TorchLean API

Docs HomeInstallationGuideExamplesGraphsTools

FloatLib.Floats.Formats.Codebook.Arithmetic.Runtime

Executable arithmetic for catalog codebooks #

The illustrative catalog encodings have direct executable bit kernels. Refinement against their numerical systems is proved in Arithmetic.Proof.

source
@[inline]
def FloatLib.Floats.Formats.Codebook.Catalog.bipolar1.neg (x : bipolar1.Code) :
bipolar1.Code

Exact additive inverse for the one-bit bipolar codebook.

Instances For
    source
    @[inline]
    def FloatLib.Floats.Formats.Codebook.Catalog.bipolar1.mul (x y : bipolar1.Code) :
    bipolar1.Code

    Exact multiplication for the one-bit bipolar codebook.

    Instances For
      source
      @[inline]
      def FloatLib.Floats.Formats.Codebook.Catalog.ternary2.neg? (x : ternary2.Code) :
      Option ternary2.Code

      Exact ternary negation, rejecting the reserved word.

      Instances For
        source
        @[inline]
        def FloatLib.Floats.Formats.Codebook.Catalog.ternary2.mul? (x y : ternary2.Code) :
        Option ternary2.Code

        Exact ternary multiplication, returning none if either operand is the reserved word.

        Instances For