LA_library/bitvector.cc

678 lines
14 KiB
C++

/*
LA: linear algebra C++ interface library
Copyright (C) 2008-2023 Jiri Pittner <jiri.pittner@jh-inst.cas.cz> or <jiri@pittnerovi.com>
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "bitvector.h"
#include <unistd.h>
#include "numbers.h"
namespace LA {
//inefficient I/O operators
std::ostream & operator<<(std::ostream &s, const bitvector &x)
{
for(unsigned int i=0; i<x.size(); ++i) s<< x[i];
return s;
}
std::istream & operator>>(std::istream &s, bitvector &x)
{
std::string str;
s >> str;
x.resize(str.size());
for(unsigned int i=0; i<x.size(); ++i) {x.assign(i,str[i]!='0');}
return s;
}
std::ostream & operator<<(std::ostream &s, const bitmatrix &x)
{
s<<x.nrows()<<" "<<x.ncols()<<std::endl;
for(unsigned int i=0; i<x.nrows(); ++i)
{
for(unsigned int j=0; j<x.ncols(); ++j) s<< x(i,j);
s<<std::endl;
}
return s;
}
std::istream & operator>>(std::istream &s, bitmatrix &x)
{
unsigned int n,m;
s >> n >> m;
x.resize(n,m);
for(unsigned int i=0; i<n; ++i)
{
for(unsigned int j=0; j<m; ++j) {int z; s>>z; if(z) x.set(i,j); else x.reset(i,j);}
}
return s;
}
void bitvector::zero_padding() const
{
if(!modulo) return;
bitvector *p = const_cast<bitvector *>(this);
p->v[nn-1] &= (1ULL<<modulo)-1;
}
bitvector& bitvector::operator++()
{
copyonwrite();
zero_padding();
int i=0;
while(i<nn) if(++v[i++]) break;
return *this;
}
bitvector& bitvector::operator--()
{
copyonwrite();
zero_padding();
int i=0;
while(i<nn) if(v[i++]--) break;
return *this;
}
//implemented so that vectors of different length are considered different automatically
bool bitvector::operator!=(const bitvector &rhs) const
{
if(nn==rhs.nn && modulo==rhs.modulo && v==rhs.v) return false;
zero_padding();
rhs.zero_padding();
int minnn=nn; if(rhs.nn<minnn) minnn=rhs.nn;
int maxnn=nn; if(rhs.nn>maxnn) maxnn=rhs.nn;
if(memcmp(v,rhs.v,minnn*sizeof(bitvector_block))) return true;
if(minnn==maxnn) return false;
if(nn==minnn) {for(int i=minnn; i<maxnn; ++i) if(rhs.v[i]) return true;}
if(rhs.nn==minnn){for(int i=minnn; i<maxnn; ++i) if(v[i]) return true;}
return false;
}
bool bitvector::operator>(const bitvector &rhs) const
{
if(nn!=rhs.nn || modulo!=rhs.modulo) laerror("at the moment only bitvectors of the same length comparable");
if(v==rhs.v) return 0;
if(!modulo) return memcmp(v,rhs.v,nn*sizeof(bitvector_block)>0);
int r;
if((r=memcmp(v,rhs.v,(nn-1)*sizeof(bitvector_block)))) return r>0;
bitvector_block a=v[nn-1];
bitvector_block b=rhs.v[nn-1];
//zero out the irrelevant bits
bitvector_block mask= ~((bitvector_block)0);
mask <<=modulo;
mask = ~mask;
a&=mask; b&=mask;
return a>b;
}
bool bitvector::operator<(const bitvector &rhs) const
{
if(nn!=rhs.nn || modulo!=rhs.modulo) laerror("at the moment only bitvectors of the same length comparable");
if(v==rhs.v) return 0;
if(!modulo) return memcmp(v,rhs.v,nn*sizeof(bitvector_block)<0);
int r;
if((r=memcmp(v,rhs.v,(nn-1)*sizeof(bitvector_block)))) return r<0;
bitvector_block a=v[nn-1];
bitvector_block b=rhs.v[nn-1];
//zero out the irrelevant bits
bitvector_block mask= ~((bitvector_block)0);
mask <<=modulo;
mask = ~mask;
a&=mask; b&=mask;
return a<b;
}
bitvector bitvector::operator~() const
{
bitvector r((*this).size());
for(int i=0; i<nn; ++i) r.v[i] = ~v[i];
r.zero_padding();
return r;
}
bitvector& bitvector::operator&=(const bitvector &rhs)
{
if(size()<rhs.size()) resize(rhs.size(),true);
copyonwrite();
for(int i=0; i<nn; ++i) v[i] &= (i>=rhs.nn? 0 : rhs.v[i]);
return *this;
}
bitvector& bitvector::operator|=(const bitvector &rhs)
{
if(size()<rhs.size()) resize(rhs.size(),true);
copyonwrite();
for(int i=0; i<nn && i<rhs.nn; ++i) v[i] |= rhs.v[i];
return *this;
}
bitvector& bitvector::operator^=(const bitvector &rhs)
{
if(size()<rhs.size()) resize(rhs.size(),true);
copyonwrite();
for(int i=0; i<nn && i<rhs.nn; ++i) v[i] ^= rhs.v[i];
return *this;
}
/*number of ones in a binary number, from "Hacker's delight" book*/
#ifdef LONG_IS_32
static unsigned int word_popul(unsigned long x)
{
x -= ((x>>1)&0x55555555);
x = (x&0x33333333) + ((x>>2)&0x33333333);
x=(x + (x>>4))&0x0f0f0f0f;
x+= (x>>8);
x+= (x>>16);
return x&0x3f;
}
#else
//@@@@ use an efficient trick too
static unsigned int word_popul(unsigned long x)
{
unsigned int s=0;
for(int i=0; i<64; ++i)
{
if(x&1) ++s;
x>>=1;
}
return s;
}
#endif
bitvector& bitvector::operator>>=(unsigned int i)
{
if(i==0) return *this;
copyonwrite();
unsigned int imod = i%blockbits;
unsigned int ishift = i/blockbits;
for(int dest=0; dest<nn; ++dest)
{
int src=dest+ishift;
if(src>=nn) v[dest]=0;
else
{
v[dest] = v[src]>>imod;
if(imod && (src+1<nn)) v[dest] |= (v[src+1]&((1ULL<<imod)-1)) <<(blockbits-imod);
}
}
return *this;
}
bitvector& bitvector::leftshift(unsigned int i, bool autoresize)
{
if(i==0) return *this;
copyonwrite();
unsigned int imod = i%blockbits;
unsigned int ishift = i/blockbits;
if(autoresize) resize(size()+i,true);
for(int dest=nn-1; dest>=0; --dest)
{
int src=dest-ishift;
if(src<0) v[dest]=0;
else
{
v[dest] = v[src]<<imod;
if(imod && (src-1>=0)) v[dest] |= (v[src-1]& (((1ULL<<imod)-1) <<(blockbits-imod)))>>(blockbits-imod);
}
}
return *this;
}
void bitvector::randomize()
{
copyonwrite();
for(int i=0; i<nn; ++i) v[i]=RANDINT64();
zero_padding();
}
unsigned int bitvector::population(const unsigned int before) const
{
if(before) laerror("before parameter in population() not implemented yet");
int i;
unsigned int s=0;
for(i=0; i<nn-1; ++i) s+=word_popul(v[i]);
bitvector_block a=v[nn-1];
if(modulo)
{
bitvector_block mask= ~((bitvector_block)0);
mask <<=modulo;
a &= ~mask;
}
return s+word_popul(a);
}
unsigned int bitvector::bitdiff(const bitvector &y) const
{
if(nn!=y.nn) laerror("incompatible size in bitdifference");
unsigned int s=0;
for(int i=0; i<nn-1; ++i) s+=word_popul(v[i]^y.v[i]);
bitvector_block a=v[nn-1]^y.v[nn-1];
if(modulo)
{
bitvector_block mask= ~((bitvector_block)0);
mask <<=modulo;
a &= ~mask;
}
return s+word_popul(a);
}
static unsigned int nlz64(uint64_t x0)
{
int64_t x=x0;
uint64_t y;
unsigned int n;
n=0;
y=x;
L: if ( x<0) return n;
if(y==0) return 64-n;
++n;
x<<=1;
y>>=1;
goto L;
}
static unsigned int ntz64(uint64_t x)
{
unsigned int n;
if(x==0) return 64;
n=1;
if((x&0xffffffff)==0) {n+=32; x>>=32;}
if((x&0xffff)==0) {n+=16; x>>=16;}
if((x&0xff)==0) {n+=8; x>>=8;}
if((x&0xf)==0) {n+=4; x>>=4;}
if((x&0x3)==0) {n+=2; x>>=2;}
return n-(x&1);
}
unsigned int bitvector::nlz() const
{
int leadblock=nn-1;
unsigned int n=0;
while(leadblock>0 && v[leadblock] == 0)
{
--leadblock;
n+=blockbits;
}
n+= nlz64(v[leadblock]);
if(modulo) n-= blockbits-modulo;
return n;
}
unsigned int bitvector::ntz() const
{
int tailblock=0;
unsigned int n=0;
if(iszero()) return size();
while(tailblock<nn-1 && v[tailblock] == 0)
{
++tailblock;
n+=blockbits;
}
n+= ntz64(v[tailblock]);
return n;
}
//NOTE: naive algorithm, just for testing
//does not perform modulo irreducible polynomial, is NOT GF(2^n) multiplication
bitvector bitvector::multiply(const bitvector &rhs, bool autoresize) const
{
int maxsize=size(); if(rhs.size()>maxsize) maxsize=rhs.size();
bitvector r(autoresize?size()+rhs.size():maxsize);
r.clear();
bitvector tmp(rhs);
if(autoresize) tmp.resize(size()+rhs.size(),true);
for(int i=0; i<=degree(); ++i)
{
if((*this)[i]) r+= tmp;
tmp.leftshift(1,false);
}
return r;
}
//this is GF(2^n) multiplication
bitvector bitvector::field_mult(const bitvector &rhs, const bitvector &irpolynom) const
{
int d=irpolynom.degree();
if(d>size()||d>rhs.size()) laerror("inconsistent dimensions in field_mult");
bitvector r(size());
r.clear();
bitvector tmp(*this);
tmp.resize(size()+1,true);
tmp.copyonwrite();
int rd=rhs.degree();
for(int i=0; i<=rd; ++i) //avoid making a working copy of rhs and shifting it
{
if(rhs[i]) r+= tmp;
tmp.leftshift(1,false);
if(tmp[d]) tmp -= irpolynom;
}
return r;
}
//this is GF(2^n) multiplicative inverseion
//cf. https://en.wikipedia.org/wiki/Extended_Euclidean_algorithm
bitvector bitvector::field_inv(const bitvector &irpolynom) const
{
int d=irpolynom.degree();
if(d>size()) laerror("inconsistent dimensions in field_inv");
bitvector t(size()); t.clear();
bitvector newt(size()); newt.clear(); newt.set(0);
bitvector r(irpolynom); r.copyonwrite();
bitvector newr(*this); if(r.size()>newr.size()) newr.resize(r.size(),true); newr.copyonwrite();
int rs=r.size();
while(!newr.is_zero())
{
//std::cout <<"r "<<r<<" newr "<<newr <<" "; std::cout <<"t "<<t<<" newt "<<newt; std::cout <<std::endl;
bitvector remainder(rs);
bitvector quotient = r.division(newr,remainder);
r=newr; newr=remainder;
remainder= t - quotient.multiply(newt,false); //avoid size growth
t=newt; newt=remainder;
}
if(r.degree()>0) laerror("field_inv: polynomial is not irreducible or input is zero modulo the polynomial");
if(!r[0]) laerror("zero in field_inv");
return t;
}
void bitvector::resize(const unsigned int n, bool preserve)
{
if(preserve) zero_padding();
int oldnn=nn;
NRVec<bitvector_block>::resize((n+blockbits-1)/blockbits,preserve);
modulo=n%blockbits;
if(preserve) //clear newly allocated memory
{
for(int i=oldnn; i<nn; ++i) v[i]=0;
}
else clear();
}
bitvector bitvector::division(const bitvector &rhs, bitvector &remainder) const
{
if(rhs.is_zero()) laerror("division by zero binary polynomial");
if(is_zero() || rhs.is_one()) {remainder.clear(); return *this;}
bitvector r(size());
r.clear();
remainder= *this;
remainder.copyonwrite();
int rhsd = rhs.degree();
int d;
while((d=remainder.degree()) >= rhsd)
{
unsigned int pos = d-rhsd;
r.set(pos);
remainder -= rhs<<pos;
}
remainder.resize(rhs.size(),true);
return r;
}
bitvector bitvector::gcd(const bitvector &rhs) const
{
bitvector big,small;
if(degree()>=rhs.degree())
{big= *this; small=rhs;}
else
{big=rhs; small= *this;}
if(big.is_zero())
{
if(small.is_zero()) laerror("two zero arguments in gcd");
return small;
}
if(small.is_zero()) return big;
if(small.is_one()) return small;
if(big.is_one()) return big;
do {
bitvector help=small;
small= big%small;
big=help;
}
while(! small.is_zero());
return big;
}
//cf. Brent & Zimmermann ANZMC08t (2008) paper
//
bool bitvector::is_irreducible() const
{
bitvector tmp(size());
tmp.clear(); tmp.set(1);
unsigned int d=degree();
//repeated squaring test
for(unsigned int j=0; j<d; ++j) tmp = tmp.field_mult(tmp,*this);
tmp.flip(1);
if(!tmp.is_zero()) return false;
FACTORIZATION<uint64_t> f = factorization((uint64_t)d);
if(f.begin()->first==d) return true; //d was prime
//additional tests needed for non-prime degrees
for(auto p=f.begin(); p!=f.end(); ++p)
{
unsigned int dm= d / p->first;
tmp.clear(); tmp.set(1);
for(unsigned int j=0; j<dm; ++j) tmp = tmp.field_mult(tmp,*this);
tmp.flip(1);
bitvector g=tmp.gcd(*this);
//std::cout << "TEST tmp, ir, gcd, is_one "<<tmp<<" "<<*this<<" "<<g<<" : "<<g.is_one()<<std::endl;
if(!g.is_one()) return false;
}
return true;
}
//horner scheme
bitvector bitvector::composition(const bitvector &x) const
{
bitvector r(size());
r.clear();
int d=degree();
for(int i=d; i>0; --i)
{
if((*this)[i]) r.flip(0);
r*=x;
}
if((*this)[0]) r.flip(0);
return r;
}
bitvector bitvector::field_composition(const bitvector &x, const bitvector &ir) const
{
bitvector r(size());
r.clear();
int d=degree();
for(int i=d; i>0; --i)
{
if((*this)[i]) r.flip(0);
r= r.field_mult(x,ir);
}
if((*this)[0]) r.flip(0);
return r;
}
void bitvector::read(int fd, bool dimensions, bool transp)
{
if(dimensions)
{
int r = ::read(fd,&modulo,sizeof(modulo));
if(r!=sizeof(modulo)) laerror("cannot read in bitvector");
}
NRVec<bitvector_block>::get(fd,dimensions,transp);
}
void bitvector::write(int fd, bool dimensions, bool transp)
{
if(dimensions)
{
int r = ::write(fd,&modulo,sizeof(modulo));
if(r!=sizeof(modulo)) laerror("cannot write in bitvector");
}
NRVec<bitvector_block>::put(fd,dimensions,transp);
}
static bitvector *tryme;
static bool irfound;
static int mynth;
static void irfinder(int nones, int top)
{
if(irfound) return;
if(nones==0) //terminate recursion
{
bool testit = tryme->is_irreducible();
//std::cout <<"candidate = "<< *tryme<< " result = "<<testit<<std::endl;
if(testit)
{
--mynth;
if(!mynth) irfound=true;
}
return;
}
for(int i=nones; i<=top; ++i)
{
tryme->set(i);
irfinder(nones-1,i-1);
if(irfound) break;
else tryme->reset(i);
}
}
bitvector find_irreducible(int deg, int pop, int nth)
{
if(deg<=0) laerror("illegal degree in find_irreducible");
if(deg==1) {bitvector r(2); r.set(1); r.reset(0); return r;}
bitvector r(deg+1);
if(pop== -1)
{
do {
r.randomize();
r.set(0);
r.set(deg);
if((r.population()&1)==0) r.flip(1+RANDINT32()%(deg-2));
}
while(!r.is_irreducible());
return r;
}
if(pop<3 || (pop&1)==0) laerror("impossible population of irreducible polynomial requested");
r.clear();
r.set(0);
r.set(deg);
pop-=2;
tryme= &r;
irfound=false;
mynth=nth;
irfinder(pop,deg-1);
if(!irfound) r.clear();
return r;
}
bitvector bitvector::pow(unsigned int n) const
{
if(n==0) {bitvector r(size()); r.clear(); r.set(0); return r;}
if(n==1) return *this;
bitvector y,z;
z= *this;
while(!(n&1))
{
z = z*z;
n >>= 1;
}
y=z;
while((n >>= 1)/*!=0*/)
{
z = z*z;
if(n&1) y *= z;
}
return y;
}
bitvector bitvector::field_pow(unsigned int n, const bitvector &ir) const
{
if(n==0) {bitvector r(size()); r.clear(); r.set(0); return r;}
if(n==1) return *this;
bitvector y,z;
z= *this;
while(!(n&1))
{
z = z.field_mult(z,ir);
n >>= 1;
}
y=z;
while((n >>= 1)/*!=0*/)
{
z = z.field_mult(z,ir);
if(n&1) y = y.field_mult(z,ir);
}
return y;
}
//sqrt(x) is x^(2^(d-1))
bitvector bitvector::field_sqrt(const bitvector &ir) const
{
int d=ir.degree();
bitvector r(*this);
for(int i=0; i<d-1; ++i)
{
r= r.field_mult(r,ir);
}
r.resize(d+1,true);
return r;
}
}//namespace