Documentation

Complexity.Algorithms.WordRAM.ThreeSum.Decision.Execute

Composing actual preprocessing with every Convolution call #

The original initialized preprocessing execution supplies its actual pointer table, count and common packed length. Its uniform polynomial word theorem bounds the pointers and signed cells. The charged native controller then runs the complete caller-safe host on every stored payload.

Main definitions #

Main results #

This native composition is original and reuses the finite Chan–He reduction correctness theorem. A fast Exact Triangle backend and the final 1.9992 endpoint remain separate obligations.

The exact common packed length already produced by native preprocessing.

Instances For

    The existing finite count bound, with the same original-input magnitude.

    Instances For

      Every preprocessing, pointer, matrix, callee and control instruction, including halt. The callee is charged at the packed side and coarse polynomial filler actually passed to it.

      Instances For
        theorem Cslib.Algorithms.WordRAM.ThreeSum.Decision.exec_large {κ w r : ℕ} (input : List ℤ) (c : Host.Callee r) (x : WordRAM.Data (registerCount r)) (hw : width κ c.widthCoefficient input.length ≤ w) (hn : 3 ≤ input.length) (hvalues : ∀ z ∈ input, |z| ≤ ↑((input.length + 2) ^ κ)) (hinit : WordRAM.Data.comap (preprocessingRegisters r) x = WordRAM.Data.init (List.map WordRAM.Signed.encode input)) (hlen : x.regs (controllerRegisters r inputLength) = input.length) (hans : x.regs (controllerRegisters r answer) = 0) :

        Complete ordinary execution after the saved-header prefix, with no solver oracle.