Error of one gadget:
ProvedOPI.QG.circ_errLet be a gadget on system qubits, with data register and resource register ( qubits each, residues mod ), a -qubit control register , an accumulator and lookup scratch, and a gradient workspace of precision . Its system unitary is
where multiplies by with computed by a lookup table. is phase estimation of the shift on with control register . reads out the rounded estimate and subtracts it from . swaps and . A separate result (circ_cat) states that the gadget's circuit realizes catalytically with the phase-gradient state.
Statement. Assume and , and let be in the domain of the gadget: on its support and are , the accumulator and scratch are clean, and holds a value . With and ,
The first term comes from rounding the phase to bits, the second from phase estimation with control qubits.
This is the basic accuracy statement for the QFT over , from which the QFT over and the decoder's error budget are built.
import Definitions.Def_TwoFoldOPI_CL_Part01
import Mathlib
import Definitions.Def_QAlgorithms_Defs_Circuit
import Definitions.Def_QAlgorithms_Defs_GradientMethods
import Definitions.Def_QAlgorithms_Defs_OPIOraclePrograms
import Definitions.Def_QAlgorithms_Defs_OPIUniformity
import Definitions.Def_QAlgorithms_Defs_Fourier
import Definitions.Def_TwoFoldOPI_elemBits
import Definitions.Def_TwoFoldOPI_schurSquare
import Definitions.Def_TwoFoldOPI_schurFailureSum
import Definitions.Def_CodingTheory
import Definitions.Def_OAIThreeMachineSched01
import Definitions.Def_CookLevin_Cost
import Definitions.Def_TwoFoldOPI_codeOf
import Definitions.Def_TwoFoldOPI_solves
import Definitions.Def_TwoFoldOPI_Input
import Definitions.Def_TwoFoldOPI_CLSerializableF
open OPI OPI.QG
open QAlgorithms Matrix
variable {n b : ℕ}
variable {p w t : ℕ} (G : QG n b p w t)
theorem OPI.QG.circ_err (hp : 0 < p) (hpw : p ≤ 2 ^ w) (ht : 1 ≤ t) (g : Qubits n → ℂ) (hg : G.Dom g) :
vnorm (G.Uq.mulVec (bfac G.bE (fvec p w 0) g) - (embedOp (qftP p w) G.aE).mulVec g) ≤
(2 * Real.pi / 2 ^ b + 2 * (p / Real.sqrt (2 ^ t))) * vnorm g := by sorry