TorchLean API

Docs HomeInstallationGuideExamplesGraphsTools

FloatLib.Floats.Formats.DecimalInterchange.DPD.Declet

Densely packed decimal declets #

The eight cases below are Cowlishaw's DPD encoding table, also specified by IEEE 754-2019 ยง3.5.2(c)(1). Three decimal digits occupy ten bits. Decoding accepts all 1024 bit patterns; the 24 redundant patterns select the same digits as their canonical encodings.

Source: https://speleotrove.com/decimal/DPDecimal.html, encoding and decoding tables.

source
def FloatLib.Floats.Formats.DecimalInterchange.DPD.encodeDeclet (n : Nat) :
Nat

Encode three decimal digits. The input contract is n < 1000.

Instances For
    source
    def FloatLib.Floats.Formats.DecimalInterchange.DPD.decodeDeclet (n : Nat) :
    Nat

    Decode a ten-bit declet, including its redundant encodings.

    Instances For