Re: constexpr is really very smart!

Liste des GroupesRevenir à cl c++ 
Sujet : Re: constexpr is really very smart!
De : already5chosen (at) *nospam* yahoo.com (Michael S)
Groupes : comp.lang.c++
Date : 20. Dec 2024, 01:17:03
Autres entêtes
Organisation : A noiseless patient Spider
Message-ID : <20241220021703.00003e9e@yahoo.com>
References : 1 2 3 4 5 6
User-Agent : Claws Mail 4.1.1 (GTK 3.24.34; x86_64-w64-mingw32)
On Thu, 19 Dec 2024 00:40:32 -0800
Tim Rentsch <tr.17687@z991.linuxsc.com> wrote:

Michael S <already5chosen@yahoo.com> writes:
 
On Wed, 18 Dec 2024 01:33:42 +0200
Michael S <already5chosen@yahoo.com> wrote:
 
I am pretty sure that better variant exists, but for tonight it's
enough. 
>
Morning fibs.
>
Recursive:
>
struct pair_t {
  long long a,b;
};
>
// return .a=fib(n), .b=fib(n+1)
static struct pair_t fib2(long n)
{
  if (n <= 0) {
    struct pair_t ret = { .a = 0, .b = n==0 ? 1 : 0 };
    return ret;
  }
  // for n > 0
  // fib(n*2)   = (fib(n-1)+fib(n+1))*fib(n)
  // fib(n*2+1) = fib(n)*fib(n)+fib(n+1)*fib(n+1)
  long m = (n-1) >> 1;
  struct pair_t ret = fib2(m);
  long long a = ret.a, b = ret.b;
  long long c = a + b;
  if ((n & 1)==0) { // (m,m+1) => ((m+1)*2,(m+1)*2*2+1)
    ret.a = (a + c)*b;
    ret.b = b*b + c*c;
  } else {          // (m,m+1) => (m*2+1,(m+1)*2)
    ret.a = a*a + b*b;
    ret.b = (a + c)*b;
  }
  return ret;
}
>
static long long fib(long n)
{
  struct pair_t x = fib2(n-1);
  return x.b;

 
I may have inadvertently misled you.  Here is a simple linear
formulation that uses recursion:
 
    typedef  unsigned long long  NN;
 
    static NN fibonacci_3( NN, NN, unsigned );
 
    NN
    fibonacci( unsigned n ){
        return  fibonacci_3( 1, 0, n );
    }
 
    NN
    fibonacci_3( NN a, NN b, unsigned n ){
        return  n == 0  ? b  :  fibonacci_3( b, a+b, n-1 );
    }
 
Can you apply this idea to your logarithmic scheme?
 
 

Not really. But I can cheat.
Below is slightly modified iterative algorithm coded as recursion.

static
long long fib_3(long long a, long long b, unsigned long rev_n) {
  if (rev_n == 1)
    return b;

  long long c = a + b;
  if ((rev_n & 1) == 0)
    return fib_3(a*a+b*b, (a+c)*b, rev_n/2);
  else
    return fib_3((a+c)*b, b*b+c*c, rev_n/2);
}

static
long long fib_bitrev(unsigned long a, unsigned long b) {
  return a==1 ? fib_3(0, 1, b) : fib_bitrev( a / 2, b * 2 + a % 2);
}

long long fib(long n) {
  return n <= 0 ? 0 : fib_bitrev(n, 1);
}












Date Sujet#  Auteur
15 Dec 24 * constexpr is really very smart!39Student Project
16 Dec 24 +* Re: constexpr is really very smart!35Michael S
16 Dec 24 i+* Re: constexpr is really very smart!8David Brown
16 Dec 24 ii`* Re: constexpr is really very smart!7Michael S
16 Dec 24 ii +* Re: constexpr is really very smart!3David Brown
17 Dec 24 ii i`* Re: constexpr is really very smart!2Michael S
17 Dec 24 ii i `- Re: constexpr is really very smart!1David Brown
18 Dec 24 ii `* Re: constexpr is really very smart!3Tim Rentsch
18 Dec 24 ii  `* Re: constexpr is really very smart!2Michael S
18 Dec 24 ii   `- Re: constexpr is really very smart!1Tim Rentsch
17 Dec 24 i`* Re: constexpr is really very smart!26Tim Rentsch
18 Dec 24 i `* Re: constexpr is really very smart!25Michael S
18 Dec 24 i  +* Re: constexpr is really very smart!13Michael S
19 Dec 24 i  i+* Re: constexpr is really very smart!11Tim Rentsch
20 Dec 24 i  ii`* Re: constexpr is really very smart!10Michael S
20 Dec 24 i  ii `* Re: constexpr is really very smart!9Tim Rentsch
21 Dec 24 i  ii  `* Re: constexpr is really very smart!8Michael S
22 Dec 24 i  ii   `* Re: constexpr is really very smart!7Tim Rentsch
23 Dec 24 i  ii    `* Re: constexpr is really very smart!6Michael S
24 Dec 24 i  ii     `* Re: constexpr is really very smart!5Tim Rentsch
25 Dec 24 i  ii      `* Bignum multiplication in Python vs GMP (was: constexpr is really very smart!)4Michael S
25 Dec 24 i  ii       +* Re: Bignum multiplication in Python vs GMP (was: constexpr is really very smart!)2Michael S
27 Dec 24 i  ii       i`- Re: Bignum multiplication in Python vs GMP (was: constexpr is really very smart!)1Tim Rentsch
27 Dec 24 i  ii       `- Re: Bignum multiplication in Python vs GMP (was: constexpr is really very smart!)1Tim Rentsch
19 Dec 24 i  i`- Re: constexpr is really very smart!1Student Project
18 Dec 24 i  `* Re: constexpr is really very smart!11Tim Rentsch
18 Dec 24 i   `* Re: constexpr is really very smart!10Michael S
19 Dec 24 i    `* Re: constexpr is really very smart!9Tim Rentsch
19 Dec 24 i     `* Re: constexpr is really very smart!8Michael S
20 Dec 24 i      `* Re: constexpr is really very smart!7Tim Rentsch
21 Dec 24 i       `* Re: constexpr is really very smart!6Michael S
21 Dec 24 i        +* Re: constexpr is really very smart!2James Kuyper
22 Dec 24 i        i`- Re: constexpr is really very smart!1Michael S
22 Dec 24 i        `* Re: constexpr is really very smart!3Tim Rentsch
22 Dec 24 i         `* Re: constexpr is really very smart!2Michael S
23 Dec 24 i          `- Re: constexpr is really very smart!1Tim Rentsch
17 Dec 24 `* Re: constexpr is really very smart!3Keith Thompson
18 Dec 24  `* Re: constexpr is really very smart!2Student Project
18 Dec 24   `- Re: constexpr is really very smart!1Michael S

Haut de la page

Les messages affichés proviennent d'usenet.

NewsPortal