Conversation

Notices

1. Adrian Cochrane (alcinnz@floss.social)'s status on Friday, 24-Mar-2023 02:12:26 JST Adrian Cochrane

   Reading through the rest of GNU Multiprecisions MPZ sublibrary, I see:

   * Wrappers around initialization & conversion from
   - doubles
   - another MPZ struct
   - signed longs
   - unsigned longs
   * Initialization multiple MPZs via variadic arguments
   * Initialize a single MPZ struct by allocating its digits or set a dummy value
   * Divide 1 by a given MPZ via `mpz_gcdext` with postprocessing
   * Inclusive bitwise-or, reallocating where ncessary as it turns it into a bitwise-and operation

   1/?

   • Elias Mårtenson (loke@functional.cafe)'s status on Friday, 24-Mar-2023 02:27:00 JST Elias Mårtenson

     in reply to

     @alcinnz yeah, I'm using it to build a multiplatform bignum api for Kotlin, and I've come to realise that the end result may be convenient (blends well with Kotlin), but there are so many features of GMP that I'm not taking advantage of.

   • Adrian Cochrane (alcinnz@floss.social)'s status on Friday, 24-Mar-2023 02:50:57 JST Adrian Cochrane

     in reply to

     * Conversion from strings without whitespace
     * Conversion from a file
     * Compute odd factors of factorials (I've also just finished skimming other factorizing code) using prime sieves, other factorizing infrastructuring, & repeated multiplication
     * Conversion to
     - doubles
     - signed types via generic template
     - signed long ints
     - strings
     * Accessor for the base2 exponent field as a double
     * Wrapper around MPN's greatest-common-divisor implementation adding memory management

     2/4?

   • Adrian Cochrane (alcinnz@floss.social)'s status on Friday, 24-Mar-2023 02:55:32 JST Adrian Cochrane

     in reply to

     * Fibonacci numbers with fastpaths
     * Wrapper around MPN's alternate Fibonacci implementation
     * Wrapper around MPN's alternate greatest-common-divisor implementation
     * Wrapper around MPN's main greatest-common-divisor implementation with own fastpaths
     * Multiple wrappers around MPN division with its own fastpaths
     * Set digits of a MPZ struct
     * Output to stdout via string serializer
     * Test whether the MPZ number is divisible by a given factor, different functions for different factor types

     3/4!

   • Adrian Cochrane (alcinnz@floss.social)'s status on Friday, 24-Mar-2023 03:13:02 JST Adrian Cochrane

     in reply to

     * MPZ to a unsigned-long power
     * Compute factorials via lookup-table optimizations
     * Get digits
     * Compute hamming distance, with own fastpaths
     * Test "congruence" against a base2-exponential as a variation around comparisons, a couple routines for this
     * Free digits from MPZs passed variadic arguments
     * Clear digits of a given MPZ
     * Wrappers around MPN comparison
     * Clear a particular bit & normalize
     * Flip a given bit
     * Flip all bits

     4/5!!

   • Adrian Cochrane (alcinnz@floss.social)'s status on Friday, 24-Mar-2023 03:23:49 JST Adrian Cochrane

     in reply to

     * Take the absolute value
     * Initialize array of MPZs
     * Compute a semi-factorial which fits in a single digit MPN digit via a selection of codepaths & possibly a lookuptable
     * Multiple common template for wrapping MPN addition & subtraction to select an inner function based on sign & adding memory management
     - One of these adds multiplication in the mix
     * Bitwise and with memory management & multiple codepaths
     * Wrappers taking unsigned long operands

     5/5 Fin for today!