The two-stream pass qop (fine cost): with G in register rG and 2·G·S in rN (S = 2·B1·|Ms| entries per…
ProvedNLogN.KFB.qop_fspThe two-stream pass qop (fine cost): with G in register rG and 2·G·S in rN (S = 2·B1·|Ms| entries per group), the first G·S Gaussian entries of stack 0 are replaced by qopL Ms B1 G zs.
The following holds (Lean statement, in the scope of module Solutions.KFBQop):
theorem NLogN.KFB.qop_fsp (s : IState HS) (w B1 G : ℕ) (Ms : List M2G) (hM : ∀ M ∈ Ms, M.Ok)
(f : Fin HS → List ℤ) (zs : List GI) (l : List ℤ)
(hlen : G * (2 * B1 * Ms.length) ≤ zs.length)
(h0 : f 0 = gflat zs ++ l) (hS : Scr f) (hc : Cnt s f B1) (hA : f qA = []) (hG : f cG = [])
{LG LN : List Item} (hrG : (encK s f).G rG = .int (G : ℤ) :: LG)
(hrN : (encK s f).G rN = .int ((2 * (G * (2 * B1 * Ms.length)) : ℕ) : ℤ) :: LN)
(hb : NLogN.KFB.GB w zs) :
FSp (qop Ms) (encK s f)
(encK s (Function.update f 0
(gflat (qopL Ms B1 G zs ++ zs.drop (G * (2 * B1 * Ms.length))) ++ l)))
(cQop w B1 Ms.length G)
Context. The lemma belongs to the part KFB: the two-stream pass qop (fine cost): On a Gaussian array viewed as consecutive blocks [u ‖ v] of two streams of B1 entries each,
qop Ms applies the small 2 × 2 Gaussian matrix M_p = Ms[p mod L] (L = |Ms|) to block p:
(u, v) ↦ (M₀₀ u + M₀₁ v, M₁₀ u + M₁₁ v) entrywise. With a single matrix this is a
selected-bit kernel on a field [2] (the C kernel (25) scaled by 2, a phase on the selected
half, a sign); with the period [1, X] it is a CNOT whose control is the field just before.
Formalization Note. This lemma belongs to the item-language (IL) layer of the integer-multiplication-kappa campaign: tape-level programs (register programs over 64 item stacks of the NLogN multitape framework, definition bundles IntMul_HvdH_M_Defs1–IntMul_HvdH_M_Defs8) that implement the data-movement primitives of the OpenAI preprint Integer multiplication below (field moves, row split/merge, shifts, the uniform Proposition 11 interchange, Lemma 12 with fine -per-element costs, Gaussian record primitives, Proposition 14 pieces) with exact list denotations from the array-model bundles IntMul_Kappa_K_Defs* and explicit cost bounds. All names refer to the definition bundles up to IntMul_Kappa_K_Defs18.
-- OpenAI-preprint / κ-improvement development (integer-multiplication-kappa), generated by skeleton subtraction. import Definitions.Def_IntMul_Kappa_K_Defs18 set_option autoImplicit false set_option relaxedAutoImplicit false set_option maxHeartbeats 1000000 set_option linter.unusedSimpArgs false set_option linter.unusedVariables false open NLogN open NLogN.KFB open IntMul RS AModel KIL KFD
theorem NLogN.KFB.qop_fsp (s : IState HS) (w B1 G : ℕ) (Ms : List M2G) (hM : ∀ M ∈ Ms, M.Ok)
(f : Fin HS → List ℤ) (zs : List GI) (l : List ℤ)
(hlen : G * (2 * B1 * Ms.length) ≤ zs.length)
(h0 : f 0 = gflat zs ++ l) (hS : Scr f) (hc : Cnt s f B1) (hA : f qA = []) (hG : f cG = [])
{LG LN : List Item} (hrG : (encK s f).G rG = .int (G : ℤ) :: LG)
(hrN : (encK s f).G rN = .int ((2 * (G * (2 * B1 * Ms.length)) : ℕ) : ℤ) :: LN)
(hb : NLogN.KFB.GB w zs) :
FSp (qop Ms) (encK s f)
(encK s (Function.update f 0
(gflat (qopL Ms B1 G zs ++ zs.drop (G * (2 * B1 * Ms.length))) ++ l)))
(cQop w B1 Ms.length G) := by sorry