• Re: Last 8 digits of fib(1000000) [Fateman still offers Chocolat Bars]

    From Mild Shock@21:1/5 to Mild Shock on Tue Mar 18 09:45:02 2025
    But I don't know what the future on browser side,
    i.e. client side CAS will be. If you ask ChatGPT
    what the last 8 digits of fib(1000000) are,

    it generates a Python program on the fly, and
    executes it on the fly on its server, and then
    displays the correct result. Quite amazing.

    On the other hand you currently you cannot ask
    DeepSeek or Grok. Well you can ask, but they
    don't produce the correct number, and its

    not clear how they do the number.

    - ChatGPT: Ok
    - DeepSeek: Nok
    - Grok: Nok

    Mild Shock schrieb:
    Hi,

    Not sure what a CAS would do, but this looks
    nice. Some measurement from a browser:

    ?- time((fib(1000000, _X), Y is _X mod 10^8)).
    % Zeit 28 ms, GC 0 ms, Lips 88857, Uhr 16.03.2025 22:48
    Y = 42546875

    The 28 ms execution time are not bad, since modulo
    was not integrated into matexp/3, making it to
    compute the full fibonacci(1000000) before taking

    the modulo. Not sure whether JavaScript bigint
    is faster or slower than GMP ?

    Bye

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Mild Shock@21:1/5 to All on Tue Mar 18 09:36:24 2025
    Hi,

    Not sure what a CAS would do, but this looks
    nice. Some measurement from a browser:

    ?- time((fib(1000000, _X), Y is _X mod 10^8)).
    % Zeit 28 ms, GC 0 ms, Lips 88857, Uhr 16.03.2025 22:48
    Y = 42546875

    The 28 ms execution time are not bad, since modulo
    was not integrated into matexp/3, making it to
    compute the full fibonacci(1000000) before taking

    the modulo. Not sure whether JavaScript bigint
    is faster or slower than GMP ?

    Bye

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From anthk@21:1/5 to Mild Shock on Wed Jun 11 21:06:30 2025
    On 2025-03-18, Mild Shock <[email protected]> wrote:
    Hi,

    Not sure what a CAS would do, but this looks
    nice. Some measurement from a browser:

    ?- time((fib(1000000, _X), Y is _X mod 10^8)).
    % Zeit 28 ms, GC 0 ms, Lips 88857, Uhr 16.03.2025 22:48
    Y = 42546875

    The 28 ms execution time are not bad, since modulo
    was not integrated into matexp/3, making it to
    compute the full fibonacci(1000000) before taking

    the modulo. Not sure whether JavaScript bigint
    is faster or slower than GMP ?

    Bye

    Try Maxima or raw SBCL computing a lisp function.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)