Configured UInt8 runtime for packed posits #
These wrappers rebuild the statically selected byte carrier around unboxed raw kernels. Range
proofs are erased during compilation; semantic refinement is isolated in Byte.Proof.
The visible monomorphic boundary is intentional: it preserves the compiler's direct UInt8
calling convention. Arithmetic itself is shared in FixedWords.Kernels, so this file contains no
byte-only algorithm or fallback, only the small adapter required by the configured carrier.
Add two posits stored directly in a UInt8 carrier.
Instances For
Subtract two posits stored directly in a UInt8 carrier.
Instances For
Multiply two posits stored directly in a UInt8 carrier.
Instances For
Divide two posits stored directly in a UInt8 carrier.
Instances For
Take the square root of a posit stored directly in a UInt8 carrier.
Instances For
Fused multiply-add for posits stored directly in a UInt8 carrier.