eeaStruct.y;
        BigInteger ret = std::move(*eeaStruct.x);
        delete eeaStruct.d; delete eeaStruct.x; delete eeaStruct.y;
        if (ret.negativity) ret += m;
        return ret;
    }

    BigInteger::_EEAstruct BigInteger::_extendedEuclid(BigInteger* a, BigInteger* b)
    {
        if (b->size == 0) return BigInteger::_EEAstruct{ new BigInteger(*a), new BigInteger(1), new BigInteger(0) };
        bool neg_ = (a->negativity != b->negativity);
        BigInteger quotient = a->absDivideAndSetThisToRemainder(*b);
        quotient.negativity = neg_;
        BigInteger::_EEAstruct ts = _extendedEuclid(b, a);
        *ts.x -= quotient * *ts.y;
        return BigInteger::_EEAstruct{ ts.d, ts.y, ts.x };
    }

    BigInteger BigInteger::gcd(const BigInteger& a, const BigInteger& b)
    {
        BigInteger a_(a), b_(b);
        BigInteger* result = _euclidGcd(&a_, &b_);
        BigInteger ret = std::move(*result);
        ret.negativity = false;
        return ret;
    }

    BigInteger* BigInteger::_euclidGcd(BigInteger* a, BigInteger* b)
    {
        if (b->size == 0) return a;
        a->absDivideAndSetThisToRemainder(*b);
        return _euclidGcd(b, a);
    }

    //Set `confidenceFactor` <= 0 to run Deterministic Robin-Miller Primality Test.
    //When testing integers contain more than 1000 bits, it makes it a LITTLE BIT faster to
    //set `confidenceFactor` to a positive number at the expense of
    //the probability of getting a wrong result is no more than 4^(-confidenceFactor)
    bool BigInteger::isPrime(int confidenceFactor) const
    {
        if (0 == size || negativity) return false;
        if (1 == size && _valPtr[0] <= 3) return (_valPtr[0] > 1);
        if ((_valPtr[0] & 1) == 0) return false;
        BigInteger thisMinus1(*this); thisMinus1 += BigInteger(-1);
        double ln_n = log(size * BITS_OF_DWORD);
        int maxTesterForDeterministicResult = (int)(2.0 * ln_n * ln_n);
        BigInteger a(1);
        if (size <= 2) goto qwordTest;
        if (maxTesterForDeterministicResult <= confidenceFactor) goto drmpt;
        if (confidenceFactor > 0)
        {
            qwordTest:	const char kTestNumbers[] = { 2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37 };
            int index = 0;
            do {
                a._valPtr[0] = kTestNumbers[index];
                if (a.compareAbsoluteValueTo(thisMinus1) >= 0) return true;
                if (!_MillerRabinPrimalityTest(a, thisMinus1)) return false;
            } while (++index < sizeof(kTestNumbers) / sizeof(char));
            if (size > 2) {
                confidenceFactor -= (sizeof(kTestNumbers) / sizeof(char));
                std::mt19937 mt(gRandomDevice());
                do {
                    a._valPtr[0] = mt();
                    if (a._valPtr[0] < 41) a._valPtr[0] += 41;
                    if (!_MillerRabinPrimalityTest(a, thisMinus1)) return false;
                } while (--confidenceFactor > 0);
            }
        }
        else
        {
            drmpt:		uint32 base = 2;
            do {
                a._valPtr[0] = base;
                if (!_MillerRabinPrimalityTest(a, thisMinus1)) return false;
            } while (++base <= (uint32)maxTesterForDeterministicResult);
        }
        return true;
    }

    //	http://www.sti15.com/nt/pseudoprimes.html
    bool BigInteger::weakerBailliePSWPrimeTest() const
    {
        if (0 == size || negativity) return false;
        if (1 == size && _valPtr[0] <= 3) return (_valPtr[0] > 1);
        if ((_valPtr[0] & 1) == 0) return false;
        BigInteger thisMinus1(*this); thisMinus1 += BigInteger(-1);
        BigInteger a(2);
        if (!_MillerRabinPrimalityTest(a, thisMinus1)) return false;
        if (_LucasPseudoprimeTest()) {
            a._valPtr[0] = 3;
            if (_MillerRabinPrimalityTest(a, thisMinus1)) return true;
            throw std::logic_error("Severe logic flaw detected during `weakerBailliePSWPrimeTest`, please report this issue.");
        }
        return false;
    }

    bool BigInteger::_MillerRabinPrimalityTest(const BigInteger& a, const BigInteger& thisMinus1) const
    {//logically wrong when *this <= 2
        if (a.modPow(thisMinus1, *this) != kBigInteger_1) return false;
        BigInteger u(thisMinus1);
        do { u >>= 1; } while ((u._valPtr[0] & 1) == 0);
        BigInteger x = a.modPow(u, *this);
        do {
            BigInteger y = _absMultiply(x, x);
            y.absDivideAndSetThisToRemainder(*this);
            if (y == kBigInteger_1 && x != kBigInteger_1 && x != thisMinus1) return false;
            x = std::move(y);
            u <<= 1;
        } while (u < *this);
        if (x != kBigInteger_1) return false;
        return true;
    }

    //Calculating the Jacobi symbol according to the properties at https://en.wikipedia.org/wiki/Jacobi_symbol
    int BigInteger::JacobiSymbol(const BigInteger& upperArgument, const BigInteger& lowerArgument)
    {
        if ((!lowerArgument.testBit(0)) | lowerArgument.negativity)
            throw std::logic_error("lowerArgument of function `JacobiSymbol` must be a positive odd number.");
        BigInteger denominator(lowerArgument);
        BigInteger numerator(upperArgument);
        if (numerator.negativity) {
            numerator.absDivideAndSetThisToRemainder(denominator);
            numerator += denominator;
        }
        return _JacobiSymbolImpl(&numerator, &denominator);
    }

    //	http://math.fau.edu/richman/jacobi.htm
    int BigInteger::_JacobiSymbolImpl(BigInteger* numerator, BigInteger* denominator)
    {
        if (denominator->size == 1 && denominator->_valPtr[0] == 1) return 1;//"Following the normal convention for the empty product, (Z / 1) = 1 for all `Z` "
        if (numerator->size == 0) return 0;
        numerator->absDivideAndSetThisToRemainder(*denominator);
        unsigned countFactorsOf2 = 0;
        for (; false == numerator->testBit(0) && numerator->size != 0; *numerator >>= 1) ++countFactorsOf2;
        if (gcd(*numerator, *denominator) != 1) return 0;
        int ret = 1;
        if (countFactorsOf2 & 1) {
            uint32 m = denominator->_valPtr[0] % 8;
            ret = ((m == 1) | (m == 7)) ? 1 : -1;
        }
        if (denominator->_valPtr[0] % 4 == 3 && numerator->_valPtr[0] % 4 == 3) ret = -ret;
        return ret * _JacobiSymbolImpl(denominator, numerator);
    }

    //	http://stackoverflow.com/questions/38343738/lucas-probable-prime-test
    std::pair<BigInteger, BigInteger> BigInteger::_getModLucasSequence(const BigInteger& k, const BigInteger& D, const BigInteger& P, const BigInteger& Q) const
    {
        BigInteger U_n(1), V_n(P), Q_n(Q);
        if (V_n < 0) { V_n.absDivideAndSetThisToRemainder(*this); V_n += *this; }
        if (Q_n < 0) { Q_n.absDivideAndSetThisToRemainder(*this); Q_n += *this;	}
        BigInteger U(!k.testBit(0) ? 0 : 1);
        BigInteger V(!k.testBit(0) ? 2 : V_n);
        BigInteger n(k);
        n >>= 1;
        while (n.size)
        {
            BigInteger U2 = _absMultiply(U_n, V_n);
            U2.absDivideAndSetThisToRemainder(*this);
            BigInteger V2 = _absMultiply(V_n, V_n);
            V2 -= (Q_n << 1);
            V2.absDivideAndSetThisToRemainder(*this);
            if (V2 < 0) V2 += *this;
            BigInteger Q2 = _absMultiply(Q_n, Q_n);
            Q2.absDivideAndSetThisToRemainder(*this);
            if ((n._valPtr[0] & 1)) {
                BigInteger newU = _absMultiply(U, V2) + _absMultiply(V, U2);
                if (newU.testBit(0)) newU += *this;
                newU >>= 1;
                newU.absDivideAndSetThisToRemainder(*this);
                BigInteger newV = _absMultiply(V, V2) + _absMultiply(U, U2) * D;
                if (newV.testBit(0)) newV += *this;
                newV >>= 1;
                newV.absDivideAndSetThisToRemainder(*this);
                if (newV < 0) newV += *this;
                U = std::move(newU);
                V = std::move(newV);
            }
            U_n = std::move(U2);
            V_n = std::move(V2);
            Q_n = std::move(Q2);
            n >>= 1;
        }
        return std::pair<BigInteger, BigInteger>(U, V);
    }

    bool BigInteger::_LucasPseudoprimeTest() const
    {//logically wrong when *this is small
        int D = 5;
        do {
            if (gcd(D, *this) != 1) return 0;
            if (JacobiSymbol(D, *this) == -1) break;
            D = -D + ((D < 0) ? 2 : -2);
        } while (true);
        const BigInteger Q((1 - D) / 4);
        BigInteger deltaThis(*this); deltaThis += 1;
        std::pair<BigInteger, BigInteger> UV = this->_getModLucasSequence(deltaThis, D, 1, Q);
        return (UV.first.size == 0);
    }

    //Warnning: not suitable when *this has more than 10k bits
    BigInteger BigInteger::nextProbablePrime() const
    {
        if (this->negativity || this->size == 0) return BigInteger(2);
        BigInteger ret(*this);
        if (ret.size == 1)
            while (true) if ((ret += kBigInteger_1).isPrime()) return ret;
        if ((ret._valPtr[0] & 1) == 0) ++ret._valPtr[0];
        else ret += 2;
        const char kTestNumbers[] = { 3, 5, 7, 11, 13, 17, 19, 23 };
        BigInteger m(2);
        static_assert(sizeof(char) == 1, "");
        uint32 modList[sizeof(kTestNumbers)];
        for (int i = 0; i < sizeof(kTestNumbers); ++i) {
            m._valPtr[0] = kTestNumbers[i];
            BigInteger x(ret);
            x.absDivideAndSetThisToRemainder(m);
            modList[i] = x.size ? x._valPtr[0] : 0;
        }
        while (true)
        {
            bool multipleOfTestNumbers = false;
            for (int i = 0; i < sizeof(kTestNumbers); ++i) {
                if (0 == modList[i] % kTestNumbers[i]) multipleOfTestNumbers = true;
                modList[i] += 2;//it may overflow until we find a next prime if *this is very large
            }
            if (!multipleOfTestNumbers && ret.weakerBailliePSWPrimeTest()) return ret;
            ret += 2;
        }
    }

    //for cryptography
    BigInteger::BigInteger(const void* dataPtr, size_t dataSize)
    {
        size = (dataSize + sizeof(uint32) - 1) / sizeof(uint32);
        _valPtr = new uint32[capacity = _getMinimumCapacity(size)];
        memcpy(_valPtr, dataPtr, dataSize);
        memset(reinterpret_cast<char*>(_valPtr) + dataSize, 0, capacity * sizeof(uint32) - dataSize);
        _setSize(size);
    }

    void BigInteger::saveBytesToBuffer(char* dst, size_t writingSize) const
    {
        size_t integerSizeInByte = size * sizeof(uint32);
        if (integerSizeInByte >= writingSize) memcpy(dst, _valPtr, writingSize);
        else {
            memcpy(dst, _valPtr, integerSizeInByte);
            memset(dst + integerSizeInByte, 0, writingSize - integerSizeInByte);
        }
    }

    void BigInteger::setLowerBitsToRandom(uint32 bitLength)
    {
        if (bitLength == 0) return;
        std::mt19937 mt(gRandomDevice());
        uint32 newSize = (bitLength + BITS_OF_DWORD - 1) / BITS_OF_DWORD;
        if (newSize > size)	{
            size = newSize;
            if (size > capacity) _valPtr = new uint32[capacity = _getMinimumCapacity(size)];
            _valPtr[size - 1] = 0;
        }
        uint32 i = 0;
        while (bitLength >= BITS_OF_DWORD) {
            _valPtr[i++] ^= mt();
            bitLength -= BITS_OF_DWORD;
        }
        if (bitLength) _valPtr[i] ^= (mt() & ((1 << bitLength) - 1));
        _setSize(size);
    }

    bool BigInteger::testBit(uint32 bitPos) const
    {
        uint32 dwordIndex = bitPos / BITS_OF_DWORD;
        if (dwordIndex >= size) return false;
        uint32 bitIndex = bitPos % BITS_OF_DWORD;
        return 0 != (_valPtr[dwordIndex] & (1 << bitIndex));
    }

    BigInteger::_SharedMemBigInteger::_SharedMemBigInteger(uint32 size_, uint32* val)
    {
        _valPtr = val;
        _setSize(size_);
        capacity = _getMinimumCapacity(size);
    }


#ifdef FFT_MULTIPLICATION
    bool BigInteger::_isRootsOfUnityInitialized = false;

    namespace BigInteger_FFT_Precedures
    {
        struct Complex {
            double re, im;
            Complex(double _re = 0.0, double _im = 0.0) : re(_re), im(_im) { }
            static Complex add(const Complex& lhs, const Complex& rhs) { return Complex(lhs.re + rhs.re, lhs.im + rhs.im); }
            static Complex subtract(const Complex& lhs, const Complex& rhs) { return Complex(lhs.re - rhs.re, lhs.im - rhs.im); }
            static Complex multiply(const Complex& lhs, const Complex& rhs) { return Complex(lhs.re * rhs.re - lhs.im * rhs.im, lhs.re * rhs.im + lhs.im * rhs.re); }
        };

#define FFT_WORD_SIZE (FFT_WORD_BITLEN / 8)
        /*const union {
        uint32 uint_value;
        char char_value[4];
        } _EndianCheckUnion = { 0x11223344 };
        constexpr bool _IS_LITTLE_ENDIAN() { return _EndianCheckUnion.char_value[0] == 0x44; }*/

        struct Coefficients {
            unsigned short *begin, *end;
            static_assert(sizeof(unsigned short) == FFT_WORD_SIZE, "Length of FFT_WORD is 16bits, which differs from sizeof(unsigned short).");
            uint32 interval;
            uint32 n;
            Coefficients(const BigInteger& src, uint32 n) : interval(1), n(n) { src._FFTgetBeginAndEndOfVal(begin, end); }
            inline Complex getComplexNumber(uint32 i) const
            {
                if (_IS_LITTLE_ENDIAN())
                    return Complex(begin + i < end ? (double)begin[i] : 0.0);
                else {//WARNING: no tests has been done for big endian machines.
                    if (begin + i >= end) return Complex();
                    if ((size_t)begin & FFT_WORD_SIZE) return (double)begin[i - 1];
                    return (double)begin[i + 1];
                }
            }
        };

        struct ValueRepresentation {
            Complex* begin;
            uint32 interval;
            uint32 n;
            ValueRepresentation(Complex* src, uint32 n) : interval(1), n(n), begin(src) { }
            inline const Complex& getComplexNumber(uint32 i) const { return begin[i]; }
        };

        Complex kRootsOfUnity[FFT_MAX_N];
        void initRootsOfUnity()
        {
            const double scale = 6.2831853071795864769 / FFT_MAX_N;
            for (int i = 0; i < FFT_MAX_N; ++i) {
                kRootsOfUnity[i].re = cos(i * scale);
                kRootsOfUnity[i].im = sin(i * scale);
            }
        }

        inline uint32 getIndexOfPower(uint32 omegaIndex, uint32 pow) { return (omegaIndex * pow) % FFT_MAX_N; }

        //`InputType` is either `Coefficients` (isReverse = false)  or  `ValueRepresentation` (isReverse = true for reverse transform)
        template <bool isReverse, typename InputType>
        void FFT(Complex* dst, InputType& src, uint32 omegaIndex)
        {//Warning: src.n cannot be smaller than 4
            if (src.n <= 4) {
                Complex x = src.getComplexNumber(0);
                Complex y = src.getComplexNumber(src.interval * 2);
                Complex buffer0 = Complex::add(x, y);
                Complex buffer1 = Complex::subtract(x, y);
                x = src.getComplexNumber(src.interval);
                y = src.getComplexNumber(src.interval * 3);
                Complex buffer2 = Complex::add(x, y);
                Complex buffer3 = Complex::subtract(x, y);
                Complex wjAo = buffer2;
                dst[0] = Complex::add(buffer0, wjAo);
                dst[2] = Complex::subtract(buffer0, wjAo);
                wjAo = !isReverse ? Complex(-buffer3.im, buffer3.re) : Complex(buffer3.im, -buffer3.re);
                dst[1] = Complex::add(buffer1, wjAo);
                dst[3] = Complex::subtract(buffer1, wjAo);
                return;
            }
            auto srcBeginBackup = src.begin + src.interval;
            src.interval <<= 1;
            src.n >>= 1;
            FFT<isReverse>(dst, src, getIndexOfPower(omegaIndex, 2));//A even
            std::swap(src.begin, srcBeginBackup);
            FFT<isReverse>(dst + src.n, src, getIndexOfPower(omegaIndex, 2));//A odd
            std::swap(src.begin, srcBeginBackup);
            for (uint32 j = 0; j < src.n; ++j) {
                Complex wjAo = Complex::multiply(kRootsOfUnity[getIndexOfPower(omegaIndex, j)], (dst + src.n)[j]);
                (dst + src.n)[j] = Complex::subtract(dst[j], wjAo);
                dst[j] = Complex::add(dst[j], wjAo);
            }
            src.interval >>= 1;
            src.n <<= 1;
        }

    }//namespace BigInteger_FFT_Precedures


    void BigInteger::_FFTMultiply(BigInteger& dst, const BigInteger& lhs, const BigInteger& rhs, void* buffer)
    {
        if ((!lhs.size) | (!rhs.size)) { dst.size = 0; return; }
        const uint32 n = _getMinimumCapacity(lhs.size + rhs.size) * BITS_OF_DWORD / FFT_WORD_BITLEN;
        const uint32 kOmegaIndex = FFT_MAX_N / n;
        if (0 == kOmegaIndex) { dst = _absMultiply(lhs, rhs); return; }
        if (!_isRootsOfUnityInitialized) {//initialize  (1 + 0i) ^ (1/n)
            BigInteger_FFT_Precedures::initRootsOfUnity();
            _isRootsOfUnityInitialized = true;
        }
        bool newMemAlloced = (nullptr == buffer);
        if (newMemAlloced) buffer = operator new(n * 2 * sizeof(BigInteger_FFT_Precedures::Complex));
        BigInteger_FFT_Precedures::Complex* _buffer = reinterpret_cast<BigInteger_FFT_Precedures::Complex*>(buffer);
        //multi-threading benefits little here
        BigInteger_FFT_Precedures::Coefficients A(lhs, n);
        BigInteger_FFT_Precedures::Complex* lhsValues = _buffer;
        BigInteger_FFT_Precedures::FFT<false>(lhsValues, A, kOmegaIndex);
        BigInteger_FFT_Precedures::Coefficients B(rhs, n);
        BigInteger_FFT_Precedures::Complex* rhsValues = lhsValues + n;
        BigInteger_FFT_Precedures::FFT<false>(rhsValues, B, kOmegaIndex);
        for (uint32 j = 0; j < n; ++j, ++rhsValues)
            *rhsValues = BigInteger_FFT_Precedures::Complex::multiply(*lhsValues++, *rhsValues);
        BigInteger_FFT_Precedures::ValueRepresentation C(rhsValues - n, n);//C[i] = A[i] * B[i]
        BigInteger_FFT_Precedures::Complex* products = _buffer;
        BigInteger_FFT_Precedures::FFT<true>(products, C, FFT_MAX_N - kOmegaIndex);

        uint32 newCapa = n / (BITS_OF_DWORD / FFT_WORD_BITLEN);
        if (newCapa > dst.capacity) {
            delete[] dst._valPtr;
            dst._valPtr = new uint32[dst.capacity = newCapa];
        }
        unsigned long long t = 0;
        static_assert(sizeof(unsigned long long) > sizeof(uint32), "");
        const double reciprocal_of_n = 1.0 / n;
        for (uint32 pieceIndex = 0; pieceIndex < newCapa; ++pieceIndex) {
            //double er = abs(round((products)->re / n) - (products)->re / n); if (er > __dbgFFTMaxError) __dbgFFTMaxError = er;
            t += (unsigned long long)((products++)->re * reciprocal_of_n + 0.5);
            dst._valPtr[pieceIndex] = t & ((1 << FFT_WORD_BITLEN) - 1);
            t >>= FFT_WORD_BITLEN;
            //er = abs(round((products)->re / n) - (products)->re / n); if (er > __dbgFFTMaxError) __dbgFFTMaxError = er;
            t += (unsigned long long)((products++)->re * reciprocal_of_n + 0.5);
            dst._valPtr[pieceIndex] |= (uint32(t) << FFT_WORD_BITLEN);
            t >>= FFT_WORD_BITLEN;
        }
        dst._setSize(newCapa);

        if (newMemAlloced) operator delete(buffer);
    }

#endif // FFT_MULTIPLICATION


}//namespace UPMath

int CharToInt(char symbol) {
    if (symbol >= 48 && symbol <= 57) {
        return symbol - 48;
    }
    if (symbol >= 65 && symbol <= 90) {
        return symbol - 55;
    }
    if (symbol >= 97 && symbol <= 122) {
        return symbol - 61;
    }
    if (symbol == 32) {
        return 62;
    }
    if (symbol == 46) {
        return 63;
    }
    return 64;
}

void MultModP(uint64_t base, uint64_t p, std::vector<uint64_t>& result) {
    for (size_t i = 0; i < result.size(); ++i) {
        result[i] *= base;
    }
    for (size_t i = 0; i < result.size() - 1; ++i) {
        if (result[i] >= p) {
            uint64_t remainder = (result[i] - result[i] % p) / p;
            result[i] %= p;
            result[i + 1] += remainder;
        }
    }
    if (result[result.size() - 1] >= p) {
        size_t i = result.size() - 1;
        int64_t remainder = (result[i] - result[i] % p) / p;
        result[i] %= p;
        result.push_back(remainder);
    }
}

void SumModP(uint64_t a, uint64_t p, std::vector<uint64_t>& result) {
    uint64_t sum = a;
    std::vector<uint64_t> number;
    while (sum != 0) {
        uint64_t remainder = sum % p;
        sum /= p;
        number.push_back(remainder);
    }
    for (size_t i = 0; i < number.size(); ++i) {
        if (i < result.size()) {
            result[i] += number[i];
        } else {
            result.push_back(number[i]);
        }
    }
    for (size_t i = 0; i < result.size() - 1; ++i) {
        if (result[i] >= p) {
            uint64_t remainder = (result[i] - result[i] % p) / p;
            result[i] %= p;
            result[i + 1] += remainder;
        }
    }
    if (result[result.size() - 1] >= p) {
        size_t i = result.size() - 1;
        uint64_t remainder = (result[i] - result[i] % p) / p;
        result[i] %= p;
        result.push_back(remainder);
    }
}

std::vector<uint64_t> Base64ToBaseP(std::vector<uint64_t>& Base64Number, uint64_t p) {
    std::vector<uint64_t> BasePNumber;
    if (p == 64) {
        return Base64Number;
    } else {
        size_t i = 0;
        uint64_t value = Base64Number[Base64Number.size() - 1 - i];
        BasePNumber.push_back(value);
        ++i;
        while (i < Base64Number.size()) {
            MultModP(64, p, BasePNumber);
            SumModP(Base64Number[Base64Number.size() - 1 - i], p, BasePNumber);
            ++i;
        }
    }
    return BasePNumber;
}


std::vector<std::pair<UPmath::BigInteger, UPmath::BigInteger>> ElGamalEncrypt(UPmath::BigInteger& p, uint64_t P, const UPmath::BigInteger& g, const UPmath::BigInteger& k, const std::vector<uint64_t>& message) {
    std::vector<std::pair<UPmath::BigInteger, UPmath::BigInteger>> encryptedMessage;
    std::srand(time(NULL));
    for (int64_t value : message) {
        int64_t B = std::rand() % P;
        UPmath::BigInteger degree_of_generator = g.modPow(B, p);
        UPmath::BigInteger EncryptMessage = k.modPow(B, p);
        UPmath::BigInteger val(std::to_string(value));
        EncryptMessage = EncryptMessage * val;
        EncryptMessage %= p;
        encryptedMessage.emplace_back(degree_of_generator, EncryptMessage);
    }

    return encryptedMessage;
}


int main() {
    uint64_t P, G, K;
    std::cin >> P >> G >> K;
    std::cin.ignore();
    std::string sms;
    std::getline(std::cin, sms);
    UPmath::BigInteger p(P);
    UPmath::BigInteger g(G);
    UPmath::BigInteger k(K);

    std::vector<uint64_t> message;
    for (char symbol : sms) {
        message.push_back(CharToInt(symbol));
    }


    std::vector<uint64_t> group_message = Base64ToBaseP(message, P);
    std::vector<std::pair<UPmath::BigInteger, UPmath::BigInteger>> encrypted_messages = ElGamalEncrypt(p, P, g, k, group_message);
//    std::vector<std::pair<uint64_t, uint64_t>> encrypted_messages = ElGamalEncrypt(P, G, K, group_message);

//    for (const auto& encryption : encrypted_messages) {
//        std::cout << encryption.first << " " << encryption.second << "\n";
//    }
//
////
//
//
    for (const auto& encryption : encrypted_messages) {
        std::cout << encryption.first.toString() << " " << encryption.second.toString() << "\n";
    }

////


}


//    uint64_t number = 0;
//    for (uint64_t i = 0; i < result.size(); ++i) {
//        number = number * 10 + result[result.size() - 1 - i];
//    }
//    return number;
//}

//    uint64_t p;
//    std::cin >> p;
//    std::vector<std::vector<uint64_t>> mult_in_base_p;
//    mult_in_base_p.assign(p, std::vector<uint64_t>(p, 0));
//    for (uint64_t i = 0; i < p; ++i) {
//        for (uint64_t j = i; j < p; ++j) {
//            mult_in_base_p[i][j] = MultModP(i, j, p);
//        }
//    }
//    for (uint64_t j = 0; j < p; ++j) {
//        for (uint64_t i = j; i < p; ++i) {
//            mult_in_base_p[i][j] = mult_in_base_p[j][i];
//        }
//    }
//    for (int i = 1; i <= p; ++i) {
//        for (int j = 1; j <= p; ++j) {
//            std::cout << mult_in_base_p[i - 1][j - 1] << "\t";
//        }
//        std::cout << std::endl;
//    }



//Solution

//    uint64_t P, G, K;
//    std::cin >> P >> G >> K;
//    std::cin.ignore();
//    std::string sms;
//    std::getline(std::cin, sms);
//    UPmath::BigInteger p(P);
//    UPmath::BigInteger g(G);
//    UPmath::BigInteger k(K);
//
//    std::vector<uint64_t> message;
//    for (char symbol : sms) {
//        message.push_back(CharToInt(symbol));
//    }


//    std::vector<uint64_t> message = {60, 36, 12};
//    uint64_t p = 83;
//    std::vector<uint64_t> mess = Base64ToBaseP(message, p);
//    for (size_t i = 0; i < mess.size(); ++i) {
//        std::cout << mess[mess.size() - 1 - i] << ' ';
//    }
//
//