From dc88e4239486966a80a2204bac75a94df1a23b58 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Mon, 2 Nov 2026 12:25:45 -0300 Subject: [PATCH 01/29] Allow pulls of multiplicity up to N when the balance check bounds the interaction count. A table that answers k identical lookups with one row pulls the facts it relies on with multiplicity k. Field-sum balance alone does not match such pulls with a push: (p - 1) / 2 pulls of multiplicity 2 and one pull of 1 sum to p with no push. With fewer than p / N interactions it does: if no interaction pushes a message, its balance is -S with 1 <= S <= length * N < p. Interaction.PullUpTo / PushUpTo, MatchedUpTo and BalancedUpTo state this, and LogUpBalanced.matchedUpTo proves it. Balance gets pullBound (default 1) and matchedUpTo_of_balanced, which takes the same fits premise as balancedInteractions_of_balanced; Balance.logUpUpTo N puts length * N < p into fits. RawChannel.ConsistentUpTo N is consistency under BalancedUpTo N, and every Consistent channel has it. Sound ensembles, partial balance and the VM theorem use ConsistentUpTo balance.pullBound, so with N = 1 nothing changes. Balance.multiset's proof moves to MultisetBalanced.balancedInteractions to reuse it. --- Clean/Air/Balance.lean | 228 +++++++++++++++++++++++++++++++--- Clean/Air/OrderedChannel.lean | 107 ++++++++-------- Clean/Air/Vm.lean | 26 ++-- 3 files changed, 278 insertions(+), 83 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 7c13ed501..8124ae739 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -35,6 +35,35 @@ def Push (i : Interaction F) : Prop := i.mult ≠ 0 ∧ (i.direction = .pull → lemma not_push_iff {i : Interaction F} : ¬i.Push ↔ i.mult = 0 ∨ i.Pull := by simp only [Push, Pull, not_and, Classical.not_imp] tauto + +/-- A pull of multiplicity `k` for some `1 ≤ k ≤ N`: a table that answers `k` identical lookups +with one row pulls the facts it relies on with the same multiplicity. `Pull` is `PullUpTo 1`. -/ +def PullUpTo (N : ℕ) (i : Interaction F) : Prop := + i.direction = .pull ∧ ∃ k : ℕ, 1 ≤ k ∧ k ≤ N ∧ i.mult = k + +/-- An interaction that pushes its message when pulls of multiplicity up to `N` are allowed: an +active push, or a pull of another nonzero multiplicity. `Push` is `PushUpTo 1`. -/ +def PushUpTo (N : ℕ) (i : Interaction F) : Prop := + i.mult ≠ 0 ∧ (i.direction = .pull → ¬∃ k : ℕ, 1 ≤ k ∧ k ≤ N ∧ i.mult = k) + +lemma exists_le_one_iff {i : Interaction F} : + (∃ k : ℕ, 1 ≤ k ∧ k ≤ 1 ∧ i.mult = k) ↔ i.mult = 1 := by + constructor + · rintro ⟨k, hk1, hk2, h⟩ + obtain rfl : k = 1 := by omega + simpa using h + · intro h + exact ⟨1, le_rfl, le_rfl, by simp [h]⟩ + +lemma pullUpTo_one_iff {i : Interaction F} : i.PullUpTo 1 ↔ i.Pull := by + simp only [PullUpTo, Pull, exists_le_one_iff] + +lemma pushUpTo_one_iff {i : Interaction F} : i.PushUpTo 1 ↔ i.Push := by + simp only [PushUpTo, Push, exists_le_one_iff] + +lemma not_pushUpTo_iff {N : ℕ} {i : Interaction F} : ¬i.PushUpTo N ↔ i.mult = 0 ∨ i.PullUpTo N := by + simp only [PushUpTo, PullUpTo, not_and, Classical.not_imp, not_not] + tauto end Interaction /-- The number of active interactions in a direction that carry a message. -/ @@ -85,6 +114,54 @@ lemma BalancedInteractions.of_perm {as bs : List (Interaction F)} : lemma BalancedInteractions.nil : BalancedInteractions ([] : List (Interaction F)) := ⟨ by simp, by simp [activeCount] ⟩ +/-- +Every pull of multiplicity up to `N` is matched by an interaction that pushes its message. For +`N = 1` this is `BalancedInteractions.exists_push_of_pull`. For larger `N` it needs more than balance +of the counts: under field-sum balance, `(p - 1) / 2` pulls of multiplicity `2` and one pull of +multiplicity `1` sum to `p ≡ 0` with no push. It holds when there are fewer than `p / N` +interactions, see `LogUpBalanced.matchedUpTo`. +-/ +def MatchedUpTo (N : ℕ) (interactions : List (Interaction F)) : Prop := + ∀ a ∈ interactions, a.PullUpTo N → ∃ b ∈ interactions, b.msg = a.msg ∧ b.PushUpTo N + +lemma MatchedUpTo.of_perm {N : ℕ} {as bs : List (Interaction F)} : + MatchedUpTo N as → List.Perm as bs → MatchedUpTo N bs := by + intro matched perm a a_mem a_pull + obtain ⟨ b, b_mem, b_msg, b_push ⟩ := matched a (perm.mem_iff.mpr a_mem) a_pull + exact ⟨ b, perm.mem_iff.mp b_mem, b_msg, b_push ⟩ + +lemma MatchedUpTo.nil (N : ℕ) : MatchedUpTo N ([] : List (Interaction F)) := by + simp [MatchedUpTo] + +/-- Matching pulls of multiplicity up to `N` matches those up to any `M ≤ N`. -/ +lemma MatchedUpTo.mono {M N : ℕ} {interactions : List (Interaction F)} (h : M ≤ N) : + MatchedUpTo N interactions → MatchedUpTo M interactions := by + rintro matched a a_mem ⟨ a_direction, k, hk1, hk2, hk ⟩ + obtain ⟨ b, b_mem, b_msg, b_ne_zero, b_not_pull ⟩ := + matched a a_mem ⟨ a_direction, k, hk1, hk2.trans h, hk ⟩ + exact ⟨ b, b_mem, b_msg, b_ne_zero, + fun b_direction ⟨ j, hj1, hj2, hj ⟩ => b_not_pull b_direction ⟨ j, hj1, hj2.trans h, hj ⟩ ⟩ + +lemma BalancedInteractions.matchedUpTo_one {interactions : List (Interaction F)} + (balanced : BalancedInteractions interactions) : MatchedUpTo 1 interactions := by + intro a a_mem a_pull + obtain ⟨ b, b_mem, b_msg, b_push ⟩ := + balanced.exists_push_of_pull a a_mem (Interaction.pullUpTo_one_iff.mp a_pull) + exact ⟨ b, b_mem, b_msg, Interaction.pushUpTo_one_iff.mpr b_push ⟩ + +/-- Balance, with pulls of multiplicity up to `N` matched by a push. -/ +structure BalancedUpTo (N : ℕ) (interactions : List (Interaction F)) : Prop + extends BalancedInteractions interactions where + matchedUpTo : MatchedUpTo N interactions + +lemma BalancedUpTo.of_perm {N : ℕ} {as bs : List (Interaction F)} : + BalancedUpTo N as → List.Perm as bs → BalancedUpTo N bs := + fun balanced perm => + ⟨ balanced.toBalancedInteractions.of_perm perm, balanced.matchedUpTo.of_perm perm ⟩ + +lemma BalancedUpTo.nil (N : ℕ) : BalancedUpTo N ([] : List (Interaction F)) := + ⟨ .nil, .nil N ⟩ + /-- A balance check: the condition that a verifier checks on the interactions with each channel, the numbers of interactions with a channel that the check is sound for, and its proof that the condition @@ -95,6 +172,16 @@ structure Balance (F : Type) [FiniteField F] where fits : ℕ → Bool balancedInteractions_of_balanced : ∀ interactions, fits interactions.length → Balanced interactions → BalancedInteractions interactions + /-- The largest pull multiplicity that the check matches with a push. -/ + pullBound : ℕ := 1 + matchedUpTo_of_balanced : ∀ interactions, fits interactions.length → + Balanced interactions → MatchedUpTo pullBound interactions + +lemma Balance.balancedUpTo_of_balanced (balance : Balance F) {interactions : List (Interaction F)} + (fits : balance.fits interactions.length) (balanced : balance.Balanced interactions) : + BalancedUpTo balance.pullBound interactions := + ⟨ balance.balancedInteractions_of_balanced _ fits balanced, + balance.matchedUpTo_of_balanced _ fits balanced ⟩ /- ## LogUp: field-sum balance @@ -181,6 +268,70 @@ theorem LogUpBalanced.exists_push_of_pull {interactions : List (Interaction F)} have := balanced.activeCount_push_eq_activeCount_pull_of_isBool length_lt a.msg isBool omega +/-- +If every interaction carrying `msg` is inactive or a pull of multiplicity at most `N`, the balance +of `msg` is minus a natural number at most `length · N`, which is positive if one of them is an +active pull. +-/ +lemma balanceOf_eq_neg_of_pullsUpTo {N : ℕ} {msg : Array F} (hN : N < ringChar F) : + ∀ {interactions : List (Interaction F)}, + (∀ i ∈ interactions, i.msg = msg → i.mult = 0 ∨ i.PullUpTo N) → + ∃ S : ℕ, balanceOf interactions msg = -(S : F) ∧ S ≤ interactions.length * N ∧ + ∀ a ∈ interactions, a.msg = msg → a.PullUpTo N → 1 ≤ S + | [], _ => ⟨ 0, by simp [balanceOf], by simp, by simp ⟩ + | i :: is, h => by + obtain ⟨ S, hS, hle, hpos ⟩ := balanceOf_eq_neg_of_pullsUpTo hN (interactions := is) + fun j hj => h j (List.mem_cons_of_mem i hj) + rw [balanceOf_cons] + simp only [List.length_cons] + by_cases h_msg : i.msg = msg + case neg => + refine ⟨ S, by simp [h_msg, hS], by nlinarith, ?_ ⟩ + intro a a_mem a_msg a_pull + rcases List.mem_cons.mp a_mem with rfl | a_mem + · exact absurd a_msg h_msg + · exact hpos a a_mem a_msg a_pull + rcases h i List.mem_cons_self h_msg with h_zero | ⟨ h_direction, k, hk1, hk2, hk ⟩ + · have signed_zero : i.signedMult = 0 := by + cases h_direction : i.direction <;> simp [Interaction.signedMult, h_direction, h_zero] + refine ⟨ S, by simp [h_msg, signed_zero, hS], by nlinarith, ?_ ⟩ + intro a a_mem a_msg a_pull + rcases List.mem_cons.mp a_mem with rfl | a_mem + · -- an inactive pull of multiplicity `k` would need `k ≡ 0`, but `0 < k < ringChar F` + obtain ⟨ _, j, hj1, hj2, hj ⟩ := a_pull + rw [h_zero] at hj + have := (Lean.Grind.IsCharP.natCast_eq_iff_of_lt (α := F) _ (lt_of_le_of_lt hj2 hN) + (lt_of_le_of_lt (Nat.zero_le N) hN)).mp (by simpa using hj.symm) + omega + · exact hpos a a_mem a_msg a_pull + · refine ⟨ S + k, ?_, by nlinarith, fun _ _ _ _ => by omega ⟩ + simp only [h_msg, ↓reduceIte, Interaction.signedMult, h_direction, hk, hS] + push_cast + ring + +/-- +Every pull of multiplicity up to `N` is matched by a push, when there are fewer than +`ringChar F / N` interactions: otherwise the multiplicities of the pulls of a message could add up +to a multiple of the characteristic. +-/ +theorem LogUpBalanced.matchedUpTo {N : ℕ} {interactions : List (Interaction F)} + (balanced : LogUpBalanced interactions) (length_mul_lt : interactions.length * N < ringChar F) : + MatchedUpTo N interactions := by + intro a a_mem a_pull + -- assuming no such push exists, all interactions with the message are inactive or pulls of + -- multiplicity at most `N`, so their balance is a negative number that can't wrap around + by_contra! no_push + have hN : N < ringChar F := + lt_of_le_of_lt (Nat.le_mul_of_pos_left N (List.length_pos_of_mem a_mem)) length_mul_lt + obtain ⟨ S, hS, hle, hpos ⟩ := balanceOf_eq_neg_of_pullsUpTo hN (msg := a.msg) + (interactions := interactions) + fun i i_mem i_msg => Interaction.not_pushUpTo_iff.mp (no_push i i_mem i_msg) + have S_pos := hpos a a_mem rfl a_pull + rw [balanced a.msg, eq_comm, neg_eq_zero] at hS + have := (Lean.Grind.IsCharP.natCast_eq_iff_of_lt (α := F) _ (lt_of_le_of_lt hle length_mul_lt) + (lt_of_le_of_lt (Nat.zero_le _) length_mul_lt)).mp (by simpa using hS) + omega + /-- LogUp: field sums, sound for fewer interactions than the characteristic. -/ def Balance.logUp [ComputableRingChar F] : Balance F where Balanced := LogUpBalanced @@ -189,6 +340,27 @@ def Balance.logUp [ComputableRingChar F] : Balance F where have length_lt := ComputableRingChar.char_eq (R := F) ▸ of_decide_eq_true fits ⟨ balanced.exists_push_of_pull length_lt, balanced.activeCount_push_eq_activeCount_pull_of_isBool length_lt ⟩ + matchedUpTo_of_balanced _ fits balanced := + have length_lt := ComputableRingChar.char_eq (R := F) ▸ of_decide_eq_true fits + balanced.matchedUpTo (by simpa using length_lt) + +/-- +LogUp with counted pulls: field sums, sound for interaction counts whose product with `N` is below +the characteristic, so that pulls of multiplicity up to `N` are matched by a push. +-/ +def Balance.logUpUpTo [ComputableRingChar F] (N : ℕ) : Balance F where + Balanced := LogUpBalanced + fits count := count < ComputableRingChar.char F && count * N < ComputableRingChar.char F + balancedInteractions_of_balanced _ fits balanced := + have length_lt := ComputableRingChar.char_eq (R := F) ▸ + (of_decide_eq_true (Bool.and_eq_true_iff.mp fits).1) + ⟨ balanced.exists_push_of_pull length_lt, + balanced.activeCount_push_eq_activeCount_pull_of_isBool length_lt ⟩ + pullBound := N + matchedUpTo_of_balanced _ fits balanced := + have length_mul_lt := ComputableRingChar.char_eq (R := F) ▸ + (of_decide_eq_true (Bool.and_eq_true_iff.mp fits).2) + balanced.matchedUpTo length_mul_lt /- ## Multiset balance with `0`/`1` activations @@ -200,25 +372,32 @@ permutation of the messages of the active pulls. The permutation gives equal cou of interactions; the soundness error of a randomized check of it, like a product argument, is up to the proof system. -/ +def MultisetBalanced (interactions : List (Interaction F)) : Prop := + (∀ i ∈ interactions, IsBool i.mult) ∧ + (activeMessages interactions .push).Perm (activeMessages interactions .pull) + +theorem MultisetBalanced.balancedInteractions {interactions : List (Interaction F)} + (h_check : MultisetBalanced interactions) : BalancedInteractions interactions := by + have counts (msg : Array F) : + activeCount interactions .push msg = activeCount interactions .pull msg := by + rw [← count_activeMessages, ← count_activeMessages] + exact h_check.2.count_eq msg + refine ⟨ ?_, fun msg _ => counts msg ⟩ + intro a a_mem a_pull + have pull_pos : 0 < activeCount interactions .pull a.msg := by + apply List.countP_pos_iff.mpr + exact ⟨ a, a_mem, by simp [a_pull.1, a_pull.2] ⟩ + rw [← counts] at pull_pos + obtain ⟨ b, b_mem, h_b ⟩ := List.countP_pos_iff.mp pull_pos + simp only [decide_eq_true_eq] at h_b + exact ⟨ b, b_mem, h_b.2.1, h_b.2.2, by simp [h_b.1] ⟩ + +/-- Multiset balance with `0`/`1` activations. -/ def Balance.multiset : Balance F where - Balanced interactions := - (∀ i ∈ interactions, IsBool i.mult) ∧ - (activeMessages interactions .push).Perm (activeMessages interactions .pull) + Balanced := MultisetBalanced fits _ := true - balancedInteractions_of_balanced interactions _ h_check := by - have counts (msg : Array F) : - activeCount interactions .push msg = activeCount interactions .pull msg := by - rw [← count_activeMessages, ← count_activeMessages] - exact h_check.2.count_eq msg - refine ⟨ ?_, fun msg _ => counts msg ⟩ - intro a a_mem a_pull - have pull_pos : 0 < activeCount interactions .pull a.msg := by - apply List.countP_pos_iff.mpr - exact ⟨ a, a_mem, by simp [a_pull.1, a_pull.2] ⟩ - rw [← counts] at pull_pos - obtain ⟨ b, b_mem, h_b ⟩ := List.countP_pos_iff.mp pull_pos - simp only [decide_eq_true_eq] at h_b - exact ⟨ b, b_mem, h_b.2.1, h_b.2.2, by simp [h_b.1] ⟩ + balancedInteractions_of_balanced _ _ h_check := h_check.balancedInteractions + matchedUpTo_of_balanced _ _ h_check := h_check.balancedInteractions.matchedUpTo_one /- ## Conditions on channels that strengthen the implications of interaction balance. -/ @@ -281,6 +460,21 @@ theorem consistent_of_normal (channel : RawChannel F) [channel.Normal] : instance (channel : RawChannel F) [channel.Normal] : channel.Consistent := consistent_of_normal channel + +/-- +Consistency on balanced lists whose pulls of multiplicity up to `N` are matched by a push +(`MatchedUpTo N`). Channels with counted pulls are consistent in this sense, under a balance check +whose `pullBound` is at least `N`. +-/ +class ConsistentUpTo (N : ℕ) (channel : RawChannel F) : Prop where + consistent : ∀ (interactions : List (Interaction F)) (data : ProverData F), + BalancedUpTo N interactions → + (∀ i ∈ interactions, i.channel = channel ∧ i.Requirements data) → + (∀ i ∈ interactions, i.Guarantees data) + +instance (N : ℕ) (channel : RawChannel F) [channel.Consistent] : channel.ConsistentUpTo N where + consistent interactions data balanced := + Consistent.consistent interactions data balanced.toBalancedInteractions end RawChannel /- diff --git a/Clean/Air/OrderedChannel.lean b/Clean/Air/OrderedChannel.lean index aace47b96..789fff0c3 100644 --- a/Clean/Air/OrderedChannel.lean +++ b/Clean/Air/OrderedChannel.lean @@ -187,7 +187,7 @@ lemma orderedChannelLt_of_no_guarantees {channel : RawChannel F} {ts : List β} intro no_grts simp_all [OrderedChannelLt] -variable [FiniteField F] +variable [FiniteField F] {N : ℕ} /-- Relaxed version of BalancedChannel that works with ensembles that aren't fully specified yet, @@ -195,11 +195,11 @@ where we assume our interactions are a subset of some larger list which is balan designed to be used for proving soundness by adding one table after another. -/ -def PartialBalancedChannel (tables : Tables F) (channel : RawChannel F) : Prop := +def PartialBalancedChannel (N : ℕ) (tables : Tables F) (channel : RawChannel F) : Prop := -- `extraInteractions` represents the unknown interactions from tables added later ∃ extraInteractions : List (Interaction F), -- the total of known + unknown interactions is balanced - BalancedInteractions (tables.interactionsWith channel ++ extraInteractions) ∧ + BalancedUpTo N (tables.interactionsWith channel ++ extraInteractions) ∧ -- the extra interactions are with the same channel. (∀ i ∈ extraInteractions, i.channel = channel) ∧ -- additionally, we _assume_ that either the requirements on future interactions hold unconditionally, @@ -211,7 +211,7 @@ def PartialBalancedChannel (tables : Tables F) (channel : RawChannel F) : Prop : /-- Partial balance is trivially weaker than balance -/ lemma partialBalancedChannel_of_balancedInteractions {tables : Tables F} {channel : RawChannel F} : - BalancedInteractions (tables.interactionsWith channel) → PartialBalancedChannel tables channel := by + BalancedUpTo N (tables.interactionsWith channel) → PartialBalancedChannel N tables channel := by intro balanced use [] simp [balanced] @@ -223,16 +223,16 @@ theorem partialBalancedChannel_of_cons_of_orderedChannelLt {table : Table F} {tables : Tables F} (same_data : table.data = tables.data) {channel : RawChannel F} : table.Constraints → - PartialBalancedChannel (.cons table tables same_data) channel → + PartialBalancedChannel N (.cons table tables same_data) channel → OrderedChannelLt channel tables.components [table.component] → - PartialBalancedChannel tables channel := by + PartialBalancedChannel N tables channel := by rintro table_constraints ⟨ extraInteractions, balanced, same_channel, extra_reqs_or_no_grts ⟩ not_in_reqs_or use table.interactionsWith channel ++ extraInteractions simp only [circuit_norm] at * simp [or_imp] at ⊢ not_in_reqs_or extra_reqs_or_no_grts constructor - · apply BalancedInteractions.of_perm balanced + · apply BalancedUpTo.of_perm balanced grw [List.perm_append_comm_assoc] constructor · intro a @@ -255,9 +255,9 @@ lemma partialBalancedChannel_of_cons_of_orderedChannel {table : Table F} {tables : Tables F} (same_data : table.data = tables.data) {channel : RawChannel F} : table.Constraints → - PartialBalancedChannel (tables.cons table same_data) channel → + PartialBalancedChannel N (tables.cons table same_data) channel → OrderedChannel channel (table.component :: tables.components) → - PartialBalancedChannel tables channel := by + PartialBalancedChannel N tables channel := by intro table_constraints partial_balance ordered_channel apply partialBalancedChannel_of_cons_of_orderedChannelLt same_data table_constraints partial_balance simp_all [circuit_norm] @@ -270,10 +270,10 @@ lemma guarantees_of_requirements_cons -- given a list of tables, and one additional table {table : Table F} {tables : Tables F} (same_data : table.data = tables.data) -- and a channel that is consistent, ordered on the new table, and partially balanced on the combined tables - {channel : RawChannel F} [channel.Consistent] : + {channel : RawChannel F} [channel.ConsistentUpTo N] : table.Constraints → OrderedChannelRefl channel table.component → - PartialBalancedChannel (tables.cons table same_data) channel → + PartialBalancedChannel N (tables.cons table same_data) channel → -- the channel requirements on the old tables imply guarantees on the new table (∀ table ∈ tables.tables, table.ChannelRequirements channel) → table.ChannelGuarantees channel := by @@ -319,24 +319,24 @@ lemma guarantees_of_requirements_cons · exact ⟨ same_channel i h_mem_extra, extra_reqs i h_mem_extra ⟩ -- consistent channels goes from requirements to guarantees -- uses `consistent_channels` and `partial_balance` - apply ‹channel.Consistent›.consistent channelInteractions tables.data balanced all_reqs + apply ‹channel.ConsistentUpTo N›.consistent channelInteractions tables.data balanced all_reqs /-- Partial balance can be specialized to a sublist (= part of a permutation), as long as none of the extra tables add requirements. -/ lemma partialBalancedChannel_of_sublist {subtables tables : Tables F} {channel : RawChannel F} : - PartialBalancedChannel tables channel → + PartialBalancedChannel N tables channel → (∃ otherTables, tables.tables.Perm (subtables.tables ++ otherTables) ∧ (∀ table ∈ otherTables, table.Constraints) ∧ ∀ table ∈ otherTables, channel ∉ table.channelsWithRequirements) → - PartialBalancedChannel subtables channel := by + PartialBalancedChannel N subtables channel := by rintro ⟨ extraInteractions, balanced, same_channel, no_grts_or_extra_reqs ⟩ subset_tables obtain ⟨ otherTables, perm, otherConstraints, otherReqs ⟩ := subset_tables by_cases subtables_empty : subtables.tables = [] · simp [subtables_empty, circuit_norm, PartialBalancedChannel, Tables.interactionsWith] use [] - simp [BalancedInteractions.nil] + simp [BalancedUpTo.nil] have subtables_subset : subtables.tables ⊆ tables.tables := by have p := perm.symm.subset simp_all @@ -374,7 +374,7 @@ lemma partialBalancedChannel_of_sublist {subtables tables : Tables F} {channel : exact otherConstraints _ ht exact otherReqs _ ht -- balance - apply BalancedInteractions.of_perm balanced + apply BalancedUpTo.of_perm balanced simp only [Tables.interactionsWith] grw [← List.append_assoc, List.perm_append_right_iff, ← List.flatMap_append, perm.flatMap] exact fun _ _ => List.Perm.refl _ @@ -391,10 +391,10 @@ lemma guarantees_of_requirements_append {ts ss : Tables F} (same_data : ts.data = ss.data) -- and a channel that is consistent, _doesn't add requirements_ on the new tables, -- and is partially balanced on the combined tables - {channel : RawChannel F} [channel.Consistent] : + {channel : RawChannel F} [channel.ConsistentUpTo N] : (∀ table ∈ ts.tables, table.Constraints) → (∀ table ∈ ts.tables, channel ∉ table.component.circuit.channelsWithRequirements) → - PartialBalancedChannel (ts.append ss same_data) channel → + PartialBalancedChannel N (ts.append ss same_data) channel → -- the channel requirements on the old tables imply guarantees on the new tables (∀ table ∈ ss.tables, table.ChannelRequirements channel) → ∀ table ∈ ts.tables, table.ChannelGuarantees channel := by @@ -402,7 +402,7 @@ lemma guarantees_of_requirements_append rintro constraints reqs partial_balance ih table h_table have same_data' : table.data = ss.data := by rw [ts.same_data _ h_table, same_data] - apply guarantees_of_requirements_cons (tables := ss) same_data' (constraints _ h_table) ?_ ?_ ih + apply guarantees_of_requirements_cons (N := N) (tables := ss) same_data' (constraints _ h_table) ?_ ?_ ih · right; exact reqs _ h_table -- get partial balance by sublist/permutation argument apply partialBalancedChannel_of_sublist partial_balance @@ -460,18 +460,18 @@ Note: In the presence of "VM-like" channels, where a circuit both pushes and pul However, in that scenario, it is still useful to establish `SoundChannels` on the subset of non-VM tables. -/ @[circuit_norm] -def SoundChannels (tables : List α) (finished : List (RawChannel F)) : Prop := +def SoundChannels (N : ℕ) (tables : List α) (finished : List (RawChannel F)) : Prop := (∀ table ∈ tables, channelsWithGuarantees table ⊆ finished) ∧ (∀ channel ∈ finished, OrderedChannel channel tables) ∧ - ∀ channel ∈ finished, channel.Consistent + ∀ channel ∈ finished, channel.ConsistentUpTo N /-- `SoundChannels` lets us prove a soundness theorem. -/ theorem spec_and_guarantees_of_soundChannels {witness : Tables F} {finished : List (RawChannel F)} : - SoundChannels (witness.tables.map (·.component)) finished → + SoundChannels N (witness.tables.map (·.component)) finished → -- constraints + partial balance witness.Assumptions → witness.Constraints → - (∀ channel ∈ finished, PartialBalancedChannel witness channel) → + (∀ channel ∈ finished, PartialBalancedChannel N witness channel) → -- implies the spec, and the guarantees and requirements on all finished channels ∀ table ∈ witness.tables, table.Spec ∧ ∀ channel ∈ finished, table.ChannelGuarantees channel ∧ table.ChannelRequirements channel := by @@ -506,7 +506,7 @@ theorem spec_and_guarantees_of_soundChannels {witness : Tables F} {finished : Li rw [iff_guarantees_of_constraints assumptions.left constraints.left subset_finished.left] -- the rest is just applying `guarantees_of_requirements_cons` intro channel h_channel - have : channel.Consistent := consistent_channels _ h_channel + have : channel.ConsistentUpTo N := consistent_channels _ h_channel have orderedChannelRefl : OrderedChannelRefl channel table.component := by simp only [circuit_norm, ordered_channels channel h_channel] apply guarantees_of_requirements_cons same_data @@ -517,10 +517,10 @@ theorem spec_and_guarantees_of_soundChannels {witness : Tables F} {finished : Li /-- `SoundChannels` is strictly increasing: you can make the finished list bigger by any consistent channels -/ lemma soundChannels_of_soundChannels_subset {tables : List α} {finished finished' : List (RawChannel F)} : - SoundChannels tables finished → + SoundChannels N tables finished → finished ⊆ finished' → - (∀ channel ∈ finished', channel.Consistent) → - SoundChannels tables finished' := by + (∀ channel ∈ finished', channel.ConsistentUpTo N) → + SoundChannels N tables finished' := by rintro ⟨ subset_finished, ordered_channels, consistent_channels ⟩ finished'_subset finished'_consistent constructor · intro table h_table @@ -538,9 +538,9 @@ lemma soundChannels_of_soundChannels_subset {tables : List α} /-- You can add one channel to the finished list and preserve `SoundChannels` -/ lemma soundChannels_cons_of_soundChannels {tables : List α} - {finished : List (RawChannel F)} {channel : RawChannel F} [channel.Consistent] : - SoundChannels tables finished → - SoundChannels tables (channel :: finished) := by + {finished : List (RawChannel F)} {channel : RawChannel F} [channel.ConsistentUpTo N] : + SoundChannels N tables finished → + SoundChannels N tables (channel :: finished) := by intro sound_channels apply soundChannels_of_soundChannels_subset sound_channels · simp @@ -561,20 +561,20 @@ abbrev OrderedChannels (ens : Ensemble F PublicIO) (finished : List (RawChannel ∀ channel ∈ finished, ens.OrderedChannel channel @[circuit_norm] -def SoundChannels (ens : Ensemble F PublicIO) (finished : List (RawChannel F)) : Prop := +def SoundChannels (ens : Ensemble F PublicIO) (N : ℕ) (finished : List (RawChannel F)) : Prop := ens.channelsWithGuarantees ⊆ finished ∧ ens.OrderedChannels finished ∧ - ∀ channel ∈ finished, channel.Consistent + ∀ channel ∈ finished, channel.ConsistentUpTo N lemma partialBalancedChannel_of_balancedChannel {ens : Ensemble F PublicIO} {witness : EnsembleWitness ens} {channel : RawChannel F} : ens.OrderedChannel channel → - witness.BalancedChannel channel → - PartialBalancedChannel witness.toTables channel := by + BalancedUpTo N (witness.interactionsWith channel) → + PartialBalancedChannel N witness.toTables channel := by intro ordered balanced use witness.verifierInteractionsWith channel constructor - · apply BalancedInteractions.of_perm balanced + · apply BalancedUpTo.of_perm balanced apply List.perm_append_comm constructor · intro interaction h_interaction @@ -591,9 +591,9 @@ lemma partialBalancedChannel_of_balancedChannel {ens : Ensemble F PublicIO} simp_all [circuit_norm] lemma verifierChannelGuarantees_of_tableRequirements {ens : Ensemble F PublicIO} - {witness : EnsembleWitness ens} {channel : RawChannel F} [channel.Consistent] : + {witness : EnsembleWitness ens} {channel : RawChannel F} [channel.ConsistentUpTo N] : OrderedChannelRefl channel ens.verifier → - witness.BalancedChannel channel → + BalancedUpTo N (witness.interactionsWith channel) → (∀ table ∈ witness.tables, table.ChannelRequirements channel) → ens.VerifierChannelGuarantees witness.data channel := by intro ordered balanced table_requirements @@ -605,7 +605,7 @@ lemma verifierChannelGuarantees_of_tableRequirements {ens : Ensemble F PublicIO} rw [witness.verifierChannelRequirements_iff_forall] at verifier_requirements -- all interactions satisfy their requirements, so by consistency they satisfy their guarantees have all_guarantees := - ‹channel.Consistent›.consistent (witness.interactionsWith channel) witness.data balanced (by + ‹channel.ConsistentUpTo N›.consistent (witness.interactionsWith channel) witness.data balanced (by intro interaction h_interaction refine ⟨witness.channel_eq_of_mem_interactionsWith h_interaction, ?_⟩ rw [witness.mem_interactionsWith] at h_interaction @@ -624,11 +624,12 @@ Main result of this section: `TableSoundness`, a complex ensemble-level soundness statement. -/ theorem tableSoundness_of_soundChannels {ens : Ensemble F PublicIO} {balance : Balance F} : - (∃ finished, finished ⊆ ens.channels ∧ ens.SoundChannels finished) → + (∃ finished, finished ⊆ ens.channels ∧ ens.SoundChannels balance.pullBound finished) → ens.TableSoundness balance := by intro ⟨ finished, finished_subset, subset_finished, ordered, consistent ⟩ witness assumptions constraints fit balanced - have balance := witness.balancedChannels_of_balanced fit balanced + have balance := fun channel h_channel => + balance.balancedUpTo_of_balanced (fit channel h_channel) (balanced channel h_channel) rw [channelsWithGuarantees_subset_iff] at subset_finished -- the tables are sound by themselves, given partial balance have table_results := spec_and_guarantees_of_soundChannels (witness := witness.toTables) @@ -649,7 +650,7 @@ theorem tableSoundness_of_soundChannels {ens : Ensemble F PublicIO} {balance : B exact fun table h_table => ((table_results table h_table).right channel h_finished).right /-- Empty ensemble satisfies SoundChannels -/ -theorem empty_soundChannels : (empty F PublicIO).SoundChannels [] := by +theorem empty_soundChannels : (empty F PublicIO).SoundChannels N [] := by simp only [circuit_norm] /-- Empty ensemble satisfies TableSoundness -/ @@ -662,7 +663,7 @@ theorem orderedChannels_of_soundChannels_addTable (ens : Ensemble F PublicIO) (table : Component F) (fresh : table.circuit.name ∉ ens.tables.map (·.circuit.name)) (table_channels : table.circuit.channels ⊆ ens.channels) {finished : List (RawChannel F)} : -- given a sound channels ensemble with empty verifier, - ens.SoundChannels finished → + ens.SoundChannels N finished → ens.verifier = .empty F PublicIO → -- assuming that the new table's channelsWithGuarantees are all finished table.circuit.channelsWithGuarantees ⊆ finished → @@ -682,7 +683,7 @@ theorem orderedChannels_of_soundChannels_merge (ens1 ens2 : Ensemble F PublicIO) (unique_channel_names : ((ens2.channels ++ ens1.channels).map (·.name)).Nodup) {finished : List (RawChannel F)} : -- given a sound channels ensemble with empty verifier, - ens1.SoundChannels finished → + ens1.SoundChannels N finished → ens1.verifier = .empty F PublicIO → -- assuming that the new tables' channelsWithRequirements contain none of the finished channels (∀ channel ∈ finished, channel ∉ ens2.channelsWithRequirements) → @@ -700,11 +701,11 @@ theorem orderedChannels_of_soundChannels_merge (ens1 ens2 : Ensemble F PublicIO) theorem soundChannels_markFinished (ens : Ensemble F PublicIO) -- given a sound channels ensemble with a list of finished channels - {finished : List (RawChannel F)} (h_sound : ens.SoundChannels finished) + {finished : List (RawChannel F)} (h_sound : ens.SoundChannels N finished) -- and given a new consistent channel to mark as finished - (channel : RawChannel F) [channel.Consistent] : + (channel : RawChannel F) [channel.ConsistentUpTo N] : -- the ensemble also satisfies SoundChannels including the new channel in the finished list - ens.SoundChannels (channel :: finished) := by + ens.SoundChannels N (channel :: finished) := by rcases h_sound with ⟨subset_finished, ordered_channels, consistent_channels⟩ refine ⟨List.subset_cons_of_subset _ subset_finished, ?_, by simp_all⟩ by_cases h_finished : channel ∈ finished @@ -720,7 +721,7 @@ end Ensemble structure SoundEnsemble (F : Type) [FiniteField F] (PublicIO : TypeMap) [ProvableType PublicIO] (balance : Balance F) extends ensemble : Ensemble F PublicIO where finished : List (RawChannel F) - finished_consistent : ∀ channel ∈ finished, channel.Consistent + finished_consistent : ∀ channel ∈ finished, channel.ConsistentUpTo balance.pullBound finished_subset : finished ⊆ channels subset_finished : ensemble.channelsWithGuarantees ⊆ finished ordered_channels : ensemble.OrderedChannels finished @@ -734,7 +735,7 @@ attribute [circuit_norm] SoundEnsemble.finished_consistent SoundEnsemble.finishe namespace SoundEnsemble variable {balance : Balance F} -lemma soundChannels (ens : SoundEnsemble F PublicIO balance) : ens.SoundChannels ens.finished := by +lemma soundChannels (ens : SoundEnsemble F PublicIO balance) : ens.SoundChannels balance.pullBound ens.finished := by exact ⟨ens.subset_finished, ens.ordered_channels, ens.finished_consistent⟩ def empty (F : Type) [FiniteField F] (PublicIO : TypeMap) [ProvableType PublicIO] @@ -834,7 +835,7 @@ variable {channel : RawChannel F} {fresh : channel.name ∉ soundEns.channels.ma (soundEns.addChannel channel fresh).verifier = soundEns.verifier := rfl end -def markFinished (soundEns : SoundEnsemble F PublicIO balance) (channel : RawChannel F) [channel.Consistent] +def markFinished (soundEns : SoundEnsemble F PublicIO balance) (channel : RawChannel F) [channel.ConsistentUpTo balance.pullBound] (h_mem : channel ∈ soundEns.channels := by simp [circuit_norm]) : SoundEnsemble F PublicIO balance where ensemble := soundEns.ensemble @@ -849,7 +850,7 @@ def markFinished (soundEns : SoundEnsemble F PublicIO balance) (channel : RawCha subset_finished := by simp [soundEns.subset_finished] ordered_channels := by intro channel' hc - have : channel'.Consistent := by + have : channel'.ConsistentUpTo balance.pullBound := by simp at hc rcases hc with rfl | hc_tail · assumption @@ -858,7 +859,7 @@ def markFinished (soundEns : SoundEnsemble F PublicIO balance) (channel : RawCha exact this.right.left channel' (by simp) verifier_empty := soundEns.verifier_empty -variable {channel : RawChannel F} [channel.Consistent] {h_mem : channel ∈ soundEns.channels} +variable {channel : RawChannel F} [channel.ConsistentUpTo balance.pullBound] {h_mem : channel ∈ soundEns.channels} @[circuit_norm] lemma markFinished_channels : (soundEns.markFinished channel h_mem).channels = soundEns.channels := rfl @@ -869,7 +870,7 @@ variable {channel : RawChannel F} [channel.Consistent] {h_mem : channel ∈ soun /-- Add a channel that no table pushes to, as finished. Its name must be new. -/ def addFinishedChannel (soundEns : SoundEnsemble F PublicIO balance) (channel : RawChannel F) - [channel.Consistent] + [channel.ConsistentUpTo balance.pullBound] (fresh : channel.name ∉ soundEns.channels.map (·.name) := by simp [circuit_norm]) : SoundEnsemble F PublicIO balance := soundEns @@ -877,7 +878,7 @@ def addFinishedChannel (soundEns : SoundEnsemble F PublicIO balance) (channel : |>.markFinished channel (List.mem_cons_self ..) section -variable {channel : RawChannel F} [channel.Consistent] +variable {channel : RawChannel F} [channel.ConsistentUpTo balance.pullBound] {fresh : channel.name ∉ soundEns.channels.map (·.name)} @[circuit_norm] lemma addFinishedChannel_channels : diff --git a/Clean/Air/Vm.lean b/Clean/Air/Vm.lean index ad2456177..c6a9b99db 100644 --- a/Clean/Air/Vm.lean +++ b/Clean/Air/Vm.lean @@ -796,8 +796,8 @@ namespace Ensemble theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balance : Balance F} -- given a sound channels ensemble with a list of finished, consistent channels - {finished : List (RawChannel F)} (soundChannels : ens.SoundChannels finished) - (consistent : ∀ channel ∈ finished, channel.Consistent) + {finished : List (RawChannel F)} (soundChannels : ens.SoundChannels balance.pullBound finished) + (consistent : ∀ channel ∈ finished, channel.ConsistentUpTo balance.pullBound) (finished_subset : finished ⊆ ens.channels) -- and given a VM channel + tables + verifier (vm : VmTables F PublicIO) @@ -814,7 +814,8 @@ theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balan (ens.addVm vm fits).SoundVmChannel balance := by intro not_mem_vm_channel grts_subset reqs_disjoint (witness : VmWitness ens vm fits) assumptions constraints fit balanced - have balance := witness.balancedChannels_of_balanced fit balanced + have balance_upTo := fun channel h_channel => + balance.balancedUpTo_of_balanced (fit channel h_channel) (balanced channel h_channel) /- the high level plan is to apply `verifier_guarantees_of_requirements_of_requirements_of_guarantees`. @@ -830,8 +831,7 @@ theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balan -/ -- the vm channel interactions are constrained to vm tables, which lets us supply the balance -- condition - have vm_balance := balance vm.channel.toRaw (by simp [Ensemble.addVm]) - simp only [circuit_norm] at vm_balance + have vm_balance := (balance_upTo vm.channel.toRaw (by simp [Ensemble.addVm])).toBalancedInteractions rw [witness.interactionsWith_eq_vmInteractionsWith not_mem_vm_channel] at vm_balance set vmChannel := vm.channel.toRaw -- next, we work on instantiating `requirements_of_partial_guarantees_of_constraints` @@ -856,13 +856,13 @@ theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balan clear * - ordered no_requirements simp_all [circuit_norm, orderedChannel_of_no_requirements, orderedChannelLt_of_no_requirements] have partial_balance : ∀ channel ∈ finished, - PartialBalancedChannel witness.toTables channel := by + PartialBalancedChannel balance.pullBound witness.toTables channel := by intro channel channel_mem apply Ensemble.partialBalancedChannel_of_balancedChannel · exact ordered_finished channel channel_mem - · exact balance channel (by simp [Ensemble.addVm, finished_subset channel_mem]) + · exact balance_upTo channel (by simp [Ensemble.addVm, finished_subset channel_mem]) have old_partial_balance : ∀ channel ∈ finished, - PartialBalancedChannel witness.oldContext channel := by + PartialBalancedChannel balance.pullBound witness.oldContext channel := by intro channel channel_mem apply partialBalancedChannel_of_sublist (partial_balance _ channel_mem) exact ⟨witness.vmContext.tables, witness.toTables_perm, vm_constraints, @@ -876,7 +876,7 @@ theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balan (witness.oldContext_constraints_of_constraints constraints) old_partial_balance table table_mem |>.right channel channel_mem |>.right - have old_sound_channels : _root_.SoundChannels ens.tables finished := ⟨ + have old_sound_channels : _root_.SoundChannels balance.pullBound ens.tables finished := ⟨ ens.channelsWithGuarantees_subset_iff.mp soundChannels.left |>.right, fun channel h_channel => (soundChannels.right.left channel h_channel).right.left, soundChannels.right.right⟩ @@ -885,7 +885,7 @@ theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balan have finished_grts : ∀ table ∈ witness.vmTables, ∀ channel ∈ finished, table.ChannelGuarantees channel := by intro table table_mem channel channel_mem - have : channel.Consistent := consistent channel channel_mem + have : channel.ConsistentUpTo balance.pullBound := consistent channel channel_mem apply guarantees_of_requirements_append (ts := witness.vmContext) (ss := witness.oldContext) rfl vm_constraints (vm_reqs_disjoint _ channel_mem) (witness.vmContext_append_oldContext ▸ partial_balance _ channel_mem) @@ -905,10 +905,10 @@ theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balan have finished_verifier_guarantees : ∀ channel ∈ finished, (ens.addVm vm fits).VerifierChannelGuarantees witness.data channel := by intro channel channel_mem - let : channel.Consistent := consistent channel channel_mem - apply Ensemble.verifierChannelGuarantees_of_tableRequirements + let : channel.ConsistentUpTo balance.pullBound := consistent channel channel_mem + apply Ensemble.verifierChannelGuarantees_of_tableRequirements (N := balance.pullBound) · exact .inr (reqs_disjoint channel channel_mem).1 - · exact balance channel (by simp [Ensemble.addVm, finished_subset channel_mem]) + · exact balance_upTo channel (by simp [Ensemble.addVm, finished_subset channel_mem]) intro table h_table by_cases h_vm : table ∈ witness.vmTables · exact table.requirements_of_not_mem_of_constraints (vm_constraints table h_vm) From edd26ee39c6f33748d177681c9295150e6d086f7 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Mon, 2 Nov 2026 12:26:01 -0300 Subject: [PATCH 02/29] Add Channel.counted N, a typed channel whose pulls may count up to N. A pull of multiplicity 1..N gets the channel's guarantee, an active push owes it, and a pull requires its multiplicity to be 0 or 1..N. The channel is ConsistentUpTo N, so a sound ensemble can finish it under a balance check with pullBound at least N, such as Balance.logUpUpTo N. pullUpTo and emitUpTo are the circuit operations. The interactions they emit are opaque definitions with circuit_norm lemmas, so that simp does not unfold the message first, and RawChannel.eq_false_of_name_ne tells channels apart by name without unfolding them. --- Clean.lean | 1 + Clean/Air/CountedChannel.lean | 161 ++++++++++++++++++++++++++++++++++ 2 files changed, 162 insertions(+) create mode 100644 Clean/Air/CountedChannel.lean diff --git a/Clean.lean b/Clean.lean index eb431921a..48c5334da 100644 --- a/Clean.lean +++ b/Clean.lean @@ -27,6 +27,7 @@ public import Clean.Examples.FemtoCairo.FemtoCairo public import Clean.Examples.FemtoCairo.FlatAir public import Clean.Air.WitnessGeneration public import Clean.Air.Lookup +public import Clean.Air.CountedChannel public import Clean.Backends.Plonky3.Print public import Clean.Tables.Fibonacci8 public import Clean.Tables.Fibonacci32 diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean new file mode 100644 index 000000000..7224a50d6 --- /dev/null +++ b/Clean/Air/CountedChannel.lean @@ -0,0 +1,161 @@ +module + +public import Clean.Air.Balance +public import Clean.Circuit.Explicit + +@[expose] public section + +/-! +# Channels with counted pulls + +A table that answers `k` identical lookups with one row pulls the facts it relies on with the same +multiplicity `k`. `Channel.counted N channel` is `channel` with pulls of multiplicity up to `N`: such +a pull gets the guarantee, and an active push owes it. It is consistent in the sense of +`RawChannel.ConsistentUpTo N`, so it can be finished in a `SoundEnsemble` whose balance check has a +`pullBound` of at least `N`, e.g. `Balance.logUpUpTo N`. +-/ + +variable {F : Type} [FiniteField F] +variable {Message : TypeMap} [ProvableType Message] + +namespace Channel + +/-- `k` for some `1 ≤ k ≤ N`. -/ +def CountUpTo (N : ℕ) (mult : F) : Prop := ∃ k : ℕ, 1 ≤ k ∧ k ≤ N ∧ mult = k + +/-- `channel`, with pulls of multiplicity up to `N`. -/ +def counted (N : ℕ) (channel : Channel F Message) : RawChannel F where + name := channel.name + arity := size Message + Guarantees mult message data := + CountUpTo N mult → channel.Guarantees (fromElements message) data + Requirements direction mult message data := + match direction with + | .push => mult ≠ 0 → channel.Guarantees (fromElements message) data + | .pull => mult = 0 ∨ CountUpTo N mult + +omit [FiniteField F] in +/-- Channels with different names are different. Lets `circuit_norm` tell a counted channel apart +from the other channels of a circuit without unfolding them. -/ +@[circuit_norm] lemma _root_.RawChannel.eq_false_of_name_ne {a b : RawChannel F} + (h : a.name ≠ b.name) : (a = b) = False := + eq_false fun h_eq => h (congrArg RawChannel.name h_eq) + +@[circuit_norm] lemma counted_name (N : ℕ) (channel : Channel F Message) : + (channel.counted N).name = channel.name := rfl +@[circuit_norm] lemma counted_arity (N : ℕ) (channel : Channel F Message) : + (channel.counted N).arity = size Message := rfl + +/-- The interaction of pulling `msg` `count` times from `channel.counted N`. -/ +def pulledUpTo (N : ℕ) (channel : Channel F Message) (count : Expression F) + (msg : Message (Expression F)) : AbstractInteraction F := + ⟨ channel.counted N, .pull, count, toElements msg ⟩ + +/-- The interaction of pushing `msg` `mult` times to `channel.counted N`. -/ +def emittedUpTo (N : ℕ) (channel : Channel F Message) (mult : Expression F) + (msg : Message (Expression F)) : AbstractInteraction F := + ⟨ channel.counted N, .push, mult, toElements msg ⟩ + +/-- Pull `msg` `count` times, for `count` up to `N`. -/ +@[circuit_norm] +def pullUpTo (N : ℕ) (channel : Channel F Message) (count : Expression F) + (msg : Message (Expression F)) : Circuit F Unit := fun _ => + ((), [.interact (pulledUpTo N channel count msg)]) + +/-- Push `msg` `mult` times to `channel.counted N`. -/ +@[circuit_norm] +def emitUpTo (N : ℕ) (channel : Channel F Message) (mult : Expression F) + (msg : Message (Expression F)) : Circuit F Unit := fun _ => + ((), [.interact (emittedUpTo N channel mult msg)]) + +instance {N : ℕ} {channel : Channel F Message} {count : Expression F} : + ExplicitCircuits (F := F) (channel.pullUpTo N count) where + output _ _ := () + localLength _ _ := 0 + operations msg _ := [.interact (pulledUpTo N channel count msg)] + channelsWithGuarantees _ _ := [channel.counted N] + +instance {N : ℕ} {channel : Channel F Message} {mult : Expression F} : + ExplicitCircuits (F := F) (channel.emitUpTo N mult) where + output _ _ := () + localLength _ _ := 0 + operations msg _ := [.interact (emittedUpTo N channel mult msg)] + channelsWithGuarantees _ _ := [] + +attribute [explicit_circuit_no_unfold] pullUpTo emitUpTo + +section +variable {N : ℕ} {channel : Channel F Message} {mult : Expression F} + {msg : Message (Expression F)} {env : Environment F} + +@[circuit_norm] lemma pulledUpTo_channel : + (pulledUpTo N channel mult msg).channel = channel.counted N := rfl +@[circuit_norm] lemma pulledUpTo_direction : + (pulledUpTo N channel mult msg).direction = .pull := rfl +@[circuit_norm] lemma emittedUpTo_channel : + (emittedUpTo N channel mult msg).channel = channel.counted N := rfl +@[circuit_norm] lemma emittedUpTo_direction : + (emittedUpTo N channel mult msg).direction = .push := rfl + +@[circuit_norm] lemma pulledUpTo_guarantees : + (pulledUpTo N channel mult msg).Guarantees env ↔ + (CountUpTo N (Expression.eval env mult) → channel.Guarantees (eval env msg) env.data) := by + simp [pulledUpTo, AbstractInteraction.Guarantees, counted, + ProvableType.fromElements_eval_toElements] + +@[circuit_norm] lemma pulledUpTo_requirements : + (pulledUpTo N channel mult msg).Requirements env ↔ + (Expression.eval env mult = 0 ∨ CountUpTo N (Expression.eval env mult)) := by + simp [pulledUpTo, AbstractInteraction.Requirements, counted] + +@[circuit_norm] lemma emittedUpTo_guarantees : + (emittedUpTo N channel mult msg).Guarantees env := by + simp [emittedUpTo, AbstractInteraction.Guarantees] + +@[circuit_norm] lemma emittedUpTo_requirements : + (emittedUpTo N channel mult msg).Requirements env ↔ + (Expression.eval env mult ≠ 0 → channel.Guarantees (eval env msg) env.data) := by + simp [emittedUpTo, AbstractInteraction.Requirements, counted, + ProvableType.fromElements_eval_toElements] +end + +omit [FiniteField F] in +/-- A nonzero multiplicity whose value is at most `N`. -/ +lemma countUpTo_of_val {p : ℕ} [Fact p.Prime] {N : ℕ} {mult : ZMod p} (h0 : mult ≠ 0) + (hN : mult.val ≤ N) : CountUpTo N mult := + ⟨ mult.val, Nat.one_le_iff_ne_zero.mpr ((ZMod.val_ne_zero mult).mpr h0), hN, + (ZMod.natCast_zmod_val mult).symm ⟩ + +/-- A pull of multiplicity up to `N` is matched by a push, whose requirement gives the +guarantee. Every other interaction gets no guarantee. -/ +instance (N : ℕ) (channel : Channel F Message) : (channel.counted N).ConsistentUpTo N where + consistent interactions data balanced reqs a a_mem := by + obtain ⟨ a_channel, - ⟩ := reqs a a_mem + by_cases a_pull : a.PullUpTo N + case neg => + rcases a with ⟨ _, a_direction, a_mult, a_msg, a_size ⟩ + subst a_channel + simp only [Interaction.Guarantees, counted] + exact fun h_direction h_count => absurd ⟨ h_direction, h_count ⟩ a_pull + obtain ⟨ b, b_mem, b_msg, b_ne_zero, b_not_pull ⟩ := balanced.matchedUpTo a a_mem a_pull + obtain ⟨ b_channel, b_reqs ⟩ := reqs b b_mem + rcases a with ⟨ _, a_direction, a_mult, a_msg, a_size ⟩ + rcases b with ⟨ _, b_direction, b_mult, b_msg', b_size ⟩ + subst a_channel b_channel + simp only at b_msg + subst b_msg + simp only [Interaction.Guarantees, Interaction.Requirements, Interaction.msgVector, + counted] at b_reqs ⊢ + intro _ _ + cases b_direction with + | push => exact b_reqs b_ne_zero + | pull => exact absurd (b_reqs.resolve_left b_ne_zero) (b_not_pull rfl) + +/-- The counted channel is also consistent for balance checks with a larger bound. -/ +theorem counted_consistentUpTo_of_le {M N : ℕ} (h : N ≤ M) (channel : Channel F Message) : + (channel.counted N).ConsistentUpTo M where + consistent interactions data balanced := + RawChannel.ConsistentUpTo.consistent (N := N) interactions data + ⟨ balanced.toBalancedInteractions, balanced.matchedUpTo.mono h ⟩ + +end Channel From b71d1832b8b5f3f7eb1e81adc1f5f8844ea1e30e Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 5 Nov 2026 18:29:22 -0300 Subject: [PATCH 03/29] Show that a counted channel with 2 <= N is not Consistent under plain balance. Channel.counted_not_consistent gives the counterexample that justifies ConsistentUpTo: two pulls of a message the channel does not guarantee, one of multiplicity N and one of multiplicity 1, and nothing pushed. BalancedInteractions holds on that list, because exists_push_of_pull only has to match pulls of multiplicity 1 and reads the pull of N as an Interaction.Push, while the count condition is vacuous since N is neither 0 nor 1. The counted channel reads both as pulls and hands the guarantee to the pull of 1, which is false. Without this the library cannot answer why RawChannel.Consistent is not enough and why MatchedUpTo, BalancedUpTo and Balance.pullBound exist. The statement is over a general field; all the counterexample needs is 2 <= N < ringChar F, which makes N distinct from both 0 and 1. --- Clean/Air/CountedChannel.lean | 65 +++++++++++++++++++++++++++++++++++ 1 file changed, 65 insertions(+) diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 7224a50d6..eb471eb7b 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -158,4 +158,69 @@ theorem counted_consistentUpTo_of_le {M N : ℕ} (h : N ≤ M) (channel : Channe RawChannel.ConsistentUpTo.consistent (N := N) interactions data ⟨ balanced.toBalancedInteractions, balanced.matchedUpTo.mono h ⟩ +/-- +Why `ConsistentUpTo` is needed: for `2 ≤ N` the counted channel is **not** +`RawChannel.Consistent`, so it could not be finished in a `SoundEnsemble` under Clean's unmodified +balance premise `BalancedInteractions`. + +The counterexample is two pulls of a message the channel does not guarantee, one of multiplicity +`N` and one of multiplicity `1`, with nothing pushed. `BalancedInteractions` is satisfied: +`exists_push_of_pull` only has to match pulls of multiplicity `1`, and it counts the pull of `N` as +an `Interaction.Push` of the message, so the pull of `1` is matched; the count condition is vacuous +because `N` is neither `0` nor `1`. The counted channel, in contrast, reads both as pulls and hands +the guarantee to the pull of `1`, which is false. + +This is exactly the gap that `MatchedUpTo N` / `BalancedUpTo N` and `Balance.pullBound` close, and +the reason `Channel.counted` carries a `RawChannel.ConsistentUpTo N` instance instead of a +`RawChannel.Consistent` one. + +The hypothesis `N < ringChar F` is what makes `N` distinct from both `0` and `1` in `F`; together +with `2 ≤ N` it is all the counterexample needs. +-/ +theorem counted_not_consistent {N : ℕ} (hN : 2 ≤ N) (h_char : N < ringChar F) + (channel : Channel F Message) (m : Message F) (data : ProverData F) + (h_m : ¬ channel.Guarantees m data) : ¬ (channel.counted N).Consistent := by + rintro ⟨ consistent ⟩ + have cast_inj {a b : ℕ} (ha : a < ringChar F) (hb : b < ringChar F) : + (a : F) = (b : F) → a = b := + fun h => (Lean.Grind.IsCharP.natCast_eq_iff_of_lt (α := F) _ ha hb).mp h + have hN0 : (N : F) ≠ 0 := fun h => by + have := cast_inj h_char (b := 0) (by omega) (by simpa using h); omega + have hN1 : (N : F) ≠ 1 := fun h => by + have := cast_inj h_char (b := 1) (by omega) (by simpa using h); omega + let mk : F → Interaction F := + fun mult => ⟨ channel.counted N, .pull, mult, (toElements m).toArray, by simp [counted] ⟩ + let l := [mk (N : F), mk 1] + have h_bal : BalancedInteractions l := by + constructor + · -- the pull of `1` is matched by the pull of `N`, which `Interaction.Push` counts as a push + intro a a_mem _ + refine ⟨ mk (N : F), by simp [l], ?_, hN0, fun _ => hN1 ⟩ + simp only [l, List.mem_cons, List.not_mem_nil, or_false] at a_mem + rcases a_mem with rfl | rfl <;> rfl + · -- the multiplicity `N` is neither `0` nor `1`, so only untouched messages qualify + intro msg is_bool + by_cases h_msg : (toElements m).toArray = msg + · exfalso + rcases is_bool (mk (N : F)) (by simp [l]) h_msg with h | h + · exact hN0 h + · exact hN1 h + · simp [activeCount, l, mk, h_msg] + have h_reqs : ∀ i ∈ l, i.channel = channel.counted N ∧ i.Requirements data := by + intro i h_i + simp only [l, List.mem_cons, List.not_mem_nil, or_false] at h_i + rcases h_i with rfl | rfl + · refine ⟨ rfl, ?_ ⟩ + simp only [Interaction.Requirements, mk, counted, Interaction.msgVector] + exact .inr ⟨ N, by omega, le_rfl, rfl ⟩ + · refine ⟨ rfl, ?_ ⟩ + simp only [Interaction.Requirements, mk, counted, Interaction.msgVector] + exact .inr ⟨ 1, le_rfl, by omega, by simp ⟩ + have h_g := consistent l data h_bal h_reqs (mk 1) (by simp [l]) + simp only [Interaction.Guarantees, mk, counted, Interaction.msgVector] at h_g + apply h_m + have h := h_g trivial ⟨ 1, le_rfl, by omega, by simp ⟩ + convert h + simp [ProvableType.fromElements_toElements] + end Channel From a58b59472cd1529f31fada088796685fef9fbe70 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 5 Nov 2026 18:32:48 -0300 Subject: [PATCH 04/29] Expose the pullBound of each balance check to simp. Balance.logUp, Balance.logUpUpTo and Balance.multiset are plain defs, so instance search, which runs at reducible transparency, cannot project pullBound out of them. A user of a SoundEnsemble over Balance.logUpUpTo N therefore had to write show channel.ConsistentUpTo N by hand before the ConsistentUpTo instance of a counted channel would be found. Balance.logUp_pullBound, Balance.logUpUpTo_pullBound and Balance.multiset_pullBound are rfl lemmas tagged simp and circuit_norm, so the goal normalises to the concrete bound first. The example in CountedChannel.lean guards this: without the lemma, infer_instance on ConsistentUpTo (Balance.logUpUpTo N).pullBound fails. --- Clean/Air/Balance.lean | 17 +++++++++++++++++ Clean/Air/CountedChannel.lean | 9 +++++++++ 2 files changed, 26 insertions(+) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 8124ae739..efcb03e97 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -344,6 +344,12 @@ def Balance.logUp [ComputableRingChar F] : Balance F where have length_lt := ComputableRingChar.char_eq (R := F) ▸ of_decide_eq_true fits balanced.matchedUpTo (by simpa using length_lt) +/-- `Balance.logUp` matches pulls of multiplicity `1` only. Stated so that `simp`/`circuit_norm` +can reduce a `ConsistentUpTo (Balance.logUp).pullBound` goal, which instance search cannot do on +its own: `Balance.logUp` is a plain `def`, so the projection is not reducible. -/ +@[simp, circuit_norm] lemma Balance.logUp_pullBound [ComputableRingChar F] : + (Balance.logUp : Balance F).pullBound = 1 := rfl + /-- LogUp with counted pulls: field sums, sound for interaction counts whose product with `N` is below the characteristic, so that pulls of multiplicity up to `N` are matched by a push. @@ -362,6 +368,13 @@ def Balance.logUpUpTo [ComputableRingChar F] (N : ℕ) : Balance F where (of_decide_eq_true (Bool.and_eq_true_iff.mp fits).2) balanced.matchedUpTo length_mul_lt +/-- `Balance.logUpUpTo N` matches pulls of multiplicity up to `N`. Without this lemma a user of a +`SoundEnsemble` over `Balance.logUpUpTo N` has to write `show channel.ConsistentUpTo N` by hand, +because instance search works at reducible transparency and cannot project `pullBound` out of the +plain `def`. -/ +@[simp, circuit_norm] lemma Balance.logUpUpTo_pullBound [ComputableRingChar F] (N : ℕ) : + (Balance.logUpUpTo N : Balance F).pullBound = N := rfl + /- ## Multiset balance with `0`/`1` activations -/ @@ -399,6 +412,10 @@ def Balance.multiset : Balance F where balancedInteractions_of_balanced _ _ h_check := h_check.balancedInteractions matchedUpTo_of_balanced _ _ h_check := h_check.balancedInteractions.matchedUpTo_one +/-- `Balance.multiset` matches pulls of multiplicity `1` only. -/ +@[simp, circuit_norm] lemma Balance.multiset_pullBound : + (Balance.multiset : Balance F).pullBound = 1 := rfl + /- ## Conditions on channels that strengthen the implications of interaction balance. -/ namespace RawChannel diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index eb471eb7b..2b7385ec3 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -151,6 +151,15 @@ instance (N : ℕ) (channel : Channel F Message) : (channel.counted N).Consisten | push => exact b_reqs b_ne_zero | pull => exact absurd (b_reqs.resolve_left b_ne_zero) (b_not_pull rfl) +/-- The instance is found for the `pullBound` of a concrete balance check, without the user +having to `show (channel.counted N).ConsistentUpTo N` first. That needs +`Balance.logUpUpTo_pullBound`: instance search runs at reducible transparency and cannot project +`pullBound` out of the plain `def` `Balance.logUpUpTo`. -/ +example [ComputableRingChar F] (N : ℕ) (channel : Channel F Message) : + (channel.counted N).ConsistentUpTo (Balance.logUpUpTo (F := F) N).pullBound := by + simp only [Balance.logUpUpTo_pullBound] + infer_instance + /-- The counted channel is also consistent for balance checks with a larger bound. -/ theorem counted_consistentUpTo_of_le {M N : ℕ} (h : N ≤ M) (channel : Channel F Message) : (channel.counted N).ConsistentUpTo M where From fd167476c13938edc79b206cd28e0ca1b11c20e3 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 5 Nov 2026 18:36:39 -0300 Subject: [PATCH 05/29] Give matchedUpTo_of_balanced a default, so a plain balance check compiles again. Adding matchedUpTo_of_balanced to Balance without a default broke every existing definition of a Balance, although for the default pullBound = 1 the field is just BalancedInteractions.matchedUpTo_one applied to balancedInteractions_of_balanced. A default has to work around how Lean elaborates them. A term default is elaborated once with pullBound still abstract, so matchedUpTo_one does not typecheck; a tactic default is elaborated per instance, against the real pullBound, but without the earlier fields in scope, so it cannot name balancedInteractions_of_balanced. Handing the conclusion of balancedInteractions_of_balanced to matchedUpTo_of_balanced as an extra premise makes the tactic form work, and costs the two checks that prove the field themselves nothing, since they ignore it. Balance.balancedUpTo_of_balanced passes the weaker balance it already computes. Setting pullBound := N for N > 1 and omitting the field is a type error, as intended, not a silently wrong default. Balance.multiset now relies on the default. --- Clean/Air/Balance.lean | 25 +++++++++++++++++++------ 1 file changed, 19 insertions(+), 6 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index efcb03e97..cbb0cd664 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -174,14 +174,27 @@ structure Balance (F : Type) [FiniteField F] where Balanced interactions → BalancedInteractions interactions /-- The largest pull multiplicity that the check matches with a push. -/ pullBound : ℕ := 1 + /-- + Pulls of multiplicity up to `pullBound` are matched by a push. + + For the default `pullBound := 1` this is `BalancedInteractions.matchedUpTo_one`, so a check that + matches nothing beyond unit pulls does not have to provide it. That is why the conclusion of + `balancedInteractions_of_balanced` is a premise here: a structure's default value is either + elaborated once with `pullBound` still abstract (a term default, which then cannot be + `matchedUpTo_one`) or elaborated per instance but without the earlier fields in scope (a tactic + default). Passing the weaker balance in makes the second form work. A check that sets + `pullBound := N` for `N > 1` and omits this field gets a type error, not a wrong default. + -/ matchedUpTo_of_balanced : ∀ interactions, fits interactions.length → - Balanced interactions → MatchedUpTo pullBound interactions + Balanced interactions → BalancedInteractions interactions → + MatchedUpTo pullBound interactions := by + exact fun _ _ _ h_weak => h_weak.matchedUpTo_one lemma Balance.balancedUpTo_of_balanced (balance : Balance F) {interactions : List (Interaction F)} (fits : balance.fits interactions.length) (balanced : balance.Balanced interactions) : BalancedUpTo balance.pullBound interactions := - ⟨ balance.balancedInteractions_of_balanced _ fits balanced, - balance.matchedUpTo_of_balanced _ fits balanced ⟩ + have weak := balance.balancedInteractions_of_balanced _ fits balanced + ⟨ weak, balance.matchedUpTo_of_balanced _ fits balanced weak ⟩ /- ## LogUp: field-sum balance @@ -340,7 +353,7 @@ def Balance.logUp [ComputableRingChar F] : Balance F where have length_lt := ComputableRingChar.char_eq (R := F) ▸ of_decide_eq_true fits ⟨ balanced.exists_push_of_pull length_lt, balanced.activeCount_push_eq_activeCount_pull_of_isBool length_lt ⟩ - matchedUpTo_of_balanced _ fits balanced := + matchedUpTo_of_balanced _ fits balanced _ := have length_lt := ComputableRingChar.char_eq (R := F) ▸ of_decide_eq_true fits balanced.matchedUpTo (by simpa using length_lt) @@ -363,7 +376,7 @@ def Balance.logUpUpTo [ComputableRingChar F] (N : ℕ) : Balance F where ⟨ balanced.exists_push_of_pull length_lt, balanced.activeCount_push_eq_activeCount_pull_of_isBool length_lt ⟩ pullBound := N - matchedUpTo_of_balanced _ fits balanced := + matchedUpTo_of_balanced _ fits balanced _ := have length_mul_lt := ComputableRingChar.char_eq (R := F) ▸ (of_decide_eq_true (Bool.and_eq_true_iff.mp fits).2) balanced.matchedUpTo length_mul_lt @@ -410,7 +423,7 @@ def Balance.multiset : Balance F where Balanced := MultisetBalanced fits _ := true balancedInteractions_of_balanced _ _ h_check := h_check.balancedInteractions - matchedUpTo_of_balanced _ _ h_check := h_check.balancedInteractions.matchedUpTo_one + -- `matchedUpTo_of_balanced` is the default: `pullBound` is `1`. /-- `Balance.multiset` matches pulls of multiplicity `1` only. -/ @[simp, circuit_norm] lemma Balance.multiset_pullBound : From 4479c995a3ac4b250dd9c45020ebecd9d280fb30 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 5 Nov 2026 18:37:38 -0300 Subject: [PATCH 06/29] Put the pull bound after the subject in the top-level SoundChannels. SoundChannels took the bound N first and the tables second, while Air.Flat.Ensemble.SoundChannels takes the ensemble first and N second, as dot notation forces it to. The two now agree: subject, bound, finished channels. Only the definition and its six call sites change; nothing else about the property moves. --- Clean/Air/OrderedChannel.lean | 12 ++++++------ Clean/Air/Vm.lean | 2 +- 2 files changed, 7 insertions(+), 7 deletions(-) diff --git a/Clean/Air/OrderedChannel.lean b/Clean/Air/OrderedChannel.lean index 789fff0c3..e37fe2d85 100644 --- a/Clean/Air/OrderedChannel.lean +++ b/Clean/Air/OrderedChannel.lean @@ -460,14 +460,14 @@ Note: In the presence of "VM-like" channels, where a circuit both pushes and pul However, in that scenario, it is still useful to establish `SoundChannels` on the subset of non-VM tables. -/ @[circuit_norm] -def SoundChannels (N : ℕ) (tables : List α) (finished : List (RawChannel F)) : Prop := +def SoundChannels (tables : List α) (N : ℕ) (finished : List (RawChannel F)) : Prop := (∀ table ∈ tables, channelsWithGuarantees table ⊆ finished) ∧ (∀ channel ∈ finished, OrderedChannel channel tables) ∧ ∀ channel ∈ finished, channel.ConsistentUpTo N /-- `SoundChannels` lets us prove a soundness theorem. -/ theorem spec_and_guarantees_of_soundChannels {witness : Tables F} {finished : List (RawChannel F)} : - SoundChannels N (witness.tables.map (·.component)) finished → + SoundChannels (witness.tables.map (·.component)) N finished → -- constraints + partial balance witness.Assumptions → witness.Constraints → @@ -517,10 +517,10 @@ theorem spec_and_guarantees_of_soundChannels {witness : Tables F} {finished : Li /-- `SoundChannels` is strictly increasing: you can make the finished list bigger by any consistent channels -/ lemma soundChannels_of_soundChannels_subset {tables : List α} {finished finished' : List (RawChannel F)} : - SoundChannels N tables finished → + SoundChannels tables N finished → finished ⊆ finished' → (∀ channel ∈ finished', channel.ConsistentUpTo N) → - SoundChannels N tables finished' := by + SoundChannels tables N finished' := by rintro ⟨ subset_finished, ordered_channels, consistent_channels ⟩ finished'_subset finished'_consistent constructor · intro table h_table @@ -539,8 +539,8 @@ lemma soundChannels_of_soundChannels_subset {tables : List α} /-- You can add one channel to the finished list and preserve `SoundChannels` -/ lemma soundChannels_cons_of_soundChannels {tables : List α} {finished : List (RawChannel F)} {channel : RawChannel F} [channel.ConsistentUpTo N] : - SoundChannels N tables finished → - SoundChannels N tables (channel :: finished) := by + SoundChannels tables N finished → + SoundChannels tables N (channel :: finished) := by intro sound_channels apply soundChannels_of_soundChannels_subset sound_channels · simp diff --git a/Clean/Air/Vm.lean b/Clean/Air/Vm.lean index c6a9b99db..e0398cfd0 100644 --- a/Clean/Air/Vm.lean +++ b/Clean/Air/Vm.lean @@ -876,7 +876,7 @@ theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balan (witness.oldContext_constraints_of_constraints constraints) old_partial_balance table table_mem |>.right channel channel_mem |>.right - have old_sound_channels : _root_.SoundChannels balance.pullBound ens.tables finished := ⟨ + have old_sound_channels : _root_.SoundChannels ens.tables balance.pullBound finished := ⟨ ens.channelsWithGuarantees_subset_iff.mp soundChannels.left |>.right, fun channel h_channel => (soundChannels.right.left channel h_channel).right.left, soundChannels.right.right⟩ From 84ca439b375a8477b720d23315e453efb86dfb43 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 11 Nov 2026 12:48:30 -0300 Subject: [PATCH 07/29] Follow upstream's vector eval normal form in the counted channel's two guarantees. Rebasing these commits onto current main crosses a2ff3c47, "Complete the vector eval normal form", which changed where `eval` sits in a `ProvableType` goal. The two lemmas about a counted pull's guarantee and a counted push's requirement were written against the old shape and called `ProvableType.fromElements_eval_toElements`, which no longer fires: the goal now reads `fromElements (toElements (eval env msg))`, with the eval already inside, so the plain `ProvableType.fromElements_toElements` is what closes it. The linter said so before the error did, reporting the simp argument as unused. No statement changes, only the two simp calls that prove them. --- Clean/Air/CountedChannel.lean | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 2b7385ec3..5589728fd 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -101,7 +101,7 @@ variable {N : ℕ} {channel : Channel F Message} {mult : Expression F} (pulledUpTo N channel mult msg).Guarantees env ↔ (CountUpTo N (Expression.eval env mult) → channel.Guarantees (eval env msg) env.data) := by simp [pulledUpTo, AbstractInteraction.Guarantees, counted, - ProvableType.fromElements_eval_toElements] + ProvableType.fromElements_toElements] @[circuit_norm] lemma pulledUpTo_requirements : (pulledUpTo N channel mult msg).Requirements env ↔ @@ -116,7 +116,7 @@ variable {N : ℕ} {channel : Channel F Message} {mult : Expression F} (emittedUpTo N channel mult msg).Requirements env ↔ (Expression.eval env mult ≠ 0 → channel.Guarantees (eval env msg) env.data) := by simp [emittedUpTo, AbstractInteraction.Requirements, counted, - ProvableType.fromElements_eval_toElements] + ProvableType.fromElements_toElements] end omit [FiniteField F] in From 1bfba6de554a9a6e4ff8e48f51b14ee0ac075ccd Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 11 Nov 2026 15:17:11 -0300 Subject: [PATCH 08/29] Give balancedChannels_of_balanced the balance check's pull bound, so it keeps its callers. Allowing counted pulls made both proofs need `BalancedUpTo balance.pullBound` where they used to need `BalancedInteractions`, and each inlined the derivation as the same lambda over the ensemble's channels. That left `balancedChannels_of_balanced` and `BalancedChannels` without users, and the step they name duplicated in two files. `BalancedChannel` and `BalancedChannels` now take the bound, the lemma concludes `BalancedChannels balance.pullBound`, and both proofs call it again. --- Clean/Air/FlatEnsemble.lean | 18 ++++++++++-------- Clean/Air/OrderedChannel.lean | 3 +-- Clean/Air/Vm.lean | 3 +-- 3 files changed, 12 insertions(+), 12 deletions(-) diff --git a/Clean/Air/FlatEnsemble.lean b/Clean/Air/FlatEnsemble.lean index fb9a3873c..187efd45f 100644 --- a/Clean/Air/FlatEnsemble.lean +++ b/Clean/Air/FlatEnsemble.lean @@ -333,16 +333,18 @@ lemma interactionsWith_of_verifier_empty {ens : Ensemble F PublicIO} {witness : simp [interactionsWith, verifierInteractionsWith, Tables.interactionsWith, circuit_norm, h_verifier_empty, Verifier.Program.empty] -/-- The ensemble interactions with a particular channel are balanced. -/ +/-- The ensemble interactions with a particular channel are balanced, with pulls of multiplicity up +to `N` matched by a push. -/ @[circuit_norm] -abbrev BalancedChannel {ens : Ensemble F PublicIO} (witness : EnsembleWitness ens) +abbrev BalancedChannel {ens : Ensemble F PublicIO} (witness : EnsembleWitness ens) (N : ℕ) (channel : RawChannel F) : Prop := - BalancedInteractions (witness.interactionsWith channel) + BalancedUpTo N (witness.interactionsWith channel) -/-- All ensemble interactions with all ensemble channels are balanced. -/ +/-- All ensemble interactions with all ensemble channels are balanced, with pulls of multiplicity up +to `N` matched by a push. -/ @[circuit_norm] -def BalancedChannels {ens : Ensemble F PublicIO} (witness : EnsembleWitness ens) : Prop := - ∀ channel ∈ ens.channels, BalancedChannel witness channel +def BalancedChannels {ens : Ensemble F PublicIO} (witness : EnsembleWitness ens) (N : ℕ) : Prop := + ∀ channel ∈ ens.channels, BalancedChannel witness N channel /-- The number of rows of each table, in the order of the ensemble's components. -/ def heights {ens : Ensemble F PublicIO} (witness : EnsembleWitness ens) : List ℕ := @@ -360,9 +362,9 @@ def InteractionsFit {ens : Ensemble F PublicIO} (witness : EnsembleWitness ens) lemma balancedChannels_of_balanced {ens : Ensemble F PublicIO} {witness : EnsembleWitness ens} {balance : Balance F} (fit : witness.InteractionsFit balance) - (balanced : witness.Balanced balance) : witness.BalancedChannels := + (balanced : witness.Balanced balance) : witness.BalancedChannels balance.pullBound := fun channel h_channel => - balance.balancedInteractions_of_balanced _ (fit channel h_channel) (balanced channel h_channel) + balance.balancedUpTo_of_balanced (fit channel h_channel) (balanced channel h_channel) end EnsembleWitness /-- diff --git a/Clean/Air/OrderedChannel.lean b/Clean/Air/OrderedChannel.lean index e37fe2d85..ec21464ea 100644 --- a/Clean/Air/OrderedChannel.lean +++ b/Clean/Air/OrderedChannel.lean @@ -628,8 +628,7 @@ theorem tableSoundness_of_soundChannels {ens : Ensemble F PublicIO} {balance : B ens.TableSoundness balance := by intro ⟨ finished, finished_subset, subset_finished, ordered, consistent ⟩ witness assumptions constraints fit balanced - have balance := fun channel h_channel => - balance.balancedUpTo_of_balanced (fit channel h_channel) (balanced channel h_channel) + have balance := witness.balancedChannels_of_balanced fit balanced rw [channelsWithGuarantees_subset_iff] at subset_finished -- the tables are sound by themselves, given partial balance have table_results := spec_and_guarantees_of_soundChannels (witness := witness.toTables) diff --git a/Clean/Air/Vm.lean b/Clean/Air/Vm.lean index e0398cfd0..4ed76a381 100644 --- a/Clean/Air/Vm.lean +++ b/Clean/Air/Vm.lean @@ -814,8 +814,7 @@ theorem addVm_soundVmChannel_of_soundChannels (ens : Ensemble F PublicIO) {balan (ens.addVm vm fits).SoundVmChannel balance := by intro not_mem_vm_channel grts_subset reqs_disjoint (witness : VmWitness ens vm fits) assumptions constraints fit balanced - have balance_upTo := fun channel h_channel => - balance.balancedUpTo_of_balanced (fit channel h_channel) (balanced channel h_channel) + have balance_upTo := witness.balancedChannels_of_balanced fit balanced /- the high level plan is to apply `verifier_guarantees_of_requirements_of_requirements_of_guarantees`. From d1f0b7adf4b9b1884e9cec18995be3a1c69e7b56 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 11 Nov 2026 16:04:58 -0300 Subject: [PATCH 09/29] State the count of a pull once, drop the unrelated multiset refactor, and fix two docstrings. `Interaction.PullUpTo` and `PushUpTo` inlined the same existential that `Channel.CountUpTo` defines, and the counted channel's `ConsistentUpTo` instance relied on the two being definitionally equal to build a `PullUpTo` from a `CountUpTo`. `CountUpTo` now lives next to the interaction predicates that use it, with `countUpTo_one_iff` in place of `exists_le_one_iff`, so that bridge is stated rather than coincidental. `Balance.multiset` goes back to its inline `Balanced` field: extracting `MultisetBalanced` was not needed for counted pulls, and the default `matchedUpTo_of_balanced` does not depend on the field being named. Both docstrings said "fewer than `p / N` interactions"; the hypothesis is `interactions.length * N < ringChar F`, which is not the same bound. --- Clean/Air/Balance.lean | 79 +++++++++++++++++------------------ Clean/Air/CountedChannel.lean | 3 -- 2 files changed, 38 insertions(+), 44 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index cbb0cd664..eb03bd005 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -20,6 +20,18 @@ soundness theory uses, and a `Balance` is a check that a verifier performs on th with a channel, together with the proof that the check implies these facts. -/ +/-- A multiplicity of `k` for some `1 ≤ k ≤ N`: what a table that answers `k` identical lookups +with one row pulls the facts it relies on with. `CountUpTo 1` is multiplicity `1`. -/ +def CountUpTo (N : ℕ) (mult : F) : Prop := ∃ k : ℕ, 1 ≤ k ∧ k ≤ N ∧ mult = k + +lemma countUpTo_one_iff {mult : F} : CountUpTo 1 mult ↔ mult = 1 := by + constructor + · rintro ⟨k, hk1, hk2, h⟩ + obtain rfl : k = 1 := by omega + simpa using h + · intro h + exact ⟨1, le_rfl, le_rfl, by simp [h]⟩ + namespace Interaction /-- A pull of multiplicity `1`: the interactions that typed channels give their guarantee. -/ def Pull (i : Interaction F) : Prop := i.direction = .pull ∧ i.mult = 1 @@ -39,27 +51,18 @@ lemma not_push_iff {i : Interaction F} : ¬i.Push ↔ i.mult = 0 ∨ i.Pull := b /-- A pull of multiplicity `k` for some `1 ≤ k ≤ N`: a table that answers `k` identical lookups with one row pulls the facts it relies on with the same multiplicity. `Pull` is `PullUpTo 1`. -/ def PullUpTo (N : ℕ) (i : Interaction F) : Prop := - i.direction = .pull ∧ ∃ k : ℕ, 1 ≤ k ∧ k ≤ N ∧ i.mult = k + i.direction = .pull ∧ CountUpTo N i.mult /-- An interaction that pushes its message when pulls of multiplicity up to `N` are allowed: an active push, or a pull of another nonzero multiplicity. `Push` is `PushUpTo 1`. -/ def PushUpTo (N : ℕ) (i : Interaction F) : Prop := - i.mult ≠ 0 ∧ (i.direction = .pull → ¬∃ k : ℕ, 1 ≤ k ∧ k ≤ N ∧ i.mult = k) - -lemma exists_le_one_iff {i : Interaction F} : - (∃ k : ℕ, 1 ≤ k ∧ k ≤ 1 ∧ i.mult = k) ↔ i.mult = 1 := by - constructor - · rintro ⟨k, hk1, hk2, h⟩ - obtain rfl : k = 1 := by omega - simpa using h - · intro h - exact ⟨1, le_rfl, le_rfl, by simp [h]⟩ + i.mult ≠ 0 ∧ (i.direction = .pull → ¬CountUpTo N i.mult) lemma pullUpTo_one_iff {i : Interaction F} : i.PullUpTo 1 ↔ i.Pull := by - simp only [PullUpTo, Pull, exists_le_one_iff] + simp only [PullUpTo, Pull, countUpTo_one_iff] lemma pushUpTo_one_iff {i : Interaction F} : i.PushUpTo 1 ↔ i.Push := by - simp only [PushUpTo, Push, exists_le_one_iff] + simp only [PushUpTo, Push, countUpTo_one_iff] lemma not_pushUpTo_iff {N : ℕ} {i : Interaction F} : ¬i.PushUpTo N ↔ i.mult = 0 ∨ i.PullUpTo N := by simp only [PushUpTo, PullUpTo, not_and, Classical.not_imp, not_not] @@ -118,8 +121,8 @@ lemma BalancedInteractions.nil : BalancedInteractions ([] : List (Interaction F) Every pull of multiplicity up to `N` is matched by an interaction that pushes its message. For `N = 1` this is `BalancedInteractions.exists_push_of_pull`. For larger `N` it needs more than balance of the counts: under field-sum balance, `(p - 1) / 2` pulls of multiplicity `2` and one pull of -multiplicity `1` sum to `p ≡ 0` with no push. It holds when there are fewer than `p / N` -interactions, see `LogUpBalanced.matchedUpTo`. +multiplicity `1` sum to `p ≡ 0` with no push. It holds when the number of interactions times `N` +is below the characteristic, see `LogUpBalanced.matchedUpTo`. -/ def MatchedUpTo (N : ℕ) (interactions : List (Interaction F)) : Prop := ∀ a ∈ interactions, a.PullUpTo N → ∃ b ∈ interactions, b.msg = a.msg ∧ b.PushUpTo N @@ -323,9 +326,9 @@ lemma balanceOf_eq_neg_of_pullsUpTo {N : ℕ} {msg : Array F} (hN : N < ringChar ring /-- -Every pull of multiplicity up to `N` is matched by a push, when there are fewer than -`ringChar F / N` interactions: otherwise the multiplicities of the pulls of a message could add up -to a multiple of the characteristic. +Every pull of multiplicity up to `N` is matched by a push, when the number of interactions times +`N` is below the characteristic: otherwise the multiplicities of the pulls of a message could add +up to a multiple of the characteristic. -/ theorem LogUpBalanced.matchedUpTo {N : ℕ} {interactions : List (Interaction F)} (balanced : LogUpBalanced interactions) (length_mul_lt : interactions.length * N < ringChar F) : @@ -398,31 +401,25 @@ permutation of the messages of the active pulls. The permutation gives equal cou of interactions; the soundness error of a randomized check of it, like a product argument, is up to the proof system. -/ -def MultisetBalanced (interactions : List (Interaction F)) : Prop := - (∀ i ∈ interactions, IsBool i.mult) ∧ - (activeMessages interactions .push).Perm (activeMessages interactions .pull) - -theorem MultisetBalanced.balancedInteractions {interactions : List (Interaction F)} - (h_check : MultisetBalanced interactions) : BalancedInteractions interactions := by - have counts (msg : Array F) : - activeCount interactions .push msg = activeCount interactions .pull msg := by - rw [← count_activeMessages, ← count_activeMessages] - exact h_check.2.count_eq msg - refine ⟨ ?_, fun msg _ => counts msg ⟩ - intro a a_mem a_pull - have pull_pos : 0 < activeCount interactions .pull a.msg := by - apply List.countP_pos_iff.mpr - exact ⟨ a, a_mem, by simp [a_pull.1, a_pull.2] ⟩ - rw [← counts] at pull_pos - obtain ⟨ b, b_mem, h_b ⟩ := List.countP_pos_iff.mp pull_pos - simp only [decide_eq_true_eq] at h_b - exact ⟨ b, b_mem, h_b.2.1, h_b.2.2, by simp [h_b.1] ⟩ - -/-- Multiset balance with `0`/`1` activations. -/ def Balance.multiset : Balance F where - Balanced := MultisetBalanced + Balanced interactions := + (∀ i ∈ interactions, IsBool i.mult) ∧ + (activeMessages interactions .push).Perm (activeMessages interactions .pull) fits _ := true - balancedInteractions_of_balanced _ _ h_check := h_check.balancedInteractions + balancedInteractions_of_balanced interactions _ h_check := by + have counts (msg : Array F) : + activeCount interactions .push msg = activeCount interactions .pull msg := by + rw [← count_activeMessages, ← count_activeMessages] + exact h_check.2.count_eq msg + refine ⟨ ?_, fun msg _ => counts msg ⟩ + intro a a_mem a_pull + have pull_pos : 0 < activeCount interactions .pull a.msg := by + apply List.countP_pos_iff.mpr + exact ⟨ a, a_mem, by simp [a_pull.1, a_pull.2] ⟩ + rw [← counts] at pull_pos + obtain ⟨ b, b_mem, h_b ⟩ := List.countP_pos_iff.mp pull_pos + simp only [decide_eq_true_eq] at h_b + exact ⟨ b, b_mem, h_b.2.1, h_b.2.2, by simp [h_b.1] ⟩ -- `matchedUpTo_of_balanced` is the default: `pullBound` is `1`. /-- `Balance.multiset` matches pulls of multiplicity `1` only. -/ diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 5589728fd..21b79638e 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -20,9 +20,6 @@ variable {Message : TypeMap} [ProvableType Message] namespace Channel -/-- `k` for some `1 ≤ k ≤ N`. -/ -def CountUpTo (N : ℕ) (mult : F) : Prop := ∃ k : ℕ, 1 ≤ k ∧ k ≤ N ∧ mult = k - /-- `channel`, with pulls of multiplicity up to `N`. -/ def counted (N : ℕ) (channel : Channel F Message) : RawChannel F where name := channel.name From 36cbaf3e5bc7873ba32d462b1eb9749615e23015 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 11 Nov 2026 17:28:18 -0300 Subject: [PATCH 10/29] Say that counting up to one is the channel Clean already had, and project the new interactions. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Every other layer states its `N = 1` coincidence — `pullUpTo_one_iff`, `pushUpTo_one_iff`, `matchedUpTo_one`, `logUp_pullBound` — but nothing said that `Channel.counted 1` is `Channel.toRaw`. `counted_one_guarantees` and `counted_one_requirements` state it field by field, which is the conservativity a reader checks first. `pulledUpTo` and `emittedUpTo` had `circuit_norm` projections for `channel` and `direction` but not for `mult` and `msg`, so simp could not reduce either, unlike the `pulled`/`pushed` family they mirror. --- Clean/Air/CountedChannel.lean | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 21b79638e..2290c431f 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -43,6 +43,23 @@ from the other channels of a circuit without unfolding them. -/ @[circuit_norm] lemma counted_arity (N : ℕ) (channel : Channel F Message) : (channel.counted N).arity = size Message := rfl +/-- `counted 1` gives what the plain typed channel gives: only a pull of multiplicity `1` gets the +guarantee. -/ +lemma counted_one_guarantees (channel : Channel F Message) (mult : F) + (message : Vector F (size Message)) (data : ProverData F) : + (channel.counted 1).Guarantees mult message data ↔ + channel.toRaw.Guarantees mult message data := by + simp [counted, toRaw, countUpTo_one_iff] + +/-- `counted 1` owes what the plain typed channel owes: an active push owes the guarantee, and a +pull has multiplicity `0` or `1`. Together with `counted_one_guarantees`, counting up to `1` is the +channel Clean already had. -/ +lemma counted_one_requirements (channel : Channel F Message) (direction : Direction) (mult : F) + (message : Vector F (size Message)) (data : ProverData F) : + (channel.counted 1).Requirements direction mult message data ↔ + channel.toRaw.Requirements direction mult message data := by + cases direction <;> simp [counted, toRaw, countUpTo_one_iff] + /-- The interaction of pulling `msg` `count` times from `channel.counted N`. -/ def pulledUpTo (N : ℕ) (channel : Channel F Message) (count : Expression F) (msg : Message (Expression F)) : AbstractInteraction F := @@ -89,10 +106,18 @@ variable {N : ℕ} {channel : Channel F Message} {mult : Expression F} (pulledUpTo N channel mult msg).channel = channel.counted N := rfl @[circuit_norm] lemma pulledUpTo_direction : (pulledUpTo N channel mult msg).direction = .pull := rfl +@[circuit_norm] lemma pulledUpTo_mult : + (pulledUpTo N channel mult msg).mult = mult := rfl +@[circuit_norm] lemma pulledUpTo_msg : + (pulledUpTo N channel mult msg).msg = toElements msg := rfl @[circuit_norm] lemma emittedUpTo_channel : (emittedUpTo N channel mult msg).channel = channel.counted N := rfl @[circuit_norm] lemma emittedUpTo_direction : (emittedUpTo N channel mult msg).direction = .push := rfl +@[circuit_norm] lemma emittedUpTo_mult : + (emittedUpTo N channel mult msg).mult = mult := rfl +@[circuit_norm] lemma emittedUpTo_msg : + (emittedUpTo N channel mult msg).msg = toElements msg := rfl @[circuit_norm] lemma pulledUpTo_guarantees : (pulledUpTo N channel mult msg).Guarantees env ↔ From 6930551cf0a98a53984b057758bf2e267e04ad31 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 12 Nov 2026 10:29:01 -0300 Subject: [PATCH 11/29] Let synthesis project the pull bound, so a counted channel can be finished. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit `SoundEnsemble.markFinished` takes `[channel.ConsistentUpTo balance.pullBound]` as an instance argument. The counted channel's instance ties that bound to the channel's own `N`, so synthesis has to see `(Balance.logUpUpTo N).pullBound` as `N`. It cannot unfold a plain `def`, and an instance argument leaves no place to `simp` first, so finishing a counted channel failed to elaborate. A plain channel was unaffected: its instance comes from `[Consistent] → ConsistentUpTo N` with `N` free, which matches an unreduced projection. `Balance.logUpUpTo` is now reducible, which is what synthesis needs. The `simp` step in the example above `counted_consistentUpTo_of_le` is no longer needed, and the lemma's docstring no longer claims to solve a problem that a `simp` lemma cannot reach. --- Clean/Air/Balance.lean | 15 ++++++++++----- Clean/Air/CountedChannel.lean | 12 +++++------- 2 files changed, 15 insertions(+), 12 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index eb03bd005..868939f84 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -369,8 +369,14 @@ its own: `Balance.logUp` is a plain `def`, so the projection is not reducible. - /-- LogUp with counted pulls: field sums, sound for interaction counts whose product with `N` is below the characteristic, so that pulls of multiplicity up to `N` are matched by a push. + +Reducible so that instance search can project `pullBound` out of it. +`SoundEnsemble.markFinished` takes `[channel.ConsistentUpTo balance.pullBound]` as an instance +argument, and the counted channel's instance ties that bound to the channel's own `N`, so synthesis +has to see `(logUpUpTo N).pullBound` as `N`. It cannot unfold a plain `def`, and an instance +argument leaves no place to `simp` first. -/ -def Balance.logUpUpTo [ComputableRingChar F] (N : ℕ) : Balance F where +@[reducible] def Balance.logUpUpTo [ComputableRingChar F] (N : ℕ) : Balance F where Balanced := LogUpBalanced fits count := count < ComputableRingChar.char F && count * N < ComputableRingChar.char F balancedInteractions_of_balanced _ fits balanced := @@ -384,10 +390,9 @@ def Balance.logUpUpTo [ComputableRingChar F] (N : ℕ) : Balance F where (of_decide_eq_true (Bool.and_eq_true_iff.mp fits).2) balanced.matchedUpTo length_mul_lt -/-- `Balance.logUpUpTo N` matches pulls of multiplicity up to `N`. Without this lemma a user of a -`SoundEnsemble` over `Balance.logUpUpTo N` has to write `show channel.ConsistentUpTo N` by hand, -because instance search works at reducible transparency and cannot project `pullBound` out of the -plain `def`. -/ +/-- `Balance.logUpUpTo N` matches pulls of multiplicity up to `N`, for `simp` and `circuit_norm`. +Instance search needs the same fact, and gets it from the check being `@[reducible]`: a `simp` +lemma plays no part in synthesis. -/ @[simp, circuit_norm] lemma Balance.logUpUpTo_pullBound [ComputableRingChar F] (N : ℕ) : (Balance.logUpUpTo N : Balance F).pullBound = N := rfl diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 2290c431f..6577d94aa 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -173,14 +173,12 @@ instance (N : ℕ) (channel : Channel F Message) : (channel.counted N).Consisten | push => exact b_reqs b_ne_zero | pull => exact absurd (b_reqs.resolve_left b_ne_zero) (b_not_pull rfl) -/-- The instance is found for the `pullBound` of a concrete balance check, without the user -having to `show (channel.counted N).ConsistentUpTo N` first. That needs -`Balance.logUpUpTo_pullBound`: instance search runs at reducible transparency and cannot project -`pullBound` out of the plain `def` `Balance.logUpUpTo`. -/ +/-- Synthesis finds the instance for the `pullBound` of a concrete balance check on its own. That is +what `SoundEnsemble.markFinished` needs, since it takes the instance as an instance argument, where +no `simp` step can run first. -/ example [ComputableRingChar F] (N : ℕ) (channel : Channel F Message) : - (channel.counted N).ConsistentUpTo (Balance.logUpUpTo (F := F) N).pullBound := by - simp only [Balance.logUpUpTo_pullBound] - infer_instance + (channel.counted N).ConsistentUpTo (Balance.logUpUpTo (F := F) N).pullBound := + inferInstance /-- The counted channel is also consistent for balance checks with a larger bound. -/ theorem counted_consistentUpTo_of_le {M N : ℕ} (h : N ≤ M) (channel : Channel F Message) : From 7ff8fa70ddd4b884796cbd8351ec8d8050aff691 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 12 Nov 2026 10:53:26 -0300 Subject: [PATCH 12/29] Name the balance lemma after the predicate it assumes. `REVIEWERS.md` asks each underscore-separated part of a theorem name to correspond to a declaration or a standard logical word. The predicate is `Interaction.PullUpTo`, so the plural `pullsUpTo` named nothing. --- Clean/Air/Balance.lean | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 868939f84..940912d1f 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -289,14 +289,14 @@ If every interaction carrying `msg` is inactive or a pull of multiplicity at mos of `msg` is minus a natural number at most `length · N`, which is positive if one of them is an active pull. -/ -lemma balanceOf_eq_neg_of_pullsUpTo {N : ℕ} {msg : Array F} (hN : N < ringChar F) : +lemma balanceOf_eq_neg_of_pullUpTo {N : ℕ} {msg : Array F} (hN : N < ringChar F) : ∀ {interactions : List (Interaction F)}, (∀ i ∈ interactions, i.msg = msg → i.mult = 0 ∨ i.PullUpTo N) → ∃ S : ℕ, balanceOf interactions msg = -(S : F) ∧ S ≤ interactions.length * N ∧ ∀ a ∈ interactions, a.msg = msg → a.PullUpTo N → 1 ≤ S | [], _ => ⟨ 0, by simp [balanceOf], by simp, by simp ⟩ | i :: is, h => by - obtain ⟨ S, hS, hle, hpos ⟩ := balanceOf_eq_neg_of_pullsUpTo hN (interactions := is) + obtain ⟨ S, hS, hle, hpos ⟩ := balanceOf_eq_neg_of_pullUpTo hN (interactions := is) fun j hj => h j (List.mem_cons_of_mem i hj) rw [balanceOf_cons] simp only [List.length_cons] @@ -339,7 +339,7 @@ theorem LogUpBalanced.matchedUpTo {N : ℕ} {interactions : List (Interaction F) by_contra! no_push have hN : N < ringChar F := lt_of_le_of_lt (Nat.le_mul_of_pos_left N (List.length_pos_of_mem a_mem)) length_mul_lt - obtain ⟨ S, hS, hle, hpos ⟩ := balanceOf_eq_neg_of_pullsUpTo hN (msg := a.msg) + obtain ⟨ S, hS, hle, hpos ⟩ := balanceOf_eq_neg_of_pullUpTo hN (msg := a.msg) (interactions := interactions) fun i i_mem i_msg => Interaction.not_pushUpTo_iff.mp (no_push i i_mem i_msg) have S_pos := hpos a a_mem rfl a_pull From f1f542da399f6e772eaf2e74098cc6b3baba249a Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 12 Nov 2026 10:58:01 -0300 Subject: [PATCH 13/29] Put countUpTo_of_val next to the definition it is about. It states a `CountUpTo` fact and sat in `namespace Channel`, where `CountUpTo` no longer lives, so it read as `Channel.countUpTo_of_val` about a root definition. Moving it leaves the lemma itself unchanged. --- Clean/Air/Balance.lean | 7 +++++++ Clean/Air/CountedChannel.lean | 7 ------- 2 files changed, 7 insertions(+), 7 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 940912d1f..7d766f735 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -32,6 +32,13 @@ lemma countUpTo_one_iff {mult : F} : CountUpTo 1 mult ↔ mult = 1 := by · intro h exact ⟨1, le_rfl, le_rfl, by simp [h]⟩ +omit [FiniteField F] in +/-- A nonzero multiplicity whose value is at most `N`. -/ +lemma countUpTo_of_val {p : ℕ} [Fact p.Prime] {N : ℕ} {mult : ZMod p} (h0 : mult ≠ 0) + (hN : mult.val ≤ N) : CountUpTo N mult := + ⟨ mult.val, Nat.one_le_iff_ne_zero.mpr ((ZMod.val_ne_zero mult).mpr h0), hN, + (ZMod.natCast_zmod_val mult).symm ⟩ + namespace Interaction /-- A pull of multiplicity `1`: the interactions that typed channels give their guarantee. -/ def Pull (i : Interaction F) : Prop := i.direction = .pull ∧ i.mult = 1 diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 6577d94aa..ef2751086 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -141,13 +141,6 @@ variable {N : ℕ} {channel : Channel F Message} {mult : Expression F} ProvableType.fromElements_toElements] end -omit [FiniteField F] in -/-- A nonzero multiplicity whose value is at most `N`. -/ -lemma countUpTo_of_val {p : ℕ} [Fact p.Prime] {N : ℕ} {mult : ZMod p} (h0 : mult ≠ 0) - (hN : mult.val ≤ N) : CountUpTo N mult := - ⟨ mult.val, Nat.one_le_iff_ne_zero.mpr ((ZMod.val_ne_zero mult).mpr h0), hN, - (ZMod.natCast_zmod_val mult).symm ⟩ - /-- A pull of multiplicity up to `N` is matched by a push, whose requirement gives the guarantee. Every other interaction gets no guarantee. -/ instance (N : ℕ) (channel : Channel F Message) : (channel.counted N).ConsistentUpTo N where From e496485d255746f8a3f8c6e67e2ffeaac567f518 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 12 Nov 2026 11:53:14 -0300 Subject: [PATCH 14/29] Exercise the counted channel end to end, in an example. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Nothing in `Clean/Examples/` used `Channel.counted`, so the API had no in-tree consumer: its ergonomics were untested, and the `pullBound` plumbing in `markFinished` and `tableSoundness_of_soundChannels` was reached by no example. `CountedSquares` is the smallest honest use. A squaring chip constrains `y = x * x` and pushes the pair as often as it is pulled; a consumer that needs the same square once or twice pulls it once with that multiplicity, and `tableSoundness` concludes for the assembled ensemble under `Balance.logUpUpTo 2`. The consumer's constraint on its count is the point: it discharges the pull's requirement `count = 0 ∨ CountUpTo 2 count`, which is what a counted channel asks of a pull before handing it the guarantee. Generation is not exercised yet, and the file says so. A following commit seeds the example so that it generates. --- Clean.lean | 1 + Clean/Examples/CountedSquares.lean | 169 +++++++++++++++++++++++++++++ 2 files changed, 170 insertions(+) create mode 100644 Clean/Examples/CountedSquares.lean diff --git a/Clean.lean b/Clean.lean index 48c5334da..e0aea6554 100644 --- a/Clean.lean +++ b/Clean.lean @@ -47,6 +47,7 @@ public import Clean.Gadgets.BLAKE3.FinalStateUpdate public import Clean.Gadgets.BLAKE3.Round public import Clean.Gadgets.BLAKE3.FinalizeChunk public import Clean.Examples.FibonacciVm.Ensemble +public import Clean.Examples.CountedSquares public import Clean.Examples.OpcodeVm.Ensemble public import Clean.Examples.BinaryFibonacci.Ensemble public import Clean.Gadgets.SHA256.SHA256Compress diff --git a/Clean/Examples/CountedSquares.lean b/Clean/Examples/CountedSquares.lean new file mode 100644 index 000000000..a608b00cd --- /dev/null +++ b/Clean/Examples/CountedSquares.lean @@ -0,0 +1,169 @@ +module + +public import Clean.Air.CountedChannel +public import Clean.Air.OrderedChannel +public import Clean.Air.WitnessGeneration + +@[expose] public section + +/-! +# A fact pulled with multiplicity + +A row that needs the same fact twice can pull it twice, or pull it once with multiplicity `2`. The +second costs one interaction instead of two, and is what `Channel.counted` is for. + +The fact here is `y = x * x`. The squaring chip enforces it with a constraint and pushes the pair; +a consumer that needs the same square once or twice pulls it once with that multiplicity. +`squares.counted 2` hands its guarantee to a pull of multiplicity `1` or `2`, and asks such a pull +to say which. That obligation is what the consumer discharges by constraining its count, and it is +what makes the channel sound: `Channel.counted_not_consistent` shows that a row free to claim any +multiplicity would break it. + +What this example does not cover: generation never runs. A `SoundEnsemble` that is not a VM has no +verifier in this version — only `addVm` attaches one — so nothing creates the first interaction and +the generator returns an empty witness. The generation modes below are therefore stated so that the +components are well formed, and are not exercised. A row's multiplicity is counted by +`Tally.signedCount`, which reads it as a natural number, so a counted pull is tallied as the number +it claims; that part needs no change for counted channels. + +Constraining the count arithmetically only scales for a small bound. A circuit counting to `510` +would range-check its multiplicity on another channel instead, which needs the finishing order of +`Ensemble.tableSoundness_of_stagedOn` rather than the single finished channel here. +-/ + +namespace CountedSquares + +open Air.Flat + +variable {p : ℕ} [Fact p.Prime] + +/-- Pulls may count up to `2`. -/ +abbrev N : ℕ := 2 + +/-- Pairs `(x, y)` with `y = x * x`. -/ +def squares : Channel (F p) (ProvablePair field field) where + name := "squares" + Guarantees | (x, y), _ => y = x * x + +/-- A count of `1` or `2` is a count up to `2`: the step from a field constraint to the channel's +`CountUpTo`, and the only arithmetic the example needs. -/ +lemma countUpTo_of_one_or_two {c : F p} (h : (c - 1) * (c - 2) = 0) : CountUpTo N c := by + rcases mul_eq_zero.mp h with h | h <;> rw [sub_eq_zero] at h + · exact ⟨1, le_rfl, by norm_num, by simp [h]⟩ + · exact ⟨2, by norm_num, le_rfl, by simp [h]⟩ + +structure ProviderInput (F : Type) where + x : F + y : F + /-- how many pulls this row answers -/ + mult : F +deriving ProvableStruct + +/-- The squaring chip: constrains `y = x * x` and pushes the pair as often as it is pulled. It owes +the guarantee on every active push, and its own constraint is what pays for it, so the chip assumes +nothing. -/ +def provider : GeneralFormalCircuit (F p) ProviderInput unit where + name := "squares-provider" + main input := do + input.y === input.x * input.x + squares.emitUpTo N input.mult (input.x, input.y) + + channelsWithRequirements := [squares.counted N] + Spec _ _ _ := True + ProverAssumptions input _ _ := input.y = input.x * input.x + + soundness := by + circuit_proof_start [squares] + simp_all + completeness := by + circuit_proof_start [squares] + simp_all + +structure ConsumerInput (F : Type) where + x : F + y : F + /-- how many times this row needs the fact: `1` or `2` -/ + count : F +deriving ProvableStruct + +/-- A row that needs `y = x * x` once or twice, and pulls it once with that multiplicity. + +The constraint on `count` is not bookkeeping: it is what discharges the pull's requirement +`count = 0 ∨ CountUpTo 2 count`. -/ +def consumer : GeneralFormalCircuit (F p) ConsumerInput unit where + name := "squares-consumer" + main input := do + assertZero ((input.count - 1) * (input.count - 2)) + squares.pullUpTo N input.count (input.x, input.y) + + Spec input _ _ := input.y = input.x * input.x + ProverAssumptions input _ _ := + (input.count = 1 ∨ input.count = 2) ∧ input.y = input.x * input.x + + requirementsChannelsLawful := by + simp only [circuit_norm, seval] + intro input env h + exact .inr (countUpTo_of_one_or_two h) + + soundness := by + circuit_proof_start [squares] + obtain ⟨h_count, h_guarantee⟩ := h_holds + have h_in := countUpTo_of_one_or_two h_count + exact ⟨h_guarantee h_in, .inr h_in⟩ + completeness := by + circuit_proof_start [squares] + obtain ⟨h_count, h_square⟩ := h_assumptions + refine ⟨?_, fun _ => h_square⟩ + rcases h_count with h | h <;> simp [h] + +/-- One provider row per square that is pulled, with the number of pulls as its multiplicity. -/ +abbrev providerComponent : Component (F p) where + circuit := provider + generation := .demand { + trigger := .pull + channel := squares + input | (x, y) => pure { x, y, mult := .multiplicity } + } + +/-- Four fixed consumer rows, each needing `4 = 2 * 2` twice. These are the rows that start +generation: they pull, and the squaring chip creates rows to answer them. A consumer driven by +demand on the same channel would feed the provider that feeds it. -/ +abbrev consumerComponent : Component (F p) where + circuit := consumer + generation := .preallocated { + channel := squares + trigger := .push + height := 4 + input _ := pure { x := .fixed 2, y := .fixed 4, count := .fixed 2 } + rowOf _ := 0 + } + +/-- The ensemble: the counted channel, the chip that serves it, the channel finished, and the chip +that reads it. `markFinished` is where the counted channel has to be `ConsistentUpTo` the balance +check's `pullBound`, which is why `Balance.logUpUpTo` must be reducible. -/ +def soundEnsemble : SoundEnsemble (F p) unit (Balance.logUpUpTo N) := + SoundEnsemble.empty (F p) unit (Balance.logUpUpTo N) + |>.addChannel (squares.counted N) + |>.addTable providerComponent + (by simp [circuit_norm, provider]) + (by simp [circuit_norm, provider]) + (by simp [circuit_norm, provider]) + |>.markFinished (squares.counted N) + |>.addTable consumerComponent + (by simp [circuit_norm, consumer]) + (by simp [circuit_norm, consumer]) + (by simp [circuit_norm, consumer, provider]) + +/-- The point of the example: every table of the ensemble is sound under `Balance.logUpUpTo 2`. For +the consumer that reads `y = x * x` out of a pull of multiplicity `2`, this is what says the fact it +read was really pushed by the squaring chip. + +The balance check's `fits` is where the bound of `LogUpBalanced.matchedUpTo` is paid: a witness only +satisfies `Ensemble.Statement` when the interaction count times `2` stays below the characteristic. +-/ +theorem tableSoundness : + (soundEnsemble (p := p)).ensemble.TableSoundness (Balance.logUpUpTo N) := + Ensemble.tableSoundness_of_soundChannels + ⟨_, (soundEnsemble (p := p)).finished_subset, (soundEnsemble (p := p)).soundChannels⟩ + +end CountedSquares From 2ee0027991a9ca378656037cde5ce726d32d611e Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Fri, 13 Nov 2026 09:29:34 -0300 Subject: [PATCH 15/29] Relate the counted definitions to the ones they generalize. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit `BalancedUpTo`, `Balance.logUpUpTo` and `RawChannel.ConsistentUpTo` each carry a bound that Clean's own definitions do not have, and nothing in the tree said what happens at bound one. A reader had to unfold three definitions to be convinced that a channel which does not count behaves exactly as it did before. `balancedUpTo_one_iff`, `Balance.logUpUpTo_one` and `consistentUpTo_one_iff` state it instead, so the claim that this generalizes Clean rather than replacing it is checked rather than asserted. `Balance.logUpUpTo` carries `count * N < ringChar F` in its `fits` field, where `Balance.logUp` only needs `count < ringChar F`. The factor reads like slack. `exists_logUpBalanced_not_matchedUpTo` shows it is not: once `ringChar F ≤ 2 * N`, two pulls whose multiplicities sum to the characteristic balance every message, the list is shorter than the characteristic, and no push occurs anywhere. Without the factor, `LogUpBalanced` would accept a balance that nothing ever pushed into. `Channel.counted_one_guarantees` and `counted_one_requirements` both state an `iff`, and `REVIEWERS.md` asks each part of a name to correspond to a declaration or a standard logical word. They are now `..._iff`. Neither is used outside its own docstring. --- Clean/Air/Balance.lean | 70 +++++++++++++++++++++++++++++++++++ Clean/Air/CountedChannel.lean | 6 +-- 2 files changed, 73 insertions(+), 3 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 7d766f735..225e6d379 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -172,6 +172,13 @@ lemma BalancedUpTo.of_perm {N : ℕ} {as bs : List (Interaction F)} : lemma BalancedUpTo.nil (N : ℕ) : BalancedUpTo N ([] : List (Interaction F)) := ⟨ .nil, .nil N ⟩ +/-- At bound `1`, balance with matched pulls is `BalancedInteractions`: a channel that does not +count keeps the balance premise its soundness proofs had before counted pulls. -/ +lemma balancedUpTo_one_iff {interactions : List (Interaction F)} : + BalancedUpTo 1 interactions ↔ BalancedInteractions interactions := + ⟨ fun balanced => balanced.toBalancedInteractions, + fun balanced => ⟨ balanced, balanced.matchedUpTo_one ⟩ ⟩ + /-- A balance check: the condition that a verifier checks on the interactions with each channel, the numbers of interactions with a channel that the check is sound for, and its proof that the condition @@ -355,6 +362,48 @@ theorem LogUpBalanced.matchedUpTo {N : ℕ} {interactions : List (Interaction F) (lt_of_le_of_lt (Nat.zero_le _) length_mul_lt)).mp (by simpa using hS) omega +/-- +The bound of `matchedUpTo` is necessary, not an artifact of its proof: once the number of +interactions times `N` reaches the characteristic, the multiplicities of the pulls of a message can +add up to a multiple of it, and the field sums see a balanced list in which nothing was pushed. + +Two pulls, of multiplicities `N` and `ringChar F - N`, are such a list whenever `N < ringChar F` +and `ringChar F ≤ 2 * N`: both multiplicities count up to `N`, they sum to the characteristic, and +a list of length `2` passes a bound on the length alone, which is all `Balance.logUp` asks. This is +what `Balance.logUpUpTo N` rules out by asking for `count * N < ringChar F` as well. +-/ +theorem exists_logUpBalanced_not_matchedUpTo {N : ℕ} (hN : 2 ≤ N) (h_lt : N < ringChar F) + (h_le : ringChar F ≤ 2 * N) (channel : RawChannel F) (msg : Array F) + (h_size : msg.size = channel.arity) : + ∃ interactions : List (Interaction F), LogUpBalanced interactions ∧ + interactions.length < ringChar F ∧ ¬ MatchedUpTo N interactions := by + have h_one_le : 1 ≤ ringChar F - N := by omega + have h_le_N : ringChar F - N ≤ N := by omega + have h_sum : (N : F) + ((ringChar F - N : ℕ) : F) = 0 := by + rw [← Nat.cast_add, show N + (ringChar F - N) = ringChar F by omega] + exact (ringChar.spec F _).mpr dvd_rfl + have h_count_rest : CountUpTo N ((ringChar F - N : ℕ) : F) := + ⟨ ringChar F - N, h_one_le, h_le_N, rfl ⟩ + let mk (mult : F) : Interaction F := ⟨ channel, .pull, mult, msg, h_size ⟩ + refine ⟨ [mk (N : F), mk ((ringChar F - N : ℕ) : F)], fun m => ?_, + by simp only [List.length_cons, List.length_nil]; omega, fun matched => ?_ ⟩ + · -- the two multiplicities sum to the characteristic, so every message balances + by_cases h_msg : msg = m + · have h_bal : balanceOf [mk (N : F), mk ((ringChar F - N : ℕ) : F)] m + = -((N : F) + ((ringChar F - N : ℕ) : F)) := by + simp [balanceOf, mk, Interaction.signedMult, h_msg] + ring + rw [h_bal, h_sum, neg_zero] + · simp [balanceOf, mk, h_msg] + · -- every interaction of the list is a pull counting up to `N`, so none of them is a push + obtain ⟨ b, b_mem, -, b_push ⟩ := + matched (mk ((ringChar F - N : ℕ) : F)) (by simp) ⟨ rfl, h_count_rest ⟩ + obtain ⟨ -, h_not ⟩ := b_push + simp only [List.mem_cons, List.not_mem_nil, or_false] at b_mem + rcases b_mem with rfl | rfl + · exact h_not rfl ⟨ N, by omega, le_rfl, rfl ⟩ + · exact h_not rfl h_count_rest + /-- LogUp: field sums, sound for fewer interactions than the characteristic. -/ def Balance.logUp [ComputableRingChar F] : Balance F where Balanced := LogUpBalanced @@ -403,6 +452,15 @@ lemma plays no part in synthesis. -/ @[simp, circuit_norm] lemma Balance.logUpUpTo_pullBound [ComputableRingChar F] (N : ℕ) : (Balance.logUpUpTo N : Balance F).pullBound = N := rfl +/-- At bound `1`, the counted check is `Balance.logUp` itself, so an ensemble that does not count +keeps the check it already had, down to the proposition its witness has to satisfy. -/ +theorem Balance.logUpUpTo_one [ComputableRingChar F] : + (Balance.logUpUpTo 1 : Balance F) = Balance.logUp := by + unfold Balance.logUpUpTo Balance.logUp + congr 1 + funext n + simp + /- ## Multiset balance with `0`/`1` activations -/ @@ -514,6 +572,18 @@ class ConsistentUpTo (N : ℕ) (channel : RawChannel F) : Prop where instance (N : ℕ) (channel : RawChannel F) [channel.Consistent] : channel.ConsistentUpTo N where consistent interactions data balanced := Consistent.consistent interactions data balanced.toBalancedInteractions + +/-- At bound `1`, consistency under counted pulls is `Consistent`: `ConsistentUpTo` generalizes +that class rather than replacing it. -/ +lemma consistentUpTo_one_iff (channel : RawChannel F) : + channel.ConsistentUpTo 1 ↔ channel.Consistent := by + constructor + · intro h + exact ⟨ fun interactions data balanced reqs => + h.consistent interactions data (balancedUpTo_one_iff.mpr balanced) reqs ⟩ + · intro h + have := h + infer_instance end RawChannel /- diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index ef2751086..ea67231b8 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -45,16 +45,16 @@ from the other channels of a circuit without unfolding them. -/ /-- `counted 1` gives what the plain typed channel gives: only a pull of multiplicity `1` gets the guarantee. -/ -lemma counted_one_guarantees (channel : Channel F Message) (mult : F) +lemma counted_one_guarantees_iff (channel : Channel F Message) (mult : F) (message : Vector F (size Message)) (data : ProverData F) : (channel.counted 1).Guarantees mult message data ↔ channel.toRaw.Guarantees mult message data := by simp [counted, toRaw, countUpTo_one_iff] /-- `counted 1` owes what the plain typed channel owes: an active push owes the guarantee, and a -pull has multiplicity `0` or `1`. Together with `counted_one_guarantees`, counting up to `1` is the +pull has multiplicity `0` or `1`. Together with `counted_one_guarantees_iff`, counting up to `1` is the channel Clean already had. -/ -lemma counted_one_requirements (channel : Channel F Message) (direction : Direction) (mult : F) +lemma counted_one_requirements_iff (channel : Channel F Message) (direction : Direction) (mult : F) (message : Vector F (size Message)) (data : ProverData F) : (channel.counted 1).Requirements direction mult message data ↔ channel.toRaw.Requirements direction mult message data := by From 9d88d778ab356f2ff39214f67f92b8fdca876fc7 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Fri, 13 Nov 2026 11:57:56 -0300 Subject: [PATCH 16/29] Seed the counted squares example, so that it generates. The module docstring said that generation never runs, and gave a reason that was never checked: that a `SoundEnsemble` which is not a VM has no verifier, so nothing creates the first interaction. Neither part held. The consumer was `.preallocated` with no `.multiplicity` cell, which `validateComponent` rejects outright, and an ensemble without a VM can seed itself: a preallocated component's rows exist before there is any demand, and a multiplicity cell starts at `0`, so an emit of `mult + 1` pushes once from every row. The provider is now the preallocated component, with four rows and that emit, and the consumer is demand-generated, one row per pushed square that nothing has pulled. The run reaches `constraintsHold`, `channelsBalanced` and `interactionsFit`. The `+ 1` is the seed and nothing more: it is field arithmetic rather than a count, so it says nothing about a row whose `mult` is `-1`, and soundness is indifferent either way, since pushing is what owes the guarantee. `x` and `y` are witness cells on both sides now. They were fixed columns on the consumer, and a fixed column is a value the verifier knows, so `y = x * x` held of the table whether or not the pull carried it, which is the one thing this example exists to show. `Clean/Examples/CountedSquaresWitnessGenerationTest.lean` pins the run, following the other `WitnessGenerationTest` modules: heights `[2, 4]` and the three checks. `heights` lists the tables in the ensemble's order, which `addTable` builds by prepending, so the consumer added last comes first. --- Clean/Examples/CountedSquares.lean | 53 ++++++++++++------- .../CountedSquaresWitnessGenerationTest.lean | 34 ++++++++++++ Clean/Test.lean | 1 + 3 files changed, 68 insertions(+), 20 deletions(-) create mode 100644 Clean/Examples/CountedSquaresWitnessGenerationTest.lean diff --git a/Clean/Examples/CountedSquares.lean b/Clean/Examples/CountedSquares.lean index a608b00cd..95910f82d 100644 --- a/Clean/Examples/CountedSquares.lean +++ b/Clean/Examples/CountedSquares.lean @@ -19,12 +19,15 @@ to say which. That obligation is what the consumer discharges by constraining it what makes the channel sound: `Channel.counted_not_consistent` shows that a row free to claim any multiplicity would break it. -What this example does not cover: generation never runs. A `SoundEnsemble` that is not a VM has no -verifier in this version — only `addVm` attaches one — so nothing creates the first interaction and -the generator returns an empty witness. The generation modes below are therefore stated so that the -components are well formed, and are not exercised. A row's multiplicity is counted by -`Tally.signedCount`, which reads it as a natural number, so a counted pull is tallied as the number -it claims; that part needs no change for counted channels. +The ensemble also generates. Four provider rows exist before any demand, and a multiplicity cell +starts at `0`, so the `+ 1` of the provider's emit turns each of them into one push. Those four +pushes are the demand the consumer answers, with two rows pulling at multiplicity `2`. The `+ 1` is +there to seed the run and nothing else: a VM seeds its channels from the verifier instead, as in +`Clean/Examples/FemtoCairo/FlatAir.lean`. `Clean/Examples/CountedSquaresWitnessGenerationTest.lean` +runs it and pins the result. + +A row's multiplicity is counted by `Tally.signedCount`, which reads it as a natural number, so a +counted pull is tallied as the number it claims; that part needs no change for counted channels. Constraining the count arithmetically only scales for a small bound. A circuit counting to `510` would range-check its multiplicity on another channel instead, which needs the finishing order of @@ -59,14 +62,18 @@ structure ProviderInput (F : Type) where mult : F deriving ProvableStruct -/-- The squaring chip: constrains `y = x * x` and pushes the pair as often as it is pulled. It owes -the guarantee on every active push, and its own constraint is what pays for it, so the chip assumes -nothing. -/ +/-- The squaring chip: constrains `y = x * x` and pushes the pair, once on top of every pull it +answers. It owes the guarantee on every active push, and its own constraint is what pays for it, so +the chip assumes nothing. + +The `+ 1` is the seed of generation, where a multiplicity cell starts at `0`. It is field +arithmetic and not a count, so it promises nothing about a row in general: a row whose `mult` is +`-1` pushes nothing. Soundness is indifferent to that, since pushing is what owes the guarantee. -/ def provider : GeneralFormalCircuit (F p) ProviderInput unit where name := "squares-provider" main input := do input.y === input.x * input.x - squares.emitUpTo N input.mult (input.x, input.y) + squares.emitUpTo N (input.mult + 1) (input.x, input.y) channelsWithRequirements := [squares.counted N] Spec _ _ _ := True @@ -116,26 +123,32 @@ def consumer : GeneralFormalCircuit (F p) ConsumerInput unit where refine ⟨?_, fun _ => h_square⟩ rcases h_count with h | h <;> simp [h] -/-- One provider row per square that is pulled, with the number of pulls as its multiplicity. -/ +/-- Four provider rows, all about the square `4 = 2 * 2`, which exist before there is any demand: +that is what lets a run start. The multiplicity counts the pulls a row answers, and the circuit's +`+ 1` turns the initial `0` of each row into one push. + +`x` and `y` are witness cells rather than fixed columns on purpose: a fixed column is a value the +verifier knows, and the verifier knowing `y` would make `y = x * x` true of the table without the +push carrying it. -/ abbrev providerComponent : Component (F p) where circuit := provider - generation := .demand { + generation := .preallocated { trigger := .pull channel := squares - input | (x, y) => pure { x, y, mult := .multiplicity } + height := 4 + input _ := pure { x := .witness 2, y := .witness 4, mult := .multiplicity } + rowOf _ := 0 } -/-- Four fixed consumer rows, each needing `4 = 2 * 2` twice. These are the rows that start -generation: they pull, and the squaring chip creates rows to answer them. A consumer driven by -demand on the same channel would feed the provider that feeds it. -/ +/-- One consumer row per pushed square that nothing has pulled, each needing it twice. Two rows +answer the provider's four pushes. Their `x` and `y` come from the message, so what makes +`y = x * x` true of a consumer row is the channel and the provider's constraint, nothing local. -/ abbrev consumerComponent : Component (F p) where circuit := consumer - generation := .preallocated { + generation := .demand { channel := squares trigger := .push - height := 4 - input _ := pure { x := .fixed 2, y := .fixed 4, count := .fixed 2 } - rowOf _ := 0 + input | (x, y) => pure { x, y, count := .witness 2 } } /-- The ensemble: the counted channel, the chip that serves it, the channel finished, and the chip diff --git a/Clean/Examples/CountedSquaresWitnessGenerationTest.lean b/Clean/Examples/CountedSquaresWitnessGenerationTest.lean new file mode 100644 index 000000000..c4f37ce0e --- /dev/null +++ b/Clean/Examples/CountedSquaresWitnessGenerationTest.lean @@ -0,0 +1,34 @@ +module + +public import Clean.Examples.CountedSquares + +-- `native_decide` runs compiled code from this module. +public meta import Clean.Examples.CountedSquares + +@[expose] public section + +namespace CountedSquares.WitnessGenerationTest + +open Air.Flat Air.Flat.WitnessGeneration + +/-- A field small enough to run in, and large enough for the bound: the run makes `6` interactions +on the channel, and `Balance.logUpUpTo 2` asks for `6 * 2 < 101`. -/ +abbrev p : ℕ := 101 + +instance : Fact (Nat.Prime p) := ⟨by decide⟩ + +/-- The heights of the generated tables, and whether the witness satisfies the constraints, the +balance check and its interaction bound. -/ +def result : Except String (List ℕ × Bool × Bool × Bool) := + match generate (soundEnsemble (p := p)).ensemble 100 () () with + | .error error => .error (toString error) + | .ok witness => + .ok (witness.heights, constraintsHold witness, channelsBalanced witness, + (soundEnsemble (p := p)).ensemble.interactionsFit (Balance.logUpUpTo N) witness.heights) + +/-- The four provider rows seed the channel with one push each, and two consumer rows answer them +at multiplicity `2`. `heights` lists the tables in the ensemble's order, which `addTable` builds by +prepending, so the consumer added last comes first. -/ +example : result = .ok ([2, 4], true, true, true) := by native_decide + +end CountedSquares.WitnessGenerationTest diff --git a/Clean/Test.lean b/Clean/Test.lean index cf6f7c6bd..3a062303f 100644 --- a/Clean/Test.lean +++ b/Clean/Test.lean @@ -16,3 +16,4 @@ public import Clean.Examples.FibonacciVm.WitnessGenerationTest public import Clean.Examples.FemtoCairo.FlatAirTest public import Clean.Examples.OpcodeVm.WitnessGenerationTest public import Clean.Examples.BinaryFibonacci.WitnessGenerationTest +public import Clean.Examples.CountedSquaresWitnessGenerationTest From 42c0991db575b039e0ee10788f69b8b0a282a179 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Fri, 13 Nov 2026 12:21:22 -0300 Subject: [PATCH 17/29] Say what CountUpTo claims about the field, which is nothing. `CountUpTo N mult` says that `mult` is named by a natural between `1` and `N`, and carries no hypothesis relating `N` to the field. In a small enough field two such naturals name the same element, so `CountUpTo 2 (0 : ZMod 2)` holds, and a reader meeting that case has no way to tell whether it was overlooked. It was not, and the docstring now says why the bound is not stated here. Carrying `N < ringChar F` in the definition would thread a condition on the field through every requirement, guarantee and proof that mentions a count, to exclude a case the balance check excludes already: `Balance.logUpUpTo N` fits an interaction count only when `count * N < ringChar F`, which over `ZMod 2` with `N = 2` leaves `count = 0`. --- Clean/Air/Balance.lean | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 225e6d379..c7d0a3095 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -21,7 +21,15 @@ with a channel, together with the proof that the check implies these facts. -/ /-- A multiplicity of `k` for some `1 ≤ k ≤ N`: what a table that answers `k` identical lookups -with one row pulls the facts it relies on with. `CountUpTo 1` is multiplicity `1`. -/ +with one row pulls the facts it relies on with. `CountUpTo 1` is multiplicity `1`. + +What this says is that the multiplicity is named by such a `k`, and nothing about the field it +lives in. It does not ask for `N < ringChar F`, so in a small enough field two of those `k` name +the same element, and `CountUpTo 2 (0 : ZMod 2)` holds because `(2 : ZMod 2) = 0`. Asking for the +bound here would carry a condition on the field through every requirement, guarantee and proof that +mentions a count, to exclude a case the balance check excludes already: `Balance.logUpUpTo N` fits +an interaction count only when `count * N < ringChar F`, which over `ZMod 2` with `N = 2` leaves +`count = 0`, so a witness holding such a pull fails the check independently of its multiplicity. -/ def CountUpTo (N : ℕ) (mult : F) : Prop := ∃ k : ℕ, 1 ≤ k ∧ k ≤ N ∧ mult = k lemma countUpTo_one_iff {mult : F} : CountUpTo 1 mult ↔ mult = 1 := by From 7ea114967fb574d04fe8ca913115261eca9ce544 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Mon, 16 Nov 2026 12:22:39 -0300 Subject: [PATCH 18/29] Say what each counted definition claims, and what it does not. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Five docstrings, no change to any proof. `exists_logUpBalanced_not_matchedUpTo` opened by calling the bound of `matchedUpTo` necessary and described the general case of many interactions, while its hypothesis `ringChar F ≤ 2 * N` confines it to fields no larger than `2 * N` and the list it builds has length `2`. It now states what it proves, and a closing paragraph says what it leaves out: `L` multiplicities in `[1, N]` reach every sum between `L` and `L * N`, so any field with `L < ringChar F ≤ L * N` admits the same counterexample on `L` interactions. The inequality is strict because at `L = ringChar F` the list no longer passes a bound on its length alone. That version is not proved here. `countUpTo_of_val` had no stated purpose. It is the step from what a chip knows about its multiplicity, `mult ≠ 0` and `mult.val ≤ N`, to the requirement a counted pull owes, and it is stated over `ZMod p` because that is where `mult.val` exists. `Channel.counted` now documents the invariant behind the shared name: witness generation tallies interactions by channel name and the generation modes take the plain `Channel`, so a counted channel under a new name would stop generating rows, and an `Ensemble` admits at most one channel of a given name. `counted_consistentUpTo_of_le` says why it is a theorem rather than an instance, and what to do instead: supply it with `haveI` before `SoundEnsemble.markFinished`, whose consistency argument is an instance argument that leaves no place for a `simp` step. `CountedSquares` pointed at `Ensemble.tableSoundness_of_stagedOn`, which this branch does not define. The sentence now describes the finishing order it means without naming it. --- Clean/Air/Balance.lean | 23 +++++++++++++++++------ Clean/Air/CountedChannel.lean | 19 +++++++++++++++++-- Clean/Examples/CountedSquares.lean | 4 ++-- 3 files changed, 36 insertions(+), 10 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index c7d0a3095..586c0f626 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -41,7 +41,12 @@ lemma countUpTo_one_iff {mult : F} : CountUpTo 1 mult ↔ mult = 1 := by exact ⟨1, le_rfl, le_rfl, by simp [h]⟩ omit [FiniteField F] in -/-- A nonzero multiplicity whose value is at most `N`. -/ +/-- A nonzero multiplicity whose value is at most `N`. + +This is how a circuit discharges the requirement of a counted pull. The pull owes +`mult = 0 ∨ CountUpTo N mult`, and a chip that range-checks its multiplicity knows exactly +`mult ≠ 0` and `mult.val ≤ N`. Stating it over `ZMod p` rather than a general field is what makes +`mult.val` available. -/ lemma countUpTo_of_val {p : ℕ} [Fact p.Prime] {N : ℕ} {mult : ZMod p} (h0 : mult ≠ 0) (hN : mult.val ≤ N) : CountUpTo N mult := ⟨ mult.val, Nat.one_le_iff_ne_zero.mpr ((ZMod.val_ne_zero mult).mpr h0), hN, @@ -371,14 +376,20 @@ theorem LogUpBalanced.matchedUpTo {N : ℕ} {interactions : List (Interaction F) omega /-- -The bound of `matchedUpTo` is necessary, not an artifact of its proof: once the number of -interactions times `N` reaches the characteristic, the multiplicities of the pulls of a message can -add up to a multiple of it, and the field sums see a balanced list in which nothing was pushed. +The bound of `matchedUpTo` is not an artifact of its proof: a bound on the number of interactions +alone, which is all `Balance.logUp` asks, accepts lists in which a pull of multiplicity up to `N` +is matched by no push. Two pulls, of multiplicities `N` and `ringChar F - N`, are such a list whenever `N < ringChar F` and `ringChar F ≤ 2 * N`: both multiplicities count up to `N`, they sum to the characteristic, and -a list of length `2` passes a bound on the length alone, which is all `Balance.logUp` asks. This is -what `Balance.logUpUpTo N` rules out by asking for `count * N < ringChar F` as well. +a list of length `2` passes a bound on the length alone. This is what `Balance.logUpUpTo N` rules +out by asking for `count * N < ringChar F` as well. + +What this covers, and what it does not: `ringChar F ≤ 2 * N` confines the statement to fields no +larger than `2 * N`, so on its own it says nothing about a field of practical size. The shape does +generalize, since `L` multiplicities in `[1, N]` reach every sum between `L` and `L * N`, so any +field with `L ≤ ringChar F ≤ L * N` admits the same counterexample on `L` interactions. That +version is not proved here. -/ theorem exists_logUpBalanced_not_matchedUpTo {N : ℕ} (hN : 2 ≤ N) (h_lt : N < ringChar F) (h_le : ringChar F ≤ 2 * N) (channel : RawChannel F) (msg : Array F) diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index ea67231b8..d52a68f20 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -20,7 +20,15 @@ variable {Message : TypeMap} [ProvableType Message] namespace Channel -/-- `channel`, with pulls of multiplicity up to `N`. -/ +/-- `channel`, with pulls of multiplicity up to `N`. + +The name is kept, so `channel.toRaw`, `channel.counted 2` and `channel.counted 3` are distinct +`RawChannel`s that share one name. That is deliberate: witness generation tallies interactions by +channel name, and `DemandMode` and `PreallocatedMode` take the plain `Channel`, so a counted +channel under a new name would silently stop generating rows. An `Ensemble` identifies channels by +name and asks those names to be distinct, so at most one of them belongs to a given ensemble: a +channel is emitted on either counted or not, not both. Adding two of them leaves a `Nodup` goal +that no longer closes. -/ def counted (N : ℕ) (channel : Channel F Message) : RawChannel F where name := channel.name arity := size Message @@ -173,7 +181,14 @@ example [ComputableRingChar F] (N : ℕ) (channel : Channel F Message) : (channel.counted N).ConsistentUpTo (Balance.logUpUpTo (F := F) N).pullBound := inferInstance -/-- The counted channel is also consistent for balance checks with a larger bound. -/ +/-- The counted channel is also consistent for balance checks with a larger bound. + +A theorem and not an instance. The instance above fixes the channel's own `N` on both sides, so +synthesis closes `ConsistentUpTo N` for a channel counted at `N` and nothing else. An ensemble +that finishes a channel counted at `N` under a check whose `pullBound` is a larger `M` has to +supply this theorem with `haveI` before calling `SoundEnsemble.markFinished`, which takes the +consistency as an instance argument and leaves no place for a `simp` step. As an instance this +would send synthesis looking for an `N ≤ M` on every goal of the shape. -/ theorem counted_consistentUpTo_of_le {M N : ℕ} (h : N ≤ M) (channel : Channel F Message) : (channel.counted N).ConsistentUpTo M where consistent interactions data balanced := diff --git a/Clean/Examples/CountedSquares.lean b/Clean/Examples/CountedSquares.lean index 95910f82d..0b31d90bd 100644 --- a/Clean/Examples/CountedSquares.lean +++ b/Clean/Examples/CountedSquares.lean @@ -30,8 +30,8 @@ A row's multiplicity is counted by `Tally.signedCount`, which reads it as a natu counted pull is tallied as the number it claims; that part needs no change for counted channels. Constraining the count arithmetically only scales for a small bound. A circuit counting to `510` -would range-check its multiplicity on another channel instead, which needs the finishing order of -`Ensemble.tableSoundness_of_stagedOn` rather than the single finished channel here. +would range-check its multiplicity on another channel instead, which needs a finishing order over +several channels rather than the single finished channel here. -/ namespace CountedSquares From 1d7cd58efea936cb8c75c61374775c45795df546 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Mon, 16 Nov 2026 12:24:39 -0300 Subject: [PATCH 19/29] Let Balance.logUp use the default proof of matchedUpTo_of_balanced. `Balance.logUp` leaves `pullBound` at `1` and still wrote `matchedUpTo_of_balanced` out through `LogUpBalanced.matchedUpTo`, where the field's default proof `BalancedInteractions.matchedUpTo_one` already closes it. `Balance.multiset` beside it takes that default, so the two checks disagreed for no reason, and the claim that an existing check compiles untouched reads better when the main one is such a check. Nothing observable changes: the field is a `Prop`, so no user of `Balance.logUp` can depend on which proof stands there. `LogUpBalanced.matchedUpTo` keeps the caller it was written for, the bound of `Balance.logUpUpTo`. --- Clean/Air/Balance.lean | 3 --- 1 file changed, 3 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 586c0f626..d6c384324 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -431,9 +431,6 @@ def Balance.logUp [ComputableRingChar F] : Balance F where have length_lt := ComputableRingChar.char_eq (R := F) ▸ of_decide_eq_true fits ⟨ balanced.exists_push_of_pull length_lt, balanced.activeCount_push_eq_activeCount_pull_of_isBool length_lt ⟩ - matchedUpTo_of_balanced _ fits balanced _ := - have length_lt := ComputableRingChar.char_eq (R := F) ▸ of_decide_eq_true fits - balanced.matchedUpTo (by simpa using length_lt) /-- `Balance.logUp` matches pulls of multiplicity `1` only. Stated so that `simp`/`circuit_norm` can reduce a `ConsistentUpTo (Balance.logUp).pullBound` goal, which instance search cannot do on From 603c4705bb85ec2629057c553ee08b2c001eaa98 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Tue, 17 Nov 2026 10:51:54 -0300 Subject: [PATCH 20/29] Narrow two counted docstrings to what they actually cover. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Two docstrings, no change to any proof. `exists_logUpBalanced_not_matchedUpTo` said that `ringChar F ≤ 2 * N` confines the statement to fields no larger than `2 * N`. It bounds the characteristic, and a field of characteristic `p` may have `p ^ k` elements. The closing paragraph also gave the general shape as `L ≤ ringChar F ≤ L * N`; at `L = ringChar F` the list has length `ringChar F` and so no longer passes a bound on the length alone, which is what it has to pass to be a counterexample to that bound. The inequality is now strict. `Channel.counted` said that `channel.counted 2` and `channel.counted 3` are distinct `RawChannel`s. Over `ZMod 2` the two bounds describe the same multiplicities and the two values are equal, so they may denote different channels rather than must. The paragraph also read as if at most one of them could belong to an ensemble, which does not hold when they are equal; what `unique_channel_names` rules out is a second entry under a name that is already taken. --- Clean/Air/Balance.lean | 9 +++++---- Clean/Air/CountedChannel.lean | 13 ++++++------- 2 files changed, 11 insertions(+), 11 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index d6c384324..e860dcac6 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -385,10 +385,11 @@ and `ringChar F ≤ 2 * N`: both multiplicities count up to `N`, they sum to the a list of length `2` passes a bound on the length alone. This is what `Balance.logUpUpTo N` rules out by asking for `count * N < ringChar F` as well. -What this covers, and what it does not: `ringChar F ≤ 2 * N` confines the statement to fields no -larger than `2 * N`, so on its own it says nothing about a field of practical size. The shape does -generalize, since `L` multiplicities in `[1, N]` reach every sum between `L` and `L * N`, so any -field with `L ≤ ringChar F ≤ L * N` admits the same counterexample on `L` interactions. That +What this covers, and what it does not: `ringChar F ≤ 2 * N` bounds the characteristic, not the +number of elements, and bounds it in terms of `N`. The shape does generalize, since `L` +multiplicities in `[1, N]` reach every sum between `L` and `L * N`, so any field with +`L < ringChar F ≤ L * N` admits the same counterexample on `L` interactions; the inequality is +strict because at `L = ringChar F` the list no longer passes a bound on the length alone. That version is not proved here. -/ theorem exists_logUpBalanced_not_matchedUpTo {N : ℕ} (hN : 2 ≤ N) (h_lt : N < ringChar F) diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index d52a68f20..e6e9a4005 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -22,13 +22,12 @@ namespace Channel /-- `channel`, with pulls of multiplicity up to `N`. -The name is kept, so `channel.toRaw`, `channel.counted 2` and `channel.counted 3` are distinct -`RawChannel`s that share one name. That is deliberate: witness generation tallies interactions by -channel name, and `DemandMode` and `PreallocatedMode` take the plain `Channel`, so a counted -channel under a new name would silently stop generating rows. An `Ensemble` identifies channels by -name and asks those names to be distinct, so at most one of them belongs to a given ensemble: a -channel is emitted on either counted or not, not both. Adding two of them leaves a `Nodup` goal -that no longer closes. -/ +The name is kept, so `channel.toRaw`, `channel.counted 2` and `channel.counted 3` may denote +different `RawChannel`s under one name. That is deliberate: witness generation tallies +interactions by channel name, and `DemandMode` and `PreallocatedMode` take the plain `Channel`, so +a counted channel under a new name would silently stop generating rows. An `Ensemble` asks the +names of its channels to be distinct, so use one raw-channel value per name: a second entry under +a name already taken fails `unique_channel_names`. -/ def counted (N : ℕ) (channel : Channel F Message) : RawChannel F where name := channel.name arity := size Message From ef49d0de1ba4128452115c542b8ce050f8d94af2 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Tue, 17 Nov 2026 10:52:02 -0300 Subject: [PATCH 21/29] Pad the counted consumer, and exercise the padding row. The consumer of the example had no padding row, so a demand table that fell short of its target height would have been padded with the generator's default of `0`. Away from characteristic `2` that does not satisfy the consumer's one constraint, `(count - 1) * (count - 2) = 0`. The component now carries a padding row that pulls, at multiplicity `1`, the one message the provider ever pushes. A padding row that interacts has to be answered, and the test now shows that it is. The existing run never pads, because its consumer table has two rows, so the new scenario builds the state directly: three consumer rows against four provider rows emitting six pushes between them, which balances, and whose consumer `ComponentState.targetHeight` asks for four. After `padAndBalance` the consumer holds four rows, the constraints, the balance and the interaction bound all hold, and the multiplicity of the provider row that `rowOf` maps the message to has risen from `2` to `3`, which is the push that covers the padding row's pull. Both scenarios go through the same `constraintsHold` and `channelsBalanced` as the existing run, by way of the tail of `generateValid`, rather than a restatement of them. The existing run is unchanged and still pins `[2, 4]`. --- Clean/Examples/CountedSquares.lean | 11 ++- .../CountedSquaresWitnessGenerationTest.lean | 71 ++++++++++++++++++- 2 files changed, 79 insertions(+), 3 deletions(-) diff --git a/Clean/Examples/CountedSquares.lean b/Clean/Examples/CountedSquares.lean index 0b31d90bd..09611ab70 100644 --- a/Clean/Examples/CountedSquares.lean +++ b/Clean/Examples/CountedSquares.lean @@ -142,13 +142,22 @@ abbrev providerComponent : Component (F p) where /-- One consumer row per pushed square that nothing has pulled, each needing it twice. Two rows answer the provider's four pushes. Their `x` and `y` come from the message, so what makes -`y = x * x` true of a consumer row is the channel and the provider's constraint, nothing local. -/ +`y = x * x` true of a consumer row is the channel and the provider's constraint, nothing local. + +Rows of this component are meant to be active: away from characteristic `2`, the default padding +of `0` does not satisfy its one constraint, `(count - 1) * (count - 2) = 0`. The padding row given +here pulls at multiplicity `1` the one message the provider ever pushes, and the generator answers +that extra pull by raising a provider multiplicity, as it does for any interacting padding row. The +generation test's own run has a power-of-two height and so never pads; its second scenario drives +the padding directly. A chip whose rows may be idle would take the `count = 0` branch of the pull's +requirement rather than pad with an active row. -/ abbrev consumerComponent : Component (F p) where circuit := consumer generation := .demand { channel := squares trigger := .push input | (x, y) => pure { x, y, count := .witness 2 } + padding := { x := 2, y := 4, count := 1 } } /-- The ensemble: the counted channel, the chip that serves it, the channel finished, and the chip diff --git a/Clean/Examples/CountedSquaresWitnessGenerationTest.lean b/Clean/Examples/CountedSquaresWitnessGenerationTest.lean index c4f37ce0e..d6397ba62 100644 --- a/Clean/Examples/CountedSquaresWitnessGenerationTest.lean +++ b/Clean/Examples/CountedSquaresWitnessGenerationTest.lean @@ -11,8 +11,8 @@ namespace CountedSquares.WitnessGenerationTest open Air.Flat Air.Flat.WitnessGeneration -/-- A field small enough to run in, and large enough for the bound: the run makes `6` interactions -on the channel, and `Balance.logUpUpTo 2` asks for `6 * 2 < 101`. -/ +/-- A field small enough to run in, and large enough for the bound: the runs below make `6` and +`8` interactions on the channel, and `Balance.logUpUpTo 2` asks for `8 * 2 < 101`. -/ abbrev p : ℕ := 101 instance : Fact (Nat.Prime p) := ⟨by decide⟩ @@ -31,4 +31,71 @@ at multiplicity `2`. `heights` lists the tables in the ensemble's order, which ` prepending, so the consumer added last comes first. -/ example : result = .ok ([2, 4], true, true, true) := by native_decide +/-! ## The padding row + +`result` never pads, because its consumer table has two rows. The scenarios below exercise +`ComponentState.pad` and the rebalancing that follows it, from a balanced state whose consumer +holds three rows. +-/ + +/-- Three consumer rows, each pulling the square twice, and four provider rows emitting six pushes +between them: the circuit's `+ 1` on the multiplicities `[2, 0, 0, 0]`. Six pulls against six +pushes balance, and `ComponentState.targetHeight` asks a three-row demand table for four. -/ +def unpadded : Array (ComponentState (F p)) := Id.run do + let data := ProverData.empty (F p) + let hint := ProverHint.empty (F p) + let mut consumer := ComponentState.initial (consumerComponent (p := p)) none + for _ in List.range 3 do + consumer := consumer.pushRow (completeRow consumer #[2, 4, 2] data hint) + let mut provider := ComponentState.initial (providerComponent (p := p)) (some 2) + for i in List.range 4 do + provider := provider.pushRow + (completeRow provider #[2, 4, if i = 0 then 2 else 0] data hint) + return #[consumer, provider] + +/-- The tail of `generateValid`, so that these states meet the same `constraintsHold` and +`channelsBalanced` as `result` rather than a restatement of them. -/ +def witnessOf (states : List (ComponentState (F p))) : + Except GenerationError (EnsembleWitness (soundEnsemble (p := p)).ensemble) := do + let assembled ← assembleTables (soundEnsemble (p := p)).ensemble.tables states + have hunique : (assembled.tables.map fun table => table.component.circuit.name).Nodup := by + simpa [← assembled.map_component, Function.comp_def] + using (soundEnsemble (p := p)).ensemble.unique_names + let publicInput : unit (F p) := () + return { + tables := withDerivedData assembled.tables hunique (toElements publicInput).toArray + publicInput + same_length := by simp [withDerivedData, assembled.same_length] + same_circuits := by simp [withDerivedData, assembled.same_circuits] + same_data := withDerivedData_data assembled.tables hunique _ + } + +/-- The heights and the three checks of `result`, and the multiplicity of the provider row that +answers the message: `providerComponent` maps every message to row `0` with `rowOf _ := 0`, so +that cell is where an extra pull has to show up. -/ +def resultOf (states : Array (ComponentState (F p))) : + Except String (List ℕ × Bool × Bool × Bool × Option (F p)) := + let answering : Option (F p) := + (states.toList[1]?).bind fun provider => + (provider.rows.toList[0]?).bind (·.values.toList[2]?) + match witnessOf states.toList with + | .error error => .error (toString error) + | .ok witness => + .ok (witness.heights, constraintsHold witness, channelsBalanced witness, + (soundEnsemble (p := p)).ensemble.interactionsFit (Balance.logUpUpTo N) witness.heights, + answering) + +/-- Before padding: three consumer rows, balanced against the provider's six pushes. -/ +example : resultOf unpadded = .ok ([3, 4], true, true, true, some 2) := by native_decide + +def padded : Except String (List ℕ × Bool × Bool × Bool × Option (F p)) := + match padAndBalance (ProverData.empty (F p)) (ProverHint.empty (F p)) 100 100 unpadded with + | .error error => .error (toString error) + | .ok states => resultOf states + +/-- After padding: the consumer reaches four rows, and the padding row's pull is answered rather +than left dangling, so the constraints, the balance and the interaction bound all still hold. The +provider's multiplicity rises from `2` to `3`, which is the extra push that covers it. -/ +example : padded = .ok ([4, 4], true, true, true, some 3) := by native_decide + end CountedSquares.WitnessGenerationTest From 3c6d5b3f73357283dc8c2c88c0e1499a5afdc643 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 18 Nov 2026 10:32:23 -0300 Subject: [PATCH 22/29] Follow two of the repo's stated conventions in the new code. `doc/conventions.md` says not to add to the default `[simp]` set, and to put simplification lemmas in `[circuit_norm]` instead. The three `pullBound` projection lemmas carried both attributes; they keep `circuit_norm`, which is what the example's proofs ask for by name, and give up `simp`. Their docstrings mentioned `simp` too. The two `ExplicitCircuits` instances for `pullUpTo` and `emitUpTo` are modelled on the ones for `Channel.emit` and `Channel.pull` in `Clean/Circuit/Explicit.lean`, which spell the named argument `(F:=F)` without spaces, as the same document asks. Also drops a comment in `Balance.multiset` that said what the docstring below it says. --- Clean/Air/Balance.lean | 19 +++++++++---------- Clean/Air/CountedChannel.lean | 4 ++-- 2 files changed, 11 insertions(+), 12 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index e860dcac6..62a361112 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -433,10 +433,10 @@ def Balance.logUp [ComputableRingChar F] : Balance F where ⟨ balanced.exists_push_of_pull length_lt, balanced.activeCount_push_eq_activeCount_pull_of_isBool length_lt ⟩ -/-- `Balance.logUp` matches pulls of multiplicity `1` only. Stated so that `simp`/`circuit_norm` -can reduce a `ConsistentUpTo (Balance.logUp).pullBound` goal, which instance search cannot do on -its own: `Balance.logUp` is a plain `def`, so the projection is not reducible. -/ -@[simp, circuit_norm] lemma Balance.logUp_pullBound [ComputableRingChar F] : +/-- `Balance.logUp` matches pulls of multiplicity `1` only. Stated so that `circuit_norm` can +reduce a `ConsistentUpTo (Balance.logUp).pullBound` goal, which instance search cannot do on its +own: `Balance.logUp` is a plain `def`, so the projection is not reducible. -/ +@[circuit_norm] lemma Balance.logUp_pullBound [ComputableRingChar F] : (Balance.logUp : Balance F).pullBound = 1 := rfl /-- @@ -463,10 +463,10 @@ argument leaves no place to `simp` first. (of_decide_eq_true (Bool.and_eq_true_iff.mp fits).2) balanced.matchedUpTo length_mul_lt -/-- `Balance.logUpUpTo N` matches pulls of multiplicity up to `N`, for `simp` and `circuit_norm`. -Instance search needs the same fact, and gets it from the check being `@[reducible]`: a `simp` -lemma plays no part in synthesis. -/ -@[simp, circuit_norm] lemma Balance.logUpUpTo_pullBound [ComputableRingChar F] (N : ℕ) : +/-- `Balance.logUpUpTo N` matches pulls of multiplicity up to `N`, for `circuit_norm`. Instance +search needs the same fact, and gets it from the check being `@[reducible]`: a simp lemma plays no +part in synthesis. -/ +@[circuit_norm] lemma Balance.logUpUpTo_pullBound [ComputableRingChar F] (N : ℕ) : (Balance.logUpUpTo N : Balance F).pullBound = N := rfl /-- At bound `1`, the counted check is `Balance.logUp` itself, so an ensemble that does not count @@ -507,10 +507,9 @@ def Balance.multiset : Balance F where obtain ⟨ b, b_mem, h_b ⟩ := List.countP_pos_iff.mp pull_pos simp only [decide_eq_true_eq] at h_b exact ⟨ b, b_mem, h_b.2.1, h_b.2.2, by simp [h_b.1] ⟩ - -- `matchedUpTo_of_balanced` is the default: `pullBound` is `1`. /-- `Balance.multiset` matches pulls of multiplicity `1` only. -/ -@[simp, circuit_norm] lemma Balance.multiset_pullBound : +@[circuit_norm] lemma Balance.multiset_pullBound : (Balance.multiset : Balance F).pullBound = 1 := rfl /- ## Conditions on channels that strengthen the implications of interaction balance. -/ diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index e6e9a4005..9ccf47fff 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -90,14 +90,14 @@ def emitUpTo (N : ℕ) (channel : Channel F Message) (mult : Expression F) ((), [.interact (emittedUpTo N channel mult msg)]) instance {N : ℕ} {channel : Channel F Message} {count : Expression F} : - ExplicitCircuits (F := F) (channel.pullUpTo N count) where + ExplicitCircuits (F:=F) (channel.pullUpTo N count) where output _ _ := () localLength _ _ := 0 operations msg _ := [.interact (pulledUpTo N channel count msg)] channelsWithGuarantees _ _ := [channel.counted N] instance {N : ℕ} {channel : Channel F Message} {mult : Expression F} : - ExplicitCircuits (F := F) (channel.emitUpTo N mult) where + ExplicitCircuits (F:=F) (channel.emitUpTo N mult) where output _ _ := () localLength _ _ := 0 operations msg _ := [.interact (emittedUpTo N channel mult msg)] From 04c26b0329bb55e791edf4d42a0f198b551bc367 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 18 Nov 2026 10:32:35 -0300 Subject: [PATCH 23/29] Keep using BalancedChannel where this branch had spelled it out. `EnsembleWitness.BalancedChannel` had two callers, both in this file. Threading the pull bound through them replaced the abbreviation with its body, which left the abbreviation used only by `BalancedChannels` and left `partialBalancedChannel_of_balancedChannel` named after a declaration its statement no longer mentioned. Both sites take the bound as an argument instead. --- Clean/Air/OrderedChannel.lean | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Clean/Air/OrderedChannel.lean b/Clean/Air/OrderedChannel.lean index ec21464ea..fffc18866 100644 --- a/Clean/Air/OrderedChannel.lean +++ b/Clean/Air/OrderedChannel.lean @@ -569,7 +569,7 @@ def SoundChannels (ens : Ensemble F PublicIO) (N : ℕ) (finished : List (RawCha lemma partialBalancedChannel_of_balancedChannel {ens : Ensemble F PublicIO} {witness : EnsembleWitness ens} {channel : RawChannel F} : ens.OrderedChannel channel → - BalancedUpTo N (witness.interactionsWith channel) → + witness.BalancedChannel N channel → PartialBalancedChannel N witness.toTables channel := by intro ordered balanced use witness.verifierInteractionsWith channel @@ -593,7 +593,7 @@ lemma partialBalancedChannel_of_balancedChannel {ens : Ensemble F PublicIO} lemma verifierChannelGuarantees_of_tableRequirements {ens : Ensemble F PublicIO} {witness : EnsembleWitness ens} {channel : RawChannel F} [channel.ConsistentUpTo N] : OrderedChannelRefl channel ens.verifier → - BalancedUpTo N (witness.interactionsWith channel) → + witness.BalancedChannel N channel → (∀ table ∈ witness.tables, table.ChannelRequirements channel) → ens.VerifierChannelGuarantees witness.data channel := by intro ordered balanced table_requirements From 7dba7f89ddc370c5ea32bee3fb4f7448c49f5cba Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 18 Nov 2026 10:32:37 -0300 Subject: [PATCH 24/29] Describe the pull bound and the counted channel in the Air README. The Organization section documents every file in `Clean/Air/` and the fields of a `Balance`, so it goes stale on both counts here: a balance check now also records the pull multiplicity it matches with a push, and `CountedChannel.lean` is a new file. The counted-channel paragraph says what the channel asks of a pull, which balance check can finish it, and that `counted_not_consistent` is why the stronger check is necessary rather than merely convenient. --- Clean/Air/README.md | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/Clean/Air/README.md b/Clean/Air/README.md index 8da75506e..ab5c171ea 100644 --- a/Clean/Air/README.md +++ b/Clean/Air/README.md @@ -21,7 +21,7 @@ In this terminology, a `Flat.Component` is a one-row AIR component: it packages It also proves the basic row-level transport lemmas: instantiated component operations agree with row operations, component soundness lifts to table soundness, and table interactions can be collected per channel. -`Balance.lean` contains the channel multiset theory. An interaction pushes a message to a channel or pulls it, a number of times given by its multiplicity. What balance means depends on the proof system: a `Balance` is the check a verifier performs on the interactions with each channel, like LogUp's field sums in prime fields or multiset equality with `0`/`1` multiplicities in binary fields, together with the numbers of interactions it is sound for (fewer than the characteristic for LogUp) and its proof of `BalancedInteractions` for those, the facts that soundness proofs use: every pull is matched by an interaction that provides its message, and messages with `0`/`1` multiplicities are pushed as often as they are pulled. The file proves permutation and counting lemmas, and provides the channel-level implication principles used by higher-level soundness proofs. It also defines `RawChannel.Consistent` and `RawChannel.Normal`; typed channels are normal by construction, and normal channels are consistent, so both properties are satisfied in practice. A highlight in `Balance.lean` is the "guarantees-to-requirements-reversal" theorem which provides the basis for soundness of VM channels. +`Balance.lean` contains the channel multiset theory. An interaction pushes a message to a channel or pulls it, a number of times given by its multiplicity. What balance means depends on the proof system: a `Balance` is the check a verifier performs on the interactions with each channel, like LogUp's field sums in prime fields or multiset equality with `0`/`1` multiplicities in binary fields, together with the numbers of interactions it is sound for (fewer than the characteristic for LogUp) and its proof of `BalancedInteractions` for those, the facts that soundness proofs use: every pull is matched by an interaction that provides its message, and messages with `0`/`1` multiplicities are pushed as often as they are pulled. The file proves permutation and counting lemmas, and provides the channel-level implication principles used by higher-level soundness proofs. It also defines `RawChannel.Consistent` and `RawChannel.Normal`; typed channels are normal by construction, and normal channels are consistent, so both properties are satisfied in practice. A highlight in `Balance.lean` is the "guarantees-to-requirements-reversal" theorem which provides the basis for soundness of VM channels. A balance check also records how high a pull multiplicity it matches with a push, in its `pullBound`, together with the proof of `MatchedUpTo` for that bound (`matchedUpTo_of_balanced`). A check that matches multiplicity `1` only takes the default proof and is unchanged; `Balance.logUpUpTo N` raises the bound to `N` by additionally asking its `fits` for `count * N < ringChar F`, which `exists_logUpBalanced_not_matchedUpTo` shows a bound on the count alone does not give. `RawChannel.ConsistentUpTo N` generalizes `RawChannel.Consistent`, which is `ConsistentUpTo 1`. `Verifier.lean` defines the verifier program language: an append-only list of public interactions, together with the verifier's spec and its soundness. @@ -29,6 +29,8 @@ It also proves the basic row-level transport lemmas: instantiated component oper `Lookup.lean` defines a convenience wrapper for static lookups. +`CountedChannel.lean` defines `Channel.counted N`, a typed channel whose pulls may carry a multiplicity of up to `N` rather than only `0` or `1`. A row that needs the same fact `N` times pulls it once at multiplicity `N`, costing one interaction instead of `N`. In exchange for the guarantee, the channel asks such a pull to account for its count (`mult = 0 ∨ CountUpTo N mult`), and it is consistent up to `N` instead of up to `1`, so finishing it in a `SoundEnsemble` needs a balance check whose `pullBound` is at least `N`. `Channel.counted_not_consistent` shows that the stronger check is necessary and not merely convenient: under plain `BalancedInteractions`, two pulls that both account for their counts can still leave a guarantee unsupported. `Clean/Examples/CountedSquares` assembles a sound ensemble that uses one. + `FlatEnsemble.lean` defines flat AIR ensembles, `Flat.Ensemble` and their witnesses, `Flat.EnsembleWitness`. An ensemble is everything a prover and a verifier implementation need: components, channels, a verifier program, and the type of the prover's private input. The verifier only contributes public interactions directly. The ensemble lists every channel its components and verifier interact with, under unique names, so a proof system can balance channels by name (`EnsembleWitness.interactionsWith_eq_filter_name`). Its `Statement` is the raw proof-system relation for a balance check: there exists a witness whose table constraints hold and whose interactions with every channel pass the check. Every table of a witness carries the prover data derived from all the witness tables and the public input, so channel guarantees and specs can refer to both. The statement also fixes the table heights. Each row of a component interacts with a channel equally often, so the heights determine the number of interactions with every channel (`Ensemble.interactionCount`, which counts by channel name), and the statement requires the balance check to be sound for those numbers, a computable condition that the verifier checks from the heights (`Ensemble.interactionsFit`). The ensemble file also defines soundness and completeness and the `FormalEnsemble` structure which bundles an ensemble with its `Spec`, `Assumptions`, the `ProverAssumptions` of witness generation, and the soundness proof under a balance check (completeness is TODO), like `FormalCircuit` does for a circuit. **Ensemble-level soundness** is more than a simple lifting of per-circuit soundness: it requires that channel guarantees, which were _assumed_ as part of local circuit proofs, are shown to hold unconditionally from global channel balance and constraints. From e0483ddb35714bfbc5087cd57f4cfcf48754123f Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 18 Nov 2026 10:32:51 -0300 Subject: [PATCH 25/29] Give the counted squares example a directory, and fix three things in it. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Every example in `Clean/Examples/` that has a witness-generation test keeps both files in its own directory: `FibonacciVm`, `OpcodeVm`, `BinaryFibonacci` as `Ensemble.lean` plus `WitnessGenerationTest.lean`, and `FemtoCairo` as `FlatAir.lean` plus `FlatAirTest.lean`. This example was the only pair at the top level, so it moves to `Clean/Examples/CountedSquares/{Ensemble,WitnessGenerationTest}.lean`. The test defined its own prime and proved it prime, where `Clean/Utils/Primes.lean` is where the repo keeps primes and their `Fact` instances, and where `OpcodeVm/WitnessGenerationTest.lean` imports them from. The bound the test needs is `8 * 2 < p`, so the existing `p1009` serves and no new prime is needed. `N` was an `ℕ`, which `REVIEWERS.md` asks to name in `lowerCamelCase`; it becomes `countBound`. `countUpTo_of_one_or_two` described its hypothesis in English, which the same document rules out; the hypothesis is a product equal to zero, so the lemma is `countUpTo_of_mul_eq_zero`. The module docstring said `counted_not_consistent` shows that a row free to claim any multiplicity would break the channel. It shows something else, and something stronger: both of its pulls account for their counts, and plain `BalancedInteractions` still leaves a guarantee unsupported. The soundness therefore comes from the balance check bounding the multiplicity, not from the obligation on the pull alone. --- Clean.lean | 2 +- .../Ensemble.lean} | 35 ++++++++++--------- .../WitnessGenerationTest.lean} | 22 ++++++------ Clean/Test.lean | 2 +- 4 files changed, 32 insertions(+), 29 deletions(-) rename Clean/Examples/{CountedSquares.lean => CountedSquares/Ensemble.lean} (87%) rename Clean/Examples/{CountedSquaresWitnessGenerationTest.lean => CountedSquares/WitnessGenerationTest.lean} (89%) diff --git a/Clean.lean b/Clean.lean index e0aea6554..ae6092d69 100644 --- a/Clean.lean +++ b/Clean.lean @@ -47,7 +47,7 @@ public import Clean.Gadgets.BLAKE3.FinalStateUpdate public import Clean.Gadgets.BLAKE3.Round public import Clean.Gadgets.BLAKE3.FinalizeChunk public import Clean.Examples.FibonacciVm.Ensemble -public import Clean.Examples.CountedSquares +public import Clean.Examples.CountedSquares.Ensemble public import Clean.Examples.OpcodeVm.Ensemble public import Clean.Examples.BinaryFibonacci.Ensemble public import Clean.Gadgets.SHA256.SHA256Compress diff --git a/Clean/Examples/CountedSquares.lean b/Clean/Examples/CountedSquares/Ensemble.lean similarity index 87% rename from Clean/Examples/CountedSquares.lean rename to Clean/Examples/CountedSquares/Ensemble.lean index 09611ab70..b2da5d3b8 100644 --- a/Clean/Examples/CountedSquares.lean +++ b/Clean/Examples/CountedSquares/Ensemble.lean @@ -15,16 +15,17 @@ second costs one interaction instead of two, and is what `Channel.counted` is fo The fact here is `y = x * x`. The squaring chip enforces it with a constraint and pushes the pair; a consumer that needs the same square once or twice pulls it once with that multiplicity. `squares.counted 2` hands its guarantee to a pull of multiplicity `1` or `2`, and asks such a pull -to say which. That obligation is what the consumer discharges by constraining its count, and it is -what makes the channel sound: `Channel.counted_not_consistent` shows that a row free to claim any -multiplicity would break it. +to say which. That obligation is what the consumer discharges by constraining its count. The +obligation by itself is not what makes the channel sound: `Channel.counted_not_consistent` builds +two pulls that both satisfy it and are still unmatched under plain `BalancedInteractions`, which is +why the bound has to come from the balance check, here `Balance.logUpUpTo 2`. The ensemble also generates. Four provider rows exist before any demand, and a multiplicity cell starts at `0`, so the `+ 1` of the provider's emit turns each of them into one push. Those four pushes are the demand the consumer answers, with two rows pulling at multiplicity `2`. The `+ 1` is there to seed the run and nothing else: a VM seeds its channels from the verifier instead, as in -`Clean/Examples/FemtoCairo/FlatAir.lean`. `Clean/Examples/CountedSquaresWitnessGenerationTest.lean` -runs it and pins the result. +`Clean/Examples/FemtoCairo/FlatAir.lean`. +`Clean/Examples/CountedSquares/WitnessGenerationTest.lean` runs it and pins the result. A row's multiplicity is counted by `Tally.signedCount`, which reads it as a natural number, so a counted pull is tallied as the number it claims; that part needs no change for counted channels. @@ -41,7 +42,7 @@ open Air.Flat variable {p : ℕ} [Fact p.Prime] /-- Pulls may count up to `2`. -/ -abbrev N : ℕ := 2 +abbrev countBound : ℕ := 2 /-- Pairs `(x, y)` with `y = x * x`. -/ def squares : Channel (F p) (ProvablePair field field) where @@ -50,7 +51,7 @@ def squares : Channel (F p) (ProvablePair field field) where /-- A count of `1` or `2` is a count up to `2`: the step from a field constraint to the channel's `CountUpTo`, and the only arithmetic the example needs. -/ -lemma countUpTo_of_one_or_two {c : F p} (h : (c - 1) * (c - 2) = 0) : CountUpTo N c := by +lemma countUpTo_of_mul_eq_zero {c : F p} (h : (c - 1) * (c - 2) = 0) : CountUpTo countBound c := by rcases mul_eq_zero.mp h with h | h <;> rw [sub_eq_zero] at h · exact ⟨1, le_rfl, by norm_num, by simp [h]⟩ · exact ⟨2, by norm_num, le_rfl, by simp [h]⟩ @@ -73,9 +74,9 @@ def provider : GeneralFormalCircuit (F p) ProviderInput unit where name := "squares-provider" main input := do input.y === input.x * input.x - squares.emitUpTo N (input.mult + 1) (input.x, input.y) + squares.emitUpTo countBound (input.mult + 1) (input.x, input.y) - channelsWithRequirements := [squares.counted N] + channelsWithRequirements := [squares.counted countBound] Spec _ _ _ := True ProverAssumptions input _ _ := input.y = input.x * input.x @@ -101,7 +102,7 @@ def consumer : GeneralFormalCircuit (F p) ConsumerInput unit where name := "squares-consumer" main input := do assertZero ((input.count - 1) * (input.count - 2)) - squares.pullUpTo N input.count (input.x, input.y) + squares.pullUpTo countBound input.count (input.x, input.y) Spec input _ _ := input.y = input.x * input.x ProverAssumptions input _ _ := @@ -110,12 +111,12 @@ def consumer : GeneralFormalCircuit (F p) ConsumerInput unit where requirementsChannelsLawful := by simp only [circuit_norm, seval] intro input env h - exact .inr (countUpTo_of_one_or_two h) + exact .inr (countUpTo_of_mul_eq_zero h) soundness := by circuit_proof_start [squares] obtain ⟨h_count, h_guarantee⟩ := h_holds - have h_in := countUpTo_of_one_or_two h_count + have h_in := countUpTo_of_mul_eq_zero h_count exact ⟨h_guarantee h_in, .inr h_in⟩ completeness := by circuit_proof_start [squares] @@ -163,14 +164,14 @@ abbrev consumerComponent : Component (F p) where /-- The ensemble: the counted channel, the chip that serves it, the channel finished, and the chip that reads it. `markFinished` is where the counted channel has to be `ConsistentUpTo` the balance check's `pullBound`, which is why `Balance.logUpUpTo` must be reducible. -/ -def soundEnsemble : SoundEnsemble (F p) unit (Balance.logUpUpTo N) := - SoundEnsemble.empty (F p) unit (Balance.logUpUpTo N) - |>.addChannel (squares.counted N) +def soundEnsemble : SoundEnsemble (F p) unit (Balance.logUpUpTo countBound) := + SoundEnsemble.empty (F p) unit (Balance.logUpUpTo countBound) + |>.addChannel (squares.counted countBound) |>.addTable providerComponent (by simp [circuit_norm, provider]) (by simp [circuit_norm, provider]) (by simp [circuit_norm, provider]) - |>.markFinished (squares.counted N) + |>.markFinished (squares.counted countBound) |>.addTable consumerComponent (by simp [circuit_norm, consumer]) (by simp [circuit_norm, consumer]) @@ -184,7 +185,7 @@ The balance check's `fits` is where the bound of `LogUpBalanced.matchedUpTo` is satisfies `Ensemble.Statement` when the interaction count times `2` stays below the characteristic. -/ theorem tableSoundness : - (soundEnsemble (p := p)).ensemble.TableSoundness (Balance.logUpUpTo N) := + (soundEnsemble (p := p)).ensemble.TableSoundness (Balance.logUpUpTo countBound) := Ensemble.tableSoundness_of_soundChannels ⟨_, (soundEnsemble (p := p)).finished_subset, (soundEnsemble (p := p)).soundChannels⟩ diff --git a/Clean/Examples/CountedSquaresWitnessGenerationTest.lean b/Clean/Examples/CountedSquares/WitnessGenerationTest.lean similarity index 89% rename from Clean/Examples/CountedSquaresWitnessGenerationTest.lean rename to Clean/Examples/CountedSquares/WitnessGenerationTest.lean index d6397ba62..93765e627 100644 --- a/Clean/Examples/CountedSquaresWitnessGenerationTest.lean +++ b/Clean/Examples/CountedSquares/WitnessGenerationTest.lean @@ -1,9 +1,11 @@ module -public import Clean.Examples.CountedSquares +public import Clean.Examples.CountedSquares.Ensemble +public import Clean.Utils.Primes --- `native_decide` runs compiled code from this module. -public meta import Clean.Examples.CountedSquares +-- `native_decide` runs compiled code from these modules. +public meta import Clean.Examples.CountedSquares.Ensemble +public meta import Clean.Utils.Primes @[expose] public section @@ -11,11 +13,9 @@ namespace CountedSquares.WitnessGenerationTest open Air.Flat Air.Flat.WitnessGeneration -/-- A field small enough to run in, and large enough for the bound: the runs below make `6` and -`8` interactions on the channel, and `Balance.logUpUpTo 2` asks for `8 * 2 < 101`. -/ -abbrev p : ℕ := 101 - -instance : Fact (Nat.Prime p) := ⟨by decide⟩ +/-- Small enough to run in, and large enough for the bound: the runs below make `6` and `8` +interactions on the channel, and `Balance.logUpUpTo 2` asks for `8 * 2 < p1009`. -/ +abbrev p : ℕ := p1009 /-- The heights of the generated tables, and whether the witness satisfies the constraints, the balance check and its interaction bound. -/ @@ -24,7 +24,8 @@ def result : Except String (List ℕ × Bool × Bool × Bool) := | .error error => .error (toString error) | .ok witness => .ok (witness.heights, constraintsHold witness, channelsBalanced witness, - (soundEnsemble (p := p)).ensemble.interactionsFit (Balance.logUpUpTo N) witness.heights) + (soundEnsemble (p := p)).ensemble.interactionsFit (Balance.logUpUpTo countBound) + witness.heights) /-- The four provider rows seed the channel with one push each, and two consumer rows answer them at multiplicity `2`. `heights` lists the tables in the ensemble's order, which `addTable` builds by @@ -82,7 +83,8 @@ def resultOf (states : Array (ComponentState (F p))) : | .error error => .error (toString error) | .ok witness => .ok (witness.heights, constraintsHold witness, channelsBalanced witness, - (soundEnsemble (p := p)).ensemble.interactionsFit (Balance.logUpUpTo N) witness.heights, + (soundEnsemble (p := p)).ensemble.interactionsFit (Balance.logUpUpTo countBound) + witness.heights, answering) /-- Before padding: three consumer rows, balanced against the provider's six pushes. -/ diff --git a/Clean/Test.lean b/Clean/Test.lean index 3a062303f..1e59a56ee 100644 --- a/Clean/Test.lean +++ b/Clean/Test.lean @@ -16,4 +16,4 @@ public import Clean.Examples.FibonacciVm.WitnessGenerationTest public import Clean.Examples.FemtoCairo.FlatAirTest public import Clean.Examples.OpcodeVm.WitnessGenerationTest public import Clean.Examples.BinaryFibonacci.WitnessGenerationTest -public import Clean.Examples.CountedSquaresWitnessGenerationTest +public import Clean.Examples.CountedSquares.WitnessGenerationTest From 3e2cb4975bc88f0ced930571aef031859a9d3d62 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Wed, 18 Nov 2026 15:16:56 -0300 Subject: [PATCH 26/29] Say why the padding scenarios run on empty prover data. `unpadded` builds its component states with `ProverData.empty` and `ProverHint.empty`, which is right here but not self-evident: a reader has to check both chips to see that nothing reads the data. Neither circuit, spec or set of assumptions mentions it, and the channel's guarantee is `y = x * x`, on the message alone. --- Clean/Examples/CountedSquares/WitnessGenerationTest.lean | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/Clean/Examples/CountedSquares/WitnessGenerationTest.lean b/Clean/Examples/CountedSquares/WitnessGenerationTest.lean index 93765e627..0b0470332 100644 --- a/Clean/Examples/CountedSquares/WitnessGenerationTest.lean +++ b/Clean/Examples/CountedSquares/WitnessGenerationTest.lean @@ -43,6 +43,10 @@ holds three rows. between them: the circuit's `+ 1` on the multiplicities `[2, 0, 0, 0]`. Six pulls against six pushes balance, and `ComponentState.targetHeight` asks a three-row demand table for four. -/ def unpadded : Array (ComponentState (F p)) := Id.run do + -- The prover data and hint are empty because nothing here reads them: neither chip's circuit, + -- spec or assumptions mentions the data, and the channel's guarantee is `y = x * x`, on the + -- message alone. `completeRow` takes them to evaluate witness programs, and the two cells these + -- rows witness are a literal and a multiplicity. let data := ProverData.empty (F p) let hint := ProverHint.empty (F p) let mut consumer := ComponentState.initial (consumerComponent (p := p)) none From 6b7eafdb31a02eb4f486c57132151cb9d7b62639 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 19 Nov 2026 11:16:47 -0300 Subject: [PATCH 27/29] Name the consumer this was written for. Several declarations here have no caller in this repo, and `countUpTo_of_val` has none at all: the example counts to `2`, where a constraint is cheaper than a range check, so it proves its own step instead. The chips that need the general form are in lambda_vm, a RISC-V zkVM built on Clean, where a multiplicity is a byte. Saying so is the answer to "where is this needed?", which `Clean/Circuit/ExpressionPolynomial` gives in the same way for leanerVM. --- Clean/Air/Balance.lean | 5 ++++- Clean/Air/CountedChannel.lean | 3 +++ 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 62a361112..7606b453e 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -46,7 +46,10 @@ omit [FiniteField F] in This is how a circuit discharges the requirement of a counted pull. The pull owes `mult = 0 ∨ CountUpTo N mult`, and a chip that range-checks its multiplicity knows exactly `mult ≠ 0` and `mult.val ≤ N`. Stating it over `ZMod p` rather than a general field is what makes -`mult.val` available. -/ +`mult.val` available. + +No example in this repo range-checks a multiplicity, because the bound there is `2` and a +constraint is cheaper. The chips that do are in lambda_vm, where the bound is a byte. -/ lemma countUpTo_of_val {p : ℕ} [Fact p.Prime] {N : ℕ} {mult : ZMod p} (h0 : mult ≠ 0) (hN : mult.val ≤ N) : CountUpTo N mult := ⟨ mult.val, Nat.one_le_iff_ne_zero.mpr ((ZMod.val_ne_zero mult).mpr h0), hN, diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 9ccf47fff..557ee49b5 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -13,6 +13,9 @@ multiplicity `k`. `Channel.counted N channel` is `channel` with pulls of multipl a pull gets the guarantee, and an active push owes it. It is consistent in the sense of `RawChannel.ConsistentUpTo N`, so it can be finished in a `SoundEnsemble` whose balance check has a `pullBound` of at least `N`, e.g. `Balance.logUpUpTo N`. + +This was written for lambda_vm, a RISC-V zkVM built on Clean, whose arithmetic chips answer several +identical lookups per row. `Clean/Examples/CountedSquares` is the same shape in miniature. -/ variable {F : Type} [FiniteField F] From 3b5b146d59589b1eee5ea088c15ab49bc904adc0 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 19 Nov 2026 12:30:57 -0300 Subject: [PATCH 28/29] Take the review fixes from the parallel round, and correct three of my own. A second pass over the same review landed in `pr1-review-fixes`. Four of its fixes are right and were missing here: - `partialBalancedChannel_of_balancedInteractions` was named for a hypothesis this branch had already generalized; it is `partialBalancedChannel_of_balancedUpTo`. - `countUpTo_of_val` named one of its two hypotheses, and named it loosely. `REVIEWERS.md` asks for the hypotheses in order, so it is `countUpTo_of_ne_zero_of_val_le`. - `RawChannel.eq_false_of_name_ne` is about `RawChannel`, so it belongs before `namespace Channel` rather than inside it behind a `_root_` escape. - `Clean/Air/README.md` said balance gives "every pull is matched by an interaction that provides its message". With counted pulls that is false: `BalancedInteractions.exists_push_of_pull` matches a pull of multiplicity `1`. Both renames are breaking for a downstream consumer, which is why they are better taken now than later. And three mistakes of mine, two of them introduced by my own earlier fixes: - The generation test said its runs make `6` and `8` interactions. There are three runs, at heights `[2, 4]`, `[3, 4]` and `[4, 4]`, so `6`, `7` and `8`. - The README called `Channel.counted N` a typed channel. It returns a `RawChannel`. - One `(F := F)` survived in the same file where the others were tightened. --- Clean/Air/Balance.lean | 2 +- Clean/Air/CountedChannel.lean | 16 ++++++++-------- Clean/Air/OrderedChannel.lean | 2 +- Clean/Air/README.md | 4 ++-- .../CountedSquares/WitnessGenerationTest.lean | 2 +- 5 files changed, 13 insertions(+), 13 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index 7606b453e..b9bdd48e1 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -50,7 +50,7 @@ This is how a circuit discharges the requirement of a counted pull. The pull owe No example in this repo range-checks a multiplicity, because the bound there is `2` and a constraint is cheaper. The chips that do are in lambda_vm, where the bound is a byte. -/ -lemma countUpTo_of_val {p : ℕ} [Fact p.Prime] {N : ℕ} {mult : ZMod p} (h0 : mult ≠ 0) +lemma countUpTo_of_ne_zero_of_val_le {p : ℕ} [Fact p.Prime] {N : ℕ} {mult : ZMod p} (h0 : mult ≠ 0) (hN : mult.val ≤ N) : CountUpTo N mult := ⟨ mult.val, Nat.one_le_iff_ne_zero.mpr ((ZMod.val_ne_zero mult).mpr h0), hN, (ZMod.natCast_zmod_val mult).symm ⟩ diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 557ee49b5..1fe60a7b7 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -21,6 +21,13 @@ identical lookups per row. `Clean/Examples/CountedSquares` is the same shape in variable {F : Type} [FiniteField F] variable {Message : TypeMap} [ProvableType Message] +omit [FiniteField F] in +/-- Channels with different names are different. Lets `circuit_norm` tell a counted channel apart +from the other channels of a circuit without unfolding them. -/ +@[circuit_norm] lemma RawChannel.eq_false_of_name_ne {a b : RawChannel F} + (h : a.name ≠ b.name) : (a = b) = False := + eq_false fun h_eq => h (congrArg RawChannel.name h_eq) + namespace Channel /-- `channel`, with pulls of multiplicity up to `N`. @@ -41,13 +48,6 @@ def counted (N : ℕ) (channel : Channel F Message) : RawChannel F where | .push => mult ≠ 0 → channel.Guarantees (fromElements message) data | .pull => mult = 0 ∨ CountUpTo N mult -omit [FiniteField F] in -/-- Channels with different names are different. Lets `circuit_norm` tell a counted channel apart -from the other channels of a circuit without unfolding them. -/ -@[circuit_norm] lemma _root_.RawChannel.eq_false_of_name_ne {a b : RawChannel F} - (h : a.name ≠ b.name) : (a = b) = False := - eq_false fun h_eq => h (congrArg RawChannel.name h_eq) - @[circuit_norm] lemma counted_name (N : ℕ) (channel : Channel F Message) : (channel.counted N).name = channel.name := rfl @[circuit_norm] lemma counted_arity (N : ℕ) (channel : Channel F Message) : @@ -180,7 +180,7 @@ instance (N : ℕ) (channel : Channel F Message) : (channel.counted N).Consisten what `SoundEnsemble.markFinished` needs, since it takes the instance as an instance argument, where no `simp` step can run first. -/ example [ComputableRingChar F] (N : ℕ) (channel : Channel F Message) : - (channel.counted N).ConsistentUpTo (Balance.logUpUpTo (F := F) N).pullBound := + (channel.counted N).ConsistentUpTo (Balance.logUpUpTo (F:=F) N).pullBound := inferInstance /-- The counted channel is also consistent for balance checks with a larger bound. diff --git a/Clean/Air/OrderedChannel.lean b/Clean/Air/OrderedChannel.lean index fffc18866..36f5fe02d 100644 --- a/Clean/Air/OrderedChannel.lean +++ b/Clean/Air/OrderedChannel.lean @@ -210,7 +210,7 @@ def PartialBalancedChannel (N : ℕ) (tables : Tables F) (channel : RawChannel F (channel ∉ tables.tables.flatMap (·.channelsWithGuarantees) ∨ ∀ i ∈ extraInteractions, i.Requirements tables.data) /-- Partial balance is trivially weaker than balance -/ -lemma partialBalancedChannel_of_balancedInteractions {tables : Tables F} {channel : RawChannel F} : +lemma partialBalancedChannel_of_balancedUpTo {tables : Tables F} {channel : RawChannel F} : BalancedUpTo N (tables.interactionsWith channel) → PartialBalancedChannel N tables channel := by intro balanced use [] diff --git a/Clean/Air/README.md b/Clean/Air/README.md index ab5c171ea..9a85c8242 100644 --- a/Clean/Air/README.md +++ b/Clean/Air/README.md @@ -21,7 +21,7 @@ In this terminology, a `Flat.Component` is a one-row AIR component: it packages It also proves the basic row-level transport lemmas: instantiated component operations agree with row operations, component soundness lifts to table soundness, and table interactions can be collected per channel. -`Balance.lean` contains the channel multiset theory. An interaction pushes a message to a channel or pulls it, a number of times given by its multiplicity. What balance means depends on the proof system: a `Balance` is the check a verifier performs on the interactions with each channel, like LogUp's field sums in prime fields or multiset equality with `0`/`1` multiplicities in binary fields, together with the numbers of interactions it is sound for (fewer than the characteristic for LogUp) and its proof of `BalancedInteractions` for those, the facts that soundness proofs use: every pull is matched by an interaction that provides its message, and messages with `0`/`1` multiplicities are pushed as often as they are pulled. The file proves permutation and counting lemmas, and provides the channel-level implication principles used by higher-level soundness proofs. It also defines `RawChannel.Consistent` and `RawChannel.Normal`; typed channels are normal by construction, and normal channels are consistent, so both properties are satisfied in practice. A highlight in `Balance.lean` is the "guarantees-to-requirements-reversal" theorem which provides the basis for soundness of VM channels. A balance check also records how high a pull multiplicity it matches with a push, in its `pullBound`, together with the proof of `MatchedUpTo` for that bound (`matchedUpTo_of_balanced`). A check that matches multiplicity `1` only takes the default proof and is unchanged; `Balance.logUpUpTo N` raises the bound to `N` by additionally asking its `fits` for `count * N < ringChar F`, which `exists_logUpBalanced_not_matchedUpTo` shows a bound on the count alone does not give. `RawChannel.ConsistentUpTo N` generalizes `RawChannel.Consistent`, which is `ConsistentUpTo 1`. +`Balance.lean` contains the channel multiset theory. An interaction pushes a message to a channel or pulls it, a number of times given by its multiplicity. What balance means depends on the proof system: a `Balance` is the check a verifier performs on the interactions with each channel, like LogUp's field sums in prime fields or multiset equality with `0`/`1` multiplicities in binary fields, together with the numbers of interactions it is sound for (fewer than the characteristic for LogUp) and its proof of `BalancedInteractions` for those, the facts that soundness proofs use: every pull of multiplicity `1` is matched by an interaction that provides its message, and messages with `0`/`1` multiplicities are pushed as often as they are pulled. The file proves permutation and counting lemmas, and provides the channel-level implication principles used by higher-level soundness proofs. It also defines `RawChannel.Consistent` and `RawChannel.Normal`; typed channels are normal by construction, and normal channels are consistent, so both properties are satisfied in practice. A highlight in `Balance.lean` is the "guarantees-to-requirements-reversal" theorem which provides the basis for soundness of VM channels. A balance check also records how high a pull multiplicity it matches with a push, in its `pullBound`, together with the proof of `MatchedUpTo` for that bound (`matchedUpTo_of_balanced`). A check that matches multiplicity `1` only takes the default proof and is unchanged; `Balance.logUpUpTo N` raises the bound to `N` by additionally asking its `fits` for `count * N < ringChar F`, which `exists_logUpBalanced_not_matchedUpTo` shows a bound on the count alone does not give. `RawChannel.ConsistentUpTo N` generalizes `RawChannel.Consistent`, which is `ConsistentUpTo 1`. `Verifier.lean` defines the verifier program language: an append-only list of public interactions, together with the verifier's spec and its soundness. @@ -29,7 +29,7 @@ It also proves the basic row-level transport lemmas: instantiated component oper `Lookup.lean` defines a convenience wrapper for static lookups. -`CountedChannel.lean` defines `Channel.counted N`, a typed channel whose pulls may carry a multiplicity of up to `N` rather than only `0` or `1`. A row that needs the same fact `N` times pulls it once at multiplicity `N`, costing one interaction instead of `N`. In exchange for the guarantee, the channel asks such a pull to account for its count (`mult = 0 ∨ CountUpTo N mult`), and it is consistent up to `N` instead of up to `1`, so finishing it in a `SoundEnsemble` needs a balance check whose `pullBound` is at least `N`. `Channel.counted_not_consistent` shows that the stronger check is necessary and not merely convenient: under plain `BalancedInteractions`, two pulls that both account for their counts can still leave a guarantee unsupported. `Clean/Examples/CountedSquares` assembles a sound ensemble that uses one. +`CountedChannel.lean` defines `Channel.counted N channel`, the raw channel of a typed `channel` whose pulls may carry a multiplicity of up to `N` rather than only `0` or `1`. A row that needs the same fact `N` times pulls it once at multiplicity `N`, costing one interaction instead of `N`. In exchange for the guarantee, the channel asks such a pull to account for its count (`mult = 0 ∨ CountUpTo N mult`), and it is consistent up to `N` instead of up to `1`, so finishing it in a `SoundEnsemble` needs a balance check whose `pullBound` is at least `N`. `Channel.counted_not_consistent` shows that the stronger check is necessary and not merely convenient: under plain `BalancedInteractions`, two pulls that both account for their counts can still leave a guarantee unsupported. `Clean/Examples/CountedSquares` assembles a sound ensemble that uses one. `FlatEnsemble.lean` defines flat AIR ensembles, `Flat.Ensemble` and their witnesses, `Flat.EnsembleWitness`. An ensemble is everything a prover and a verifier implementation need: components, channels, a verifier program, and the type of the prover's private input. The verifier only contributes public interactions directly. The ensemble lists every channel its components and verifier interact with, under unique names, so a proof system can balance channels by name (`EnsembleWitness.interactionsWith_eq_filter_name`). Its `Statement` is the raw proof-system relation for a balance check: there exists a witness whose table constraints hold and whose interactions with every channel pass the check. Every table of a witness carries the prover data derived from all the witness tables and the public input, so channel guarantees and specs can refer to both. The statement also fixes the table heights. Each row of a component interacts with a channel equally often, so the heights determine the number of interactions with every channel (`Ensemble.interactionCount`, which counts by channel name), and the statement requires the balance check to be sound for those numbers, a computable condition that the verifier checks from the heights (`Ensemble.interactionsFit`). The ensemble file also defines soundness and completeness and the `FormalEnsemble` structure which bundles an ensemble with its `Spec`, `Assumptions`, the `ProverAssumptions` of witness generation, and the soundness proof under a balance check (completeness is TODO), like `FormalCircuit` does for a circuit. diff --git a/Clean/Examples/CountedSquares/WitnessGenerationTest.lean b/Clean/Examples/CountedSquares/WitnessGenerationTest.lean index 0b0470332..c817d0624 100644 --- a/Clean/Examples/CountedSquares/WitnessGenerationTest.lean +++ b/Clean/Examples/CountedSquares/WitnessGenerationTest.lean @@ -13,7 +13,7 @@ namespace CountedSquares.WitnessGenerationTest open Air.Flat Air.Flat.WitnessGeneration -/-- Small enough to run in, and large enough for the bound: the runs below make `6` and `8` +/-- Small enough to run in, and large enough for the bound: the runs below make `6`, `7` and `8` interactions on the channel, and `Balance.logUpUpTo 2` asks for `8 * 2 < p1009`. -/ abbrev p : ℕ := p1009 From bebc0cda63e69c75f3f93213f4a0eeab5a994cd7 Mon Sep 17 00:00:00 2001 From: jotabulacios Date: Thu, 19 Nov 2026 15:06:48 -0300 Subject: [PATCH 29/29] Name the declarations the docstrings describe, and restore the generation caveat. A parallel review round on the earlier version of this PR made these changes; the port of that round took the ones it listed and left these. `MatchedUpTo` cited `BalancedInteractions.exists_push_of_pull`, a field of the structure, where `matchedUpTo_one` is the lemma that produces `MatchedUpTo 1`; the two are separated by `Interaction.pullUpTo_one_iff`. `LogUpBalanced.matchedUpTo` now points at the counterexample that shows its bound is sharp, which already points back. The `CountedChannel.lean` module docstring says how consistency above `N` is obtained, the only mention `counted_consistentUpTo_of_le` has. `pullUpTo` claimed `count` had to count up to `N`, excluding the `0` that `Direction.Requirements` also allows. `CountedSquares` regains the caveat that generation does not split a row at `N`: a demand component keeping one row per message (`DemandMode.answer`) would, if it pulled with that count, give a message demanded more than `N` times a row whose counted pull fails its constraint. `countUpTo_of_mul_eq_zero` named a hypothesis it does not have: `a * b = 0` is weaker than `(c - 1) * (c - 2) = 0`. --- Clean/Air/Balance.lean | 12 ++++++------ Clean/Air/CountedChannel.lean | 6 ++++-- Clean/Examples/CountedSquares/Ensemble.lean | 14 +++++++++----- 3 files changed, 19 insertions(+), 13 deletions(-) diff --git a/Clean/Air/Balance.lean b/Clean/Air/Balance.lean index b9bdd48e1..5f9c06e5e 100644 --- a/Clean/Air/Balance.lean +++ b/Clean/Air/Balance.lean @@ -141,11 +141,11 @@ lemma BalancedInteractions.nil : BalancedInteractions ([] : List (Interaction F) ⟨ by simp, by simp [activeCount] ⟩ /-- -Every pull of multiplicity up to `N` is matched by an interaction that pushes its message. For -`N = 1` this is `BalancedInteractions.exists_push_of_pull`. For larger `N` it needs more than balance -of the counts: under field-sum balance, `(p - 1) / 2` pulls of multiplicity `2` and one pull of -multiplicity `1` sum to `p ≡ 0` with no push. It holds when the number of interactions times `N` -is below the characteristic, see `LogUpBalanced.matchedUpTo`. +Every pull of multiplicity up to `N` is matched by an interaction that pushes its message. +`BalancedInteractions` implies it for `N = 1` (`BalancedInteractions.matchedUpTo_one`). For larger +`N` it needs more than balance of the counts: under field-sum balance, `(p - 1) / 2` pulls of +multiplicity `2` and one pull of multiplicity `1` sum to `p ≡ 0` with no push. It holds when the +number of interactions times `N` is below the characteristic, see `LogUpBalanced.matchedUpTo`. -/ def MatchedUpTo (N : ℕ) (interactions : List (Interaction F)) : Prop := ∀ a ∈ interactions, a.PullUpTo N → ∃ b ∈ interactions, b.msg = a.msg ∧ b.PushUpTo N @@ -358,7 +358,7 @@ lemma balanceOf_eq_neg_of_pullUpTo {N : ℕ} {msg : Array F} (hN : N < ringChar /-- Every pull of multiplicity up to `N` is matched by a push, when the number of interactions times `N` is below the characteristic: otherwise the multiplicities of the pulls of a message could add -up to a multiple of the characteristic. +up to a multiple of the characteristic (`exists_logUpBalanced_not_matchedUpTo`). -/ theorem LogUpBalanced.matchedUpTo {N : ℕ} {interactions : List (Interaction F)} (balanced : LogUpBalanced interactions) (length_mul_lt : interactions.length * N < ringChar F) : diff --git a/Clean/Air/CountedChannel.lean b/Clean/Air/CountedChannel.lean index 1fe60a7b7..ca722a3a8 100644 --- a/Clean/Air/CountedChannel.lean +++ b/Clean/Air/CountedChannel.lean @@ -12,7 +12,8 @@ A table that answers `k` identical lookups with one row pulls the facts it relie multiplicity `k`. `Channel.counted N channel` is `channel` with pulls of multiplicity up to `N`: such a pull gets the guarantee, and an active push owes it. It is consistent in the sense of `RawChannel.ConsistentUpTo N`, so it can be finished in a `SoundEnsemble` whose balance check has a -`pullBound` of at least `N`, e.g. `Balance.logUpUpTo N`. +`pullBound` of at least `N`, such as `Balance.logUpUpTo N`: at exactly `N` by instance, above it +through `counted_consistentUpTo_of_le`. This was written for lambda_vm, a RISC-V zkVM built on Clean, whose arithmetic chips answer several identical lookups per row. `Clean/Examples/CountedSquares` is the same shape in miniature. @@ -80,7 +81,8 @@ def emittedUpTo (N : ℕ) (channel : Channel F Message) (mult : Expression F) (msg : Message (Expression F)) : AbstractInteraction F := ⟨ channel.counted N, .push, mult, toElements msg ⟩ -/-- Pull `msg` `count` times, for `count` up to `N`. -/ +/-- Pull `msg` `count` times from `channel.counted N`, which requires `count` to be `0` or to count +up to `N`. -/ @[circuit_norm] def pullUpTo (N : ℕ) (channel : Channel F Message) (count : Expression F) (msg : Message (Expression F)) : Circuit F Unit := fun _ => diff --git a/Clean/Examples/CountedSquares/Ensemble.lean b/Clean/Examples/CountedSquares/Ensemble.lean index b2da5d3b8..44ee08a7d 100644 --- a/Clean/Examples/CountedSquares/Ensemble.lean +++ b/Clean/Examples/CountedSquares/Ensemble.lean @@ -27,8 +27,11 @@ there to seed the run and nothing else: a VM seeds its channels from the verifie `Clean/Examples/FemtoCairo/FlatAir.lean`. `Clean/Examples/CountedSquares/WitnessGenerationTest.lean` runs it and pins the result. -A row's multiplicity is counted by `Tally.signedCount`, which reads it as a natural number, so a -counted pull is tallied as the number it claims; that part needs no change for counted channels. +Generation tallies a counted pull as the natural number its multiplicity claims +(`Tally.signedCount`), but does not split a row at `N`. A demand component with a multiplicity cell +keeps one row per message (`DemandMode.answer`), so if it also pulled with that count, a message +demanded more than `N` times would get a row whose counted pull fails its constraint. The consumer +here pulls a fixed count per row instead. Constraining the count arithmetically only scales for a small bound. A circuit counting to `510` would range-check its multiplicity on another channel instead, which needs a finishing order over @@ -51,7 +54,8 @@ def squares : Channel (F p) (ProvablePair field field) where /-- A count of `1` or `2` is a count up to `2`: the step from a field constraint to the channel's `CountUpTo`, and the only arithmetic the example needs. -/ -lemma countUpTo_of_mul_eq_zero {c : F p} (h : (c - 1) * (c - 2) = 0) : CountUpTo countBound c := by +lemma countUpTo_of_sub_one_mul_sub_two_eq_zero {c : F p} (h : (c - 1) * (c - 2) = 0) : + CountUpTo countBound c := by rcases mul_eq_zero.mp h with h | h <;> rw [sub_eq_zero] at h · exact ⟨1, le_rfl, by norm_num, by simp [h]⟩ · exact ⟨2, by norm_num, le_rfl, by simp [h]⟩ @@ -111,12 +115,12 @@ def consumer : GeneralFormalCircuit (F p) ConsumerInput unit where requirementsChannelsLawful := by simp only [circuit_norm, seval] intro input env h - exact .inr (countUpTo_of_mul_eq_zero h) + exact .inr (countUpTo_of_sub_one_mul_sub_two_eq_zero h) soundness := by circuit_proof_start [squares] obtain ⟨h_count, h_guarantee⟩ := h_holds - have h_in := countUpTo_of_mul_eq_zero h_count + have h_in := countUpTo_of_sub_one_mul_sub_two_eq_zero h_count exact ⟨h_guarantee h_in, .inr h_in⟩ completeness := by circuit_proof_start [squares]