TorchLean API

FloatLib.Floats.Formats.OCP.FP8.E5M2

OCP FP8 E5M2 #

OCP E5M2 has a nominal format identity and a direct UInt8 runtime carrier. Its IEEE-style infinity and NaN encodings remain properties of this format package rather than assumptions of the universal executable carrier.

For addition, subtraction, multiplication, division, and square root, balanced and throughput planning select exhaustive byte tables; latency planning selects the arithmetic kernels. Fused multiply-add uses the proved generic single-rounding kernel, avoiding a 16,777,216-entry ternary table while preserving the same independent specification.

Reference #

OCP E5M2 FP8 with direct byte storage.

    Instances For
      @[instance_reducible, always_inline]

      Package the five byte tables with the proved arithmetic FMA baseline.

      @[inline]

      Construct OCP E5M2 from a natural bit pattern, reduced to eight bits.

      Instances For
        @[inline]

        Read the complete OCP E5M2 encoding byte.

        Instances For