diff --git a/research/clifford-unknown-support-decoder/.gitattributes b/research/clifford-unknown-support-decoder/.gitattributes new file mode 100644 index 0000000000..5067585a37 --- /dev/null +++ b/research/clifford-unknown-support-decoder/.gitattributes @@ -0,0 +1 @@ +tables/*.json -text diff --git a/research/clifford-unknown-support-decoder/CREDITS.md b/research/clifford-unknown-support-decoder/CREDITS.md new file mode 100644 index 0000000000..57cf927044 --- /dev/null +++ b/research/clifford-unknown-support-decoder/CREDITS.md @@ -0,0 +1,17 @@ +# Sources and credit + +Ali Romman (@aliromman) prepared the new reconstruction algorithm, support certificates, two-register example, implementation and campaign synthesis with substantial OpenAI Codex assistance. New material is Apache-2.0. Separate Codex reviewers attacked the cut and phase arguments, implemented independent character, Pauli and dense-minor computations and tested the executable. Their checks do not constitute independent human review or formal proof-kernel acceptance. + +## Mathematical and data dependencies + +- Nick Petalas (@petalas) supplied the four-qubit quotient calculation and CTblLib projection in PR 7747. The 7,054-byte projection here is byte-exact. His proof credits alephnan for the odd-sector reduction, Ali Romman for the pointwise framework, and his own earlier lower bound. The published ordinary table is credited to the ATLAS and Thomas Breuer's CTblLib 1.3.11. [PR 7747](https://github.com/CrocSwap/integer-mult-bounds/pull/7747) · [Projection](https://github.com/petalas/integer-mult-bounds/blob/ff12ca358d601a21cd4b1f016b8a3a969ac44016/research/clifford-four-qubit-width-nine/PROJECTION.json) · [Source proof and notice](https://github.com/petalas/integer-mult-bounds/blob/ff12ca358d601a21cd4b1f016b8a3a969ac44016/research/clifford-four-qubit-width-nine/README.md) +- alephnan (@alephnan) developed the Pauli central-character and projective Gram framework in PR 6714. Its predecessor record preserves Ali Romman's pointwise definition, MatShep's small-qubit results and the Lee–Yu–Peng–Lai table. This package uses the four-qubit implications identified by the credited successors. [PR 6714](https://github.com/CrocSwap/integer-mult-bounds/pull/6714) · [Proof](https://github.com/alephnan/integer-mult-bounds/blob/f7f32a18dde54e67624a9ed9cc7e1b266d9585c0/research/clifford-three-qubit-threshold/PROOF.md) +- Ali Romman supplied the full width-13 threshold and width-11 decomposition, the complete width-12 projector and its four-row 10,561-byte projection, and the exact width-13 conditioning. The new work derives restricted-support and fibre guarantees from these results. It does not originate the tables or the predecessor thresholds. [PR 7766](https://github.com/CrocSwap/integer-mult-bounds/pull/7766) · [Odd-width proof](https://github.com/aliromman/integer-mult-bounds/blob/d7c2dd1868c70f0addce2c0a4d3220149452f14b/research/clifford-four-qubit-threshold/PROOF.md) · [PR 7770](https://github.com/CrocSwap/integer-mult-bounds/pull/7770) · [Even-width proof](https://github.com/aliromman/integer-mult-bounds/blob/ff9bdc0e82e75ab5ed39f3ff034ac8ddb3ea06d2/research/clifford-width-twelve-nullity/PROOF.md) · [Four character rows](https://github.com/aliromman/integer-mult-bounds/blob/ff9bdc0e82e75ab5ed39f3ff034ac8ddb3ea06d2/research/clifford-width-twelve-nullity/CHARACTER-COLUMNS.json) · [PR 7772](https://github.com/CrocSwap/integer-mult-bounds/pull/7772) +- The column recovery-norm comparison uses the full-spectrum bound and its odd and Bell-sector refinements. The odd-division obstruction concerns the canonical inverse. This algorithm has an exact Gaussian-rational arithmetic contract. [Odd spectrum](https://github.com/CrocSwap/integer-mult-bounds/pull/7788) · [Bell-sector spectrum](https://github.com/CrocSwap/integer-mult-bounds/pull/7789) · [Canonical divisor obstruction](https://github.com/CrocSwap/integer-mult-bounds/pull/7792) +- S. E. Leurgans, R. T. Ross and R. B. Abel proved three-way uniqueness with two independent factors and projectively distinct third factors. Ivan Oseledets developed tensor-train representations. Jeroen Dehaene and Bart De Moor gave binary and quadratic descriptions of Clifford operations and stabilizer states. These form the classical background. The new cut procedure exploits the finite phase ratios, exact Clifford test, restricted kernels and full residual on a supplied train. [Three-way decomposition](https://epubs.siam.org/doi/10.1137/0614071) · [Tensor-train decomposition](https://epubs.siam.org/doi/10.1137/090752286) · [Clifford normal forms](https://arxiv.org/abs/quant-ph/0304125) + +## Distinct pending constructions + +apoursh (@apoursh) supplied a sharp Gaussian-dyadic Bell decoder, 4,265 reads on a named four-qubit fibre, a compact identity-row contraction, the 256-output identity-fibre implementation, canonical denominator results, and a 51/52-pass Gram schedule. The chart contraction credits Ali Romman's earlier circuit evaluators. These constructions motivate the access-model comparison. Their row implementation, operator schedule and denominator computations are not dependencies of the new executable. [Sharp Bell width](https://github.com/CrocSwap/integer-mult-bounds/pull/7799) · [Compiled row](https://github.com/CrocSwap/integer-mult-bounds/pull/7800) · [Compact chart](https://github.com/CrocSwap/integer-mult-bounds/pull/7801) · [Optimal norm and denominator](https://github.com/CrocSwap/integer-mult-bounds/pull/7802) · [Canonical inverse denominator](https://github.com/CrocSwap/integer-mult-bounds/pull/7805) · [Gram schedules](https://github.com/CrocSwap/integer-mult-bounds/pull/7809) · [Compact full identity fibre](https://github.com/CrocSwap/integer-mult-bounds/pull/7813) + +Clear1nsight (@Clear1nsight) supplied the pending P104 rational and binary construction used for the complete-cost comparison. The cooperative receiver-cut classification belongs to apoursh. Each row in the campaign inventory links its original scientific source and contributor evidence. [P104 application](https://github.com/CrocSwap/integer-mult-bounds/pull/7786) · [Cooperative cut](https://github.com/CrocSwap/integer-mult-bounds/pull/7808) · [Campaign index](https://github.com/CrocSwap/integer-mult-bounds/discussions/7794) diff --git a/research/clifford-unknown-support-decoder/INVENTORY.md b/research/clifford-unknown-support-decoder/INVENTORY.md new file mode 100644 index 0000000000..dd4df27d07 --- /dev/null +++ b/research/clifford-unknown-support-decoder/INVENTORY.md @@ -0,0 +1,135 @@ +# Inventory of the 100 campaign results + +Each row states the contribution made at its frozen head, the class in which it applies and the remaining cost or open interface. The evidence link names the source revision. The credit link gives its inputs and their authors. A graph, character table, compiler or claimed exponent used as an input remains the work of its source authors. An audit adds a check of the stated claim. It does not originate the audited construction. The public index also links the handoff for each result. [Campaign index](https://github.com/CrocSwap/integer-mult-bounds/discussions/7794) + +Results 1–10 start at `7c49fe32c55de02c2c873655c804b16e4c3fef0a`. Results 11–100 start at `e54523517d0f03e58352e647c403a5af3e525a57`. Historical uses of “selected” or “current” refer to the word in that row. Only row 50 supplies a complete conditional κ application. See [the direction comparison](SYNTHESIS.md) for the later pending work, cross-group contracts and the sources used by the new decoder. + +## Native precision, tapes and ownership + +| Result | Established result, cost and boundary | Frozen record | +|---:|---|---| +| 2 | The fixed-grid C0 callee and spectator planner compose if operations stay within complete rows and undo the final permutation. A fast recursive child still owes this return. | [PR 6480](https://github.com/CrocSwap/integer-mult-bounds/pull/6480) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/ecc322ebd18978e5eddb6e140081b07d7aee33b4/research/rounded-spectator-callee/README.md) · [Lineage 5485, 6070](https://github.com/aliromman/integer-mult-bounds/blob/ecc322ebd18978e5eddb6e140081b07d7aee33b4/research/rounded-spectator-callee/README.md) | +| 3 | Three modular-inverse lift domains give different Lipschitz bounds. The three-coordinate query has a corrected paid SWAP word. A native residue producer still owes conversion. | [PR 6491](https://github.com/CrocSwap/integer-mult-bounds/pull/6491) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/3d9dab36c3cc58d3a160f06e5f24185bb7fa0d4a/research/modular-normalization-boundary/README.md) · [Lineage 1159](https://github.com/aliromman/integer-mult-bounds/blob/3d9dab36c3cc58d3a160f06e5f24185bb7fa0d4a/research/modular-normalization-boundary/PROOF.md) | +| 5 | The actual three-event warp, climb and half-ADD join returns correctly on 1,024 columns in both directions. Other emitted joins and the complete word require their own proof. | [PR 6508](https://github.com/CrocSwap/integer-mult-bounds/pull/6508) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/320c3bb73c5bdf231b6504dbe0cfe19410a78b8b/research/emitted-native-boundary-join/README.md) · [Lineage 4088, 4424](https://github.com/aliromman/integer-mult-bounds/blob/320c3bb73c5bdf231b6504dbe0cfe19410a78b8b/research/emitted-native-boundary-join/PROOF.md) | +| 6 | For the selected historical #6030 word, completed conditional mean-zero errors give 13 individual RMS guard bits and a 16-bit stopped tail bound. The noise contract must hold at each completed edge. | [PR 6526](https://github.com/CrocSwap/integer-mult-bounds/pull/6526) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/f2431f81bfc1a7b418ce3a1b02b958d5f259f50b/research/selected-whole-word-stochastic/README.md) · [Lineage 6030, 5819](https://github.com/aliromman/integer-mult-bounds/blob/f2431f81bfc1a7b418ce3a1b02b958d5f259f50b/research/selected-whole-word-stochastic/PROOF.md) | +| 7 | At the first native MOVE one fair bit per butterfly layer gives optimal pointwise grid error and a retained-seed inverse. A full recursive word also needs the discarded information. | [PR 6532](https://github.com/CrocSwap/integer-mult-bounds/pull/6532) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/9327b8f6b7f94728f188f8a27595bd8b14775446/research/selected-native-unbiased-rounding/README.md) · [Lineage 6509, 6438, 6526](https://github.com/aliromman/integer-mult-bounds/blob/9327b8f6b7f94728f188f8a27595bd8b14775446/research/selected-native-unbiased-rounding/PROOF.md) | +| 8 | At grid ratio 2^k with k at least half the active axes rounded up, a Gaussian child coarsens without bias with 2kN retained bits. The literal direct invoice includes that stock. | [PR 6540](https://github.com/CrocSwap/integer-mult-bounds/pull/6540) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/5f7c893b6af9a75cb19ad22c6db06ed22eea6f54/research/native-coarse-grid-residue-codec/README.md) · [Lineage 5815, 5821, 6506](https://github.com/aliromman/integer-mult-bounds/blob/5f7c893b6af9a75cb19ad22c6db06ed22eea6f54/research/native-coarse-grid-residue-codec/PROOF.md) | +| 9 | All 22,803 events of the historical word admit a same-grid stochastic caller and literal retained-state reversal. Its absolute error is bounded at every prefix. A faster native child and a priced history remain. | [PR 6546](https://github.com/CrocSwap/integer-mult-bounds/pull/6546) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/ff038070b8d122a43993f2b2524154323cc6b4d7/research/selected-absolute-stochastic-caller/README.md) · [Lineage 6030, 6526, 6532, 6540](https://github.com/aliromman/integer-mult-bounds/blob/ff038070b8d122a43993f2b2524154323cc6b4d7/research/selected-absolute-stochastic-caller/PROOF.md) | +| 10 | At fixed seed, 2^(86N) distinct inputs share one complete output of that caller. Exact reversal of this family needs at least 86N hidden bits. A changed output or input promise changes the obstruction. | [PR 6547](https://github.com/CrocSwap/integer-mult-bounds/pull/6547) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/317e4a099488794c372f8a4ad1e70deef3626bbf/research/selected-caller-zero-fiber/README.md) · [Lineage 6546, 6030](https://github.com/aliromman/integer-mult-bounds/blob/317e4a099488794c372f8a4ad1e70deef3626bbf/research/selected-caller-zero-fiber/PROOF.md) | +| 11 | Literal instructions and output owners of #7113 give both constants in the 17-stage paid and productive bills. A caller with another native interface must derive its own constants. | [PR 7661](https://github.com/CrocSwap/integer-mult-bounds/pull/7661) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/1e6fd08a7c4fb616ea802c5f432f58594d14a107/research/caller-8x9-derived-residuals/README.md) · [Lineage 7113](https://github.com/aliromman/integer-mult-bounds/blob/1e6fd08a7c4fb616ea802c5f432f58594d14a107/research/caller-8x9-derived-residuals/PROOF.md) | +| 12 | The actual clock-100504 half-ADD accepts the pinned rounded service, with a terminal error identity. A second site and the full dispatcher need compatible charts and grids. | [PR 7662](https://github.com/CrocSwap/integer-mult-bounds/pull/7662) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/68b18142473c518102cbe2c151645a60dc9eb7f6/research/current-half-add-service/README.md) · [Lineage 7660, 6988](https://github.com/aliromman/integer-mult-bounds/blob/68b18142473c518102cbe2c151645a60dc9eb7f6/research/current-half-add-service/PROOF.md) | +| 13 | All 143,376 ADD/SCALE occurrences in #7660 now have local rounded tables and a weighted scalar bill. Recursive field delivery and ownership remain. | [PR 7665](https://github.com/CrocSwap/integer-mult-bounds/pull/7665) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/7b4e792e91706dd0e00f51a10402696fef5998ec/research/current-native-rounded-catalogue/README.md) · [Lineage 7660, 6988, 7662](https://github.com/aliromman/integer-mult-bounds/blob/7b4e792e91706dd0e00f51a10402696fef5998ec/research/current-native-rounded-catalogue/PROOF.md) | +| 14 | The #7663 owned MOVE can pay actual travel to and from a private bank and preserve the others. The proof takes bank lengths and owners as inputs. | [PR 7666](https://github.com/CrocSwap/integer-mult-bounds/pull/7666) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/93710078d292f6702aaa12d47a1c97a5e3051674/research/current-native-bank-selector/README.md) · [Lineage 7663, 7586](https://github.com/aliromman/integer-mult-bounds/blob/93710078d292f6702aaa12d47a1c97a5e3051674/research/current-native-bank-selector/PROOF.md) | +| 15 | A seven-head program constructs, uses and retires the flat guide from physical descriptors. A compiled child must supply those descriptors. | [PR 7667](https://github.com/CrocSwap/integer-mult-bounds/pull/7667) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/776e58c3879193d68d0da586d0ff5c07bdb96010/research/current-flat-guide-lifecycle/README.md) · [Lineage 7663, 7666](https://github.com/aliromman/integer-mult-bounds/blob/776e58c3879193d68d0da586d0ff5c07bdb96010/research/current-flat-guide-lifecycle/PROOF.md) | +| 16 | One shared guide serves the 234-parent/234-child tuple across 468 resets. All travel and cleanup are charged. A literal numerical child preserving that tuple is still missing. | [PR 7670](https://github.com/CrocSwap/integer-mult-bounds/pull/7670) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/677d23ef1610db9921547ab5d58eecc8c8d4c08b/research/current-shared-guide-banks/README.md) · [Lineage 7667, 7666, 7663](https://github.com/aliromman/integer-mult-bounds/blob/677d23ef1610db9921547ab5d58eecc8c8d4c08b/research/current-shared-guide-banks/PROOF.md) | +| 17 | The #7669 exact triangle can change rounded tapes by (0,h,2h) on one integer input. Reachability in a native caller is a separate premise. | [PR 7671](https://github.com/CrocSwap/integer-mult-bounds/pull/7671) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/f34696319eb19d05d1b09f7372c3146be3ea1635/research/current-triangle-rounded-contract/README.md) · [Lineage 7669](https://github.com/aliromman/integer-mult-bounds/blob/f34696319eb19d05d1b09f7372c3146be3ea1635/research/current-triangle-rounded-contract/PROOF.md) | +| 18 | The three rounding errors and three native MOVE errors give an explicit chart-aware output operator. The application needs reachable allowances at the six sites. | [PR 7672](https://github.com/CrocSwap/integer-mult-bounds/pull/7672) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/22f492a68bfa5c8d64bc08330b2e2227fa9190d0/research/current-triangle-native-error/README.md) · [Lineage 7669, 7671](https://github.com/aliromman/integer-mult-bounds/blob/22f492a68bfa5c8d64bc08330b2e2227fa9190d0/research/current-triangle-native-error/PROOF.md) | +| 19 | Exact sum, injectivity and pair error at most one force outward slack at least sqrt(13)-3. Result 36 supersedes this lower bound. | [PR 7673](https://github.com/CrocSwap/integer-mult-bounds/pull/7673) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/8ebf5eec9188f29c98be86eb1247f2ec6b41f19f/research/gaussian-exact-sum-next-obstruction/README.md) · [Lineage 7652, 7570](https://github.com/aliromman/integer-mult-bounds/blob/8ebf5eec9188f29c98be86eb1247f2ec6b41f19f/research/gaussian-exact-sum-next-obstruction/PROOF.md) | +| 24 | The literal products of the fixed #7674 DAG need odd factor 225 and 12 binary bits. Complete root and chart precision still need a caller. | [PR 7680](https://github.com/CrocSwap/integer-mult-bounds/pull/7680) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/533cdca9225d1c0af1ae8d3b79cb527f59afaf0f/research/banked-native-minimum-grid/README.md) · [Lineage 7674, 7669](https://github.com/aliromman/integer-mult-bounds/blob/533cdca9225d1c0af1ae8d3b79cb527f59afaf0f/research/banked-native-minimum-grid/PROOF.md) | +| 25 | With fixed row order and distinct snapshots, peak stock 2,425 is attained and necessary. Other legal orders can change the live cut. | [PR 7681](https://github.com/CrocSwap/integer-mult-bounds/pull/7681) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/1753a3ff56be34b09a217b12b590f51eba15defb/research/banked-native-snapshot-reuse/README.md) · [Lineage 7674, 7680](https://github.com/aliromman/integer-mult-bounds/blob/1753a3ff56be34b09a217b12b590f51eba15defb/research/banked-native-snapshot-reuse/PROOF.md) | +| 36 | The all-lattice injective exact-sum error-one optimum is outward slack 1/sqrt(2). The attaining bijection still needs the caller's symmetries and ports. | [PR 7696](https://github.com/CrocSwap/integer-mult-bounds/pull/7696) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/d7be8d9131a948ea1d25833e96906ea3da979764/research/gaussian-exact-sum-global-optimum/README.md) · [Lineage 7673, 7652, 7570](https://github.com/aliromman/integer-mult-bounds/blob/d7be8d9131a948ea1d25833e96906ea3da979764/research/gaussian-exact-sum-global-optimum/CREDITS.md) | +| 37 | A square-ring matching attains that slack with swap, translation and Gaussian phase covariance. Indexed ports and a physical tape implementation still cost work. | [PR 7698](https://github.com/CrocSwap/integer-mult-bounds/pull/7698) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/f821eb5ea909e5c56118c5042f7740d87f0355b2/research/phase-covariant-radial-ports/README.md) · [Lineage 7696, 6509](https://github.com/aliromman/integer-mult-bounds/blob/f821eb5ea909e5c56118c5042f7740d87f0355b2/research/phase-covariant-radial-ports/CREDITS.md) | +| 38 | Every optimal error-one exact-sum injection has unbounded cycle lengths. A caller demanding finite iterate order must pay extra radial slack. | [PR 7699](https://github.com/CrocSwap/integer-mult-bounds/pull/7699) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/2a5dcb314ccdfeed24bd86ec1f56f519eead56c1/research/gaussian-radial-order-classification/README.md) · [Lineage 7570, 7696, 7698](https://github.com/aliromman/integer-mult-bounds/blob/2a5dcb314ccdfeed24bd86ec1f56f519eead56c1/research/gaussian-radial-order-classification/CREDITS.md) | +| 39 | Independent basis replay of chart 595 gives literal scale 45,158,400 and stored scale 161,280. The factor 49 comes from two input columns. | [PR 7701](https://github.com/CrocSwap/integer-mult-bounds/pull/7701) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/cd3819c70c6417bc1d3cc138294d973c58735229/research/chart595-factor-seven-audit/README.md) · [Lineage 7697, 7680](https://github.com/aliromman/integer-mult-bounds/blob/cd3819c70c6417bc1d3cc138294d973c58735229/research/chart595-factor-seven-audit/CREDITS.md) | +| 40 | One integer numerator per row realizes the same 7,644 stored rows at scale 161,280. Deferred division changes accumulator range and the scalar implementation. | [PR 7702](https://github.com/CrocSwap/integer-mult-bounds/pull/7702) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/785860501b03893eaa9f5279f978287e75d3f3af/research/chart595-row-atomic-arithmetic/README.md) · [Lineage 7697, 7701, 7680](https://github.com/aliromman/integer-mult-bounds/blob/785860501b03893eaa9f5279f978287e75d3f3af/research/chart595-row-atomic-arithmetic/CREDITS.md) | +| 44 | For the fixed role-1256 scalar graph and order, the sharp prefix bound is 53,895,757/256. Transported physical components require result 59. | [PR 7709](https://github.com/CrocSwap/integer-mult-bounds/pull/7709) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/762a4d8c541081be987f2fdf9053d4d7b52990f4/research/relay1256-sharp-scalar-range/README.md) · [Lineage 7704, 7700, 7702](https://github.com/aliromman/integer-mult-bounds/blob/762a4d8c541081be987f2fdf9053d4d7b52990f4/research/relay1256-sharp-scalar-range/CREDITS.md) | +| 47 | An error-one exact-sum pair map with F^m=I exists exactly for 4 dividing m. A formula gives the sharp outward slack for each such m. An actual finite-order caller is needed. | [PR 7714](https://github.com/CrocSwap/integer-mult-bounds/pull/7714) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/ba5e006ed70119b23d6ef9e2e8627e43a3f4f18b/research/gaussian-finite-iterate-optimum/README.md) · [Lineage 7570, 7699, 7696](https://github.com/aliromman/integer-mult-bounds/blob/ba5e006ed70119b23d6ef9e2e8627e43a3f4f18b/research/gaussian-finite-iterate-optimum/CREDITS.md) | +| 56 | At width 1,076, the #7721 half-sum CLEAR uses 25 data tapes and 11,206,982 interpreter steps with full return. Uniform generation and global stock remain. | [PR 7728](https://github.com/CrocSwap/integer-mult-bounds/pull/7728) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/fae44c3ea2bc24e42f678365282f3737014fe9b3/research/half-sum-clear-tape/README.md) · [Lineage 7721, 7669](https://github.com/aliromman/integer-mult-bounds/blob/fae44c3ea2bc24e42f678365282f3737014fe9b3/research/half-sum-clear-tape/CREDITS.md) | +| 59 | One role-1256 ordered prefix gives the sharp physical component envelope for every positive tensor repetition. Global caller correlations may lower attainable ranges. | [PR 7731](https://github.com/CrocSwap/integer-mult-bounds/pull/7731) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/5834f421f2751f63c1ddb459b5c2d30a3f8a9a7b/research/relay1256-tensor-prefix-range/README.md) · [Lineage 7725, 7709, 7704](https://github.com/aliromman/integer-mult-bounds/blob/5834f421f2751f63c1ddb459b5c2d30a3f8a9a7b/research/relay1256-tensor-prefix-range/CREDITS.md) | +| 60 | Every permitted order of the fifteen-pivot graph needs at least 1,507 snapshot banks. A known order uses 2,106. The intervening gap remains. | [PR 7732](https://github.com/CrocSwap/integer-mult-bounds/pull/7732) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/54dd8940c86fe6b423e4b6c55e08764b1985eedf/research/banked-native-order-lower-bound/README.md) · [Lineage 7724, 7674](https://github.com/aliromman/integer-mult-bounds/blob/54dd8940c86fe6b423e4b6c55e08764b1985eedf/research/banked-native-order-lower-bound/CREDITS.md) | +| 61 | A canonical padded-rotation switch bit takes Θ(h) bit work and O(h) scratch. Literal ownership and head travel require result 62. | [PR 7734](https://github.com/CrocSwap/integer-mult-bounds/pull/7734) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/57bb6efed1be1dda9eb1a0f622ec2773bc8b0047/research/rotation-switch-closed-descent/README.md) · [Lineage 7504, 4100](https://github.com/aliromman/integer-mult-bounds/blob/57bb6efed1be1dda9eb1a0f622ec2773bc8b0047/research/rotation-switch-closed-descent/CREDITS.md) | +| 62 | One fourteen-tape program returns the switch in at most 1024(h+4) transitions with borrowed inputs and cleared scratch. A caller must supply its descriptors. | [PR 7735](https://github.com/CrocSwap/integer-mult-bounds/pull/7735) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/460eb0c026649b28ab3b00f45de136599d0570fa/research/linear-tape-rotation-switch/README.md) · [Lineage 7593, 7734, 7504](https://github.com/aliromman/integer-mult-bounds/blob/460eb0c026649b28ab3b00f45de136599d0570fa/research/linear-tape-rotation-switch/CREDITS.md) | +| 63 | A complete canonical rotation row uses at most three quotient/remainder rectangles, and some need all three. Physical emission and descriptor movement still cost work. | [PR 7740](https://github.com/CrocSwap/integer-mult-bounds/pull/7740) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/913d61fd9bd5d3d65a11c04f1baf50fc24ceb858/research/rotation-row-rectangles/README.md) · [Lineage 7734, 7504, 4100](https://github.com/aliromman/integer-mult-bounds/blob/913d61fd9bd5d3d65a11c04f1baf50fc24ceb858/research/rotation-row-rectangles/CREDITS.md) | +| 67 | A fifteen-tape owner emits a full padded-rotation row with T≤1280N(h+4) and additional stock ≤N+20(h+4). This accepts four descriptors, not an arbitrary recursive owner map. | [PR 7752](https://github.com/CrocSwap/integer-mult-bounds/pull/7752) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/857a826e162cf9d4ccaa20e02bee6a3389e67bb0/research/owned-canonical-row-producer/README.md) · [Lineage 7735, 7593, 7504, 7740](https://github.com/aliromman/integer-mult-bounds/blob/857a826e162cf9d4ccaa20e02bee6a3389e67bb0/research/owned-canonical-row-producer/CREDITS.md) | + +## Rational P104 and binary suppliers + +| Result | Established result, cost and boundary | Frozen record | +|---:|---|---| +| 1 | On the #5350 compressed Cao word, no single #5426 level drop lowers the paid moment on [0.99,1]. A coupled drop changes the search. | [PR 6475](https://github.com/CrocSwap/integer-mult-bounds/pull/6475) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/c9f27ddbee648bc190696ecfcd2370fbf4b55d3c/research/compressed-word-drop-invoice/README.md) · [Lineage 5350, 5426](https://github.com/aliromman/integer-mult-bounds/blob/c9f27ddbee648bc190696ecfcd2370fbf4b55d3c/research/compressed-word-drop-invoice/README.md) | +| 48 | A separate basis audit confirms 33 #7713 restrictions, all 67 receiver relations and every surrendered helper. These checks support the changed caller in 49. | [PR 7715](https://github.com/CrocSwap/integer-mult-bounds/pull/7715) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/4517ad2a3e9f06ca369a2625f652de4f57be2fb9/research/helper-quotient-basis-audit/README.md) · [Lineage 7713, 7708, 7700](https://github.com/aliromman/integer-mult-bounds/blob/4517ad2a3e9f06ca369a2625f652de4f57be2fb9/research/helper-quotient-basis-audit/CREDITS.md) | +| 49 | Two literal #7712 causal splits survive those restrictions with 1,470 independent fields and the same 6,054 paid keys. Result 50 prices this fixed graph. | [PR 7716](https://github.com/CrocSwap/integer-mult-bounds/pull/7716) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/1636b3c9b922d23b222b60dd52e005e60c592491/research/quotient-causal-split-composition/README.md) · [Lineage 7712, 7713, 7715](https://github.com/aliromman/integer-mult-bounds/blob/1636b3c9b922d23b222b60dd52e005e60c592491/research/quotient-causal-split-composition/CREDITS.md) | +| 50 | The full fixed two-split application gives conditional κ≥0.00376313686053 with all five inherited interfaces. Newer #7786 is the relevant proposed comparison. | [PR 7717](https://github.com/CrocSwap/integer-mult-bounds/pull/7717) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/586f4f81a8bd09a65a8e6d0bfff57a0fec8925e8/research/quotient-split-physical-application/README.md) · [Lineage 7716, 7713, 7704](https://github.com/aliromman/integer-mult-bounds/blob/586f4f81a8bd09a65a8e6d0bfff57a0fec8925e8/research/quotient-split-physical-application/CREDITS.md) | +| 51 | For that graph the native odd scale falls from 2,083,725 to 11,025 at binary height 20. The scalar-rate allowance falls from 2^29 to 2^22 without a new exponent. | [PR 7718](https://github.com/CrocSwap/integer-mult-bounds/pull/7718) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/ceed7f9c4050060aa8a9c4149c704e6151555e5b/research/quotient-split-minimal-odd-scale/README.md) · [Lineage 7717, 866](https://github.com/aliromman/integer-mult-bounds/blob/ceed7f9c4050060aa8a9c4149c704e6151555e5b/research/quotient-split-minimal-odd-scale/CREDITS.md) | +| 52 | A separate basis evaluator confirms all 46 #7708 relay cuts and 1,501 public outputs, including the division by five. Uniform physical operations remain inherited. | [PR 7719](https://github.com/CrocSwap/integer-mult-bounds/pull/7719) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/6de456fff16721021227f9e5464cd9ba4a366348/research/relay-recoding-basis-audit/README.md) · [Lineage 7708, 7704](https://github.com/aliromman/integer-mult-bounds/blob/6de456fff16721021227f9e5464cd9ba4a366348/research/relay-recoding-basis-audit/CREDITS.md) | +| 53 | One relation in a cocharted BIT helper group cannot improve the signed bill on [0,0.01] with productive debits. Result 54 covers simultaneous relations. | [PR 7720](https://github.com/CrocSwap/integer-mult-bounds/pull/7720) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/3e56de8c160f259612f9766e06bf6990c7b5a3ef/research/bit-single-relation-bank-obstruction/README.md) · [Lineage 7001, 7304, 7573](https://github.com/aliromman/integer-mult-bounds/blob/3e56de8c160f259612f9766e06bf6990c7b5a3ef/research/bit-single-relation-bank-obstruction/CREDITS.md) | +| 54 | All subspaces of relations in one fixed cocharted BIT group fail the same two-orientation bill, including codimension 410. A changed paid source escapes this premise. | [PR 7722](https://github.com/CrocSwap/integer-mult-bounds/pull/7722) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/8a70046cafae35d29627017983d55cec71e06e27/research/bit-joint-relation-bank-obstruction/README.md) · [Lineage 7720, 7001](https://github.com/aliromman/integer-mult-bounds/blob/8a70046cafae35d29627017983d55cec71e06e27/research/bit-joint-relation-bank-obstruction/CREDITS.md) | +| 55 | For all original BIT arrivals one earlier original library and one nested send cannot lower rank. Results 64–66 expand the allowed deliveries. | [PR 7723](https://github.com/CrocSwap/integer-mult-bounds/pull/7723) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/a4fd22c08058c0fef830306b2d46e4a51f091939/research/bit-causal-delivery-rank-obstruction/README.md) · [Lineage 7720, 7001, 7293](https://github.com/aliromman/integer-mult-bounds/blob/a4fd22c08058c0fef830306b2d46e4a51f091939/research/bit-causal-delivery-rank-obstruction/CREDITS.md) | +| 57 | Fifty-four rational certificates check the nine restriction/split comparisons in #7726, including failed alternatives and full signed changes. The physical realization is inherited. | [PR 7729](https://github.com/CrocSwap/integer-mult-bounds/pull/7729) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/93483d1dbd8a8d7b7b199960b998f6daef6934e8/research/interleaved-basis-residual-audit/README.md) · [Lineage 7726, 7713, 7716](https://github.com/aliromman/integer-mult-bounds/blob/93483d1dbd8a8d7b7b199960b998f6daef6934e8/research/interleaved-basis-residual-audit/CREDITS.md) | +| 58 | The interleaved graph uses native scale 176,715 at binary height 21. Its scalar allowance falls from 2^30 to 2^26 with the same root and invoice. | [PR 7730](https://github.com/CrocSwap/integer-mult-bounds/pull/7730) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/a5a71d8f0b2ae30f72b8d7f616048127c06e8dc5/research/interleaved-minimal-odd-scale/README.md) · [Lineage 7726, 7718](https://github.com/aliromman/integer-mult-bounds/blob/a5a71d8f0b2ae30f72b8d7f616048127c06e8dc5/research/interleaved-minimal-odd-scale/CREDITS.md) | +| 64 | Every original BIT arrival retains its price under all finite direct bundles from strictly earlier original libraries. Internal relay networks change the hypothesis. | [PR 7741](https://github.com/CrocSwap/integer-mult-bounds/pull/7741) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/33817d65c801232f1e07355f43ec6040a41e7c7e/research/bit-all-bundle-price-floor/README.md) · [Lineage 7739, 7723, 7001](https://github.com/aliromman/integer-mult-bounds/blob/33817d65c801232f1e07355f43ec6040a41e7c7e/research/bit-all-bundle-price-floor/CREDITS.md) | +| 65 | For nine rank-seven deadlines the price 7^t persists with arbitrary paid relays on retained charts, 00 costs at least H4+H2+H1. A new exterior source can evade this boundary. | [PR 7790](https://github.com/CrocSwap/integer-mult-bounds/pull/7790) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/8cae9a6b7fed315f704144f27a89bf9b0ad9fd6e/research/binary-rooted-routing-optimum/README.md) · [Lineage 7786, 7781, 7776](https://github.com/aliromman/integer-mult-bounds/blob/8cae9a6b7fed315f704144f27a89bf9b0ad9fd6e/research/binary-rooted-routing-optimum/CREDITS.md) | +| 98 | The last #7786 bypass preserves 1,284 owners. Its packet scale and recoding are forced in the two-parameter family, and 901 retained banks cannot reconstruct it. | [PR 7791](https://github.com/CrocSwap/integer-mult-bounds/pull/7791) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/189c882b6aa3d9adac3e30fe228446141a7c94f9/research/p104-final-bypass-audit/README.md) · [Lineage 7786, 7781, 7776](https://github.com/aliromman/integer-mult-bounds/blob/189c882b6aa3d9adac3e30fe228446141a7c94f9/research/p104-final-bypass-audit/CREDITS.md) | +| 100 | One source on a strictly nested terminal chain pays exactly the sum over adjacent rank gaps, with arbitrary auxiliary charts, linear mixing and local storage. Another informative source changes the cut. | [PR 7793](https://github.com/CrocSwap/integer-mult-bounds/pull/7793) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/cd282ce24fb3eea707f4342cecd1935ed367a1a3/research/nested-broadcast-auxiliary-optimum/README.md) · [Lineage 7790, 7786, 7750](https://github.com/aliromman/integer-mult-bounds/blob/cd282ce24fb3eea707f4342cecd1935ed367a1a3/research/nested-broadcast-auxiliary-optimum/CREDITS.md) | + +## Call algebra, channel rank and projector words + +| Result | Established result, cost and boundary | Frozen record | +|---:|---|---| +| 4 | Two coordinate 3-cycles cost twelve for commuting rank-one projectors. The connected incidence census excludes the specified alternatives. Disconnected noncommuting words remain. | [PR 6504](https://github.com/CrocSwap/integer-mult-bounds/pull/6504) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/c98e5b03a7d9001a155ba10bc9faad0ceb7cad68/research/two-cycle-incidence-certificate/README.md) · [Lineage 6012, 2317](https://github.com/aliromman/integer-mult-bounds/blob/c98e5b03a7d9001a155ba10bc9faad0ceb7cad68/research/two-cycle-incidence-certificate/PROOF.md) | +| 20 | Two-role paired dyadic completion can require more correction rank than field completion, with an unbounded gap in the extended family. Arbitrary dyadic interfaces still need a denominator proof. | [PR 7675](https://github.com/CrocSwap/integer-mult-bounds/pull/7675) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/67d5652c917a1ba6086f8867512e0aa9d02492ca/research/dyadic-correction-rank-gap/README.md) · [Lineage 7654, 6016](https://github.com/aliromman/integer-mult-bounds/blob/67d5652c917a1ba6086f8867512e0aa9d02492ca/research/dyadic-correction-rank-gap/PROOF.md) | +| 21 | One fresh link on the two fixed binary blocks is irreducible exactly for the shared hub and odd fresh pairing. Its seven-dimensional action feeds result 31. | [PR 7676](https://github.com/CrocSwap/integer-mult-bounds/pull/7676) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/943c355c233649f77f018e6247e5456584ef715e/research/two-block-fresh-link-classification/README.md) · [Lineage 7648, 7237, 6921](https://github.com/aliromman/integer-mult-bounds/blob/943c355c233649f77f018e6247e5456584ef715e/research/two-block-fresh-link-classification/PROOF.md) | +| 22 | Kind A removal at every address width requires the same middle frame and a singular binomial. The retained finishing test then excludes the single-output combination. Joint outputs remain. | [PR 7677](https://github.com/CrocSwap/integer-mult-bounds/pull/7677) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/2bb938975cf34de19d1e1db820f4c9b42c486cb4/research/kind-a-removability-coset/README.md) · [Lineage 7629, 7649](https://github.com/aliromman/integer-mult-bounds/blob/2bb938975cf34de19d1e1db820f4c9b42c486cb4/research/kind-a-removability-coset/PROOF.md) | +| 23 | Kind B removal is equivalent to a frame exchange and coefficient ±1 in the specified other-array call. It does not supply a joint decoder. | [PR 7679](https://github.com/CrocSwap/integer-mult-bounds/pull/7679) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/7c3b229d0788d6e980f75cbaabe07287899263a2/research/kind-b-frame-swap-removal/README.md) · [Lineage 7629, 7649, 7677](https://github.com/aliromman/integer-mult-bounds/blob/7c3b229d0788d6e980f75cbaabe07287899263a2/research/kind-b-frame-swap-removal/PROOF.md) | +| 26 | One last call cannot finish a nonplain kind B part at any width or nonzero rational coefficient. Result 27 extends the test to common postprocessing. | [PR 7683](https://github.com/CrocSwap/integer-mult-bounds/pull/7683) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/9b99d7b30972a78e1ad0e90dd6ff3187bd3cc86c/research/kind-b-last-call-obstruction/README.md) · [Lineage 7629, 7679](https://github.com/aliromman/integer-mult-bounds/blob/9b99d7b30972a78e1ad0e90dd6ff3187bd3cc86c/research/kind-b-last-call-obstruction/PROOF.md) | +| 27 | A common free map on a single last output can be moved before the call under the stated leading-frame hypothesis. Two jointly decoded outputs remain outside the proof. | [PR 7684](https://github.com/CrocSwap/integer-mult-bounds/pull/7684) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/723b9b489d564411bde6b2c3eb0451d072b58276/research/last-call-common-postmap/README.md) · [Lineage 7683, 7629](https://github.com/aliromman/integer-mult-bounds/blob/723b9b489d564411bde6b2c3eb0451d072b58276/research/last-call-common-postmap/PROOF.md) | +| 28 | Independent images in an irreducible no-edge binary projector collection satisfy 2n≤3k−1. Image dependence matters for complete pools, as result 29 shows. | [PR 7686](https://github.com/CrocSwap/integer-mult-bounds/pull/7686) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/9da06cbe0b888293cbb249cb29eef0f37addaa4f/research/no-edge-image-basis-bound/README.md) · [Lineage 6921, 7237, 7281, 7648](https://github.com/aliromman/integer-mult-bounds/blob/9da06cbe0b888293cbb249cb29eef0f37addaa4f/research/no-edge-image-basis-bound/CREDITS.md) | +| 29 | For each n≥3 an irreducible complete n+1 projector pool has reducible image-basis subcollections. A basis-only search loses valid candidates. | [PR 7688](https://github.com/CrocSwap/integer-mult-bounds/pull/7688) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/d9aba79eaed566640353efcb0a27b0d5649f705a/research/no-edge-basis-extraction-failure/README.md) · [Lineage 7686, 6962](https://github.com/aliromman/integer-mult-bounds/blob/d9aba79eaed566640353efcb0a27b0d5649f705a/research/no-edge-basis-extraction-failure/CREDITS.md) | +| 30 | No-edge pattern dimension k permits at most floor(3k/2) joint image dimensions, with attaining families. Irreducible equality needs its own classification. | [PR 7690](https://github.com/CrocSwap/integer-mult-bounds/pull/7690) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/b725be51349cce4e0a262c76233838992042dd62/research/no-edge-total-image-bound/README.md) · [Lineage 6838, 7139, 7237, 7686](https://github.com/aliromman/integer-mult-bounds/blob/b725be51349cce4e0a262c76233838992042dd62/research/no-edge-total-image-bound/CREDITS.md) | +| 31 | At k=5 every irreducible seven-image no-edge complete pool has the two-triples-plus-link form. Realizing it as a paid word is an additional condition. | [PR 7691](https://github.com/CrocSwap/integer-mult-bounds/pull/7691) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/69e70b1426c32b19da8334930815800a2c95ffca/research/no-edge-dim5-extremal-classification/README.md) · [Lineage 7690, 6921, 7648](https://github.com/aliromman/integer-mult-bounds/blob/69e70b1426c32b19da8334930815800a2c95ffca/research/no-edge-dim5-extremal-classification/CREDITS.md) | +| 32 | Matching number ν gives the total bound 2n≤3k+ν and a strengthened irreducible bound. Rank bounds alone provide no profitable word. | [PR 7692](https://github.com/CrocSwap/integer-mult-bounds/pull/7692) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/c2870a38841f515db7203a0761cd77f014f7fb7a/research/edge-matching-image-bound/README.md) · [Lineage 7690, 7691, 6782](https://github.com/aliromman/integer-mult-bounds/blob/c2870a38841f515db7203a0761cd77f014f7fb7a/research/edge-matching-image-bound/CREDITS.md) | +| 33 | An irreducible image of size 2k−1 fixes the entire extremal noncoordinate edge pool. Result 34 tests the words on that structure. | [PR 7693](https://github.com/CrocSwap/integer-mult-bounds/pull/7693) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/dab13e010216570a00e22490b379b31b610f2890/research/edge-extremal-pool-classification/README.md) · [Lineage 7692, 7690, 7691](https://github.com/aliromman/integer-mult-bounds/blob/dab13e010216570a00e22490b379b31b610f2890/research/edge-extremal-pool-classification/CREDITS.md) | +| 34 | In that extremal pool, a diagonal monomial word with at most k saturated relations cannot gain productive rank. Changed relation budgets are open. | [PR 7694](https://github.com/CrocSwap/integer-mult-bounds/pull/7694) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/96b8b22f5925c4cf4514ed27575f465ebc95e51a/research/extremal-core-word-obstruction/README.md) · [Lineage 6782, 7693, 7692](https://github.com/aliromman/integer-mult-bounds/blob/96b8b22f5925c4cf4514ed27575f465ebc95e51a/research/extremal-core-word-obstruction/CREDITS.md) | +| 35 | An even orthogonal old basis plus one prescribed fresh bridge gives no saving under the stated incidences and relation budget. Results 45–46 examine extensions. | [PR 7695](https://github.com/CrocSwap/integer-mult-bounds/pull/7695) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/996c002d82420141c19797e4e3b699ed88aee330/research/orthogonal-basis-bridge-obstruction/README.md) · [Lineage 7694, 6782, 7691](https://github.com/aliromman/integer-mult-bounds/blob/996c002d82420141c19797e4e3b699ed88aee330/research/orthogonal-basis-bridge-obstruction/CREDITS.md) | +| 41 | Irreducible equality 2n=3k+ν−1 requires and attains k−ν=1 mod 4 in the stated binary class. Even differences need result 43. | [PR 7703](https://github.com/CrocSwap/integer-mult-bounds/pull/7703) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/2523029b27d40526b811591ce0fb85f9bc1b1df7/research/edge-matching-equality-classification/README.md) · [Lineage 7692, 7691, 6921](https://github.com/aliromman/integer-mult-bounds/blob/2523029b27d40526b811591ce0fb85f9bc1b1df7/research/edge-matching-equality-classification/CREDITS.md) | +| 42 | Over F2, every idempotent first factor has a classification and count of all admissible middle factors in the three-factor identity. Higher-rank implementation costs remain. | [PR 7706](https://github.com/CrocSwap/integer-mult-bounds/pull/7706) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/79ff656972d66915977e816b23ffdcef42fb0e12/research/idempotent-first-factor-kernel/README.md) · [Lineage 7223, 7040, 7027](https://github.com/aliromman/integer-mult-bounds/blob/79ff656972d66915977e816b23ffdcef42fb0e12/research/idempotent-first-factor-kernel/CREDITS.md) | +| 43 | For positive even k−ν the irreducible maximum is (3k+ν−2)/2 apart from (2,0). Classifying all equality pools and pricing their words are separate. | [PR 7707](https://github.com/CrocSwap/integer-mult-bounds/pull/7707) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/57e24923b609d2588b83321f530fa2a1e75ed364/research/edge-matching-even-attainment/README.md) · [Lineage 7692, 7703](https://github.com/aliromman/integer-mult-bounds/blob/57e24923b609d2588b83321f530fa2a1e75ed364/research/edge-matching-even-attainment/CREDITS.md) | +| 45 | An orthogonal old basis and the specified orthogonal family of b fresh links cannot save rank under the relation budget. The shared-hub exception needs result 46. | [PR 7710](https://github.com/CrocSwap/integer-mult-bounds/pull/7710) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/c396e9ee03b331f3b46aa87526817b237f5ad6f8/research/orthogonal-bridge-family-obstruction/README.md) · [Lineage 7707, 7695](https://github.com/aliromman/integer-mult-bounds/blob/c396e9ee03b331f3b46aa87526817b237f5ad6f8/research/orthogonal-bridge-family-obstruction/CREDITS.md) | +| 46 | For the exceptional six-projector pool, all six types and at most four saturated relations force a nonprofitable endpoint. Larger relation budgets remain. | [PR 7711](https://github.com/CrocSwap/integer-mult-bounds/pull/7711) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/ac185e4529a01156f7b574555b50827f8cb645bb/research/shared-hub-triple-word-obstruction/README.md) · [Lineage 7707, 7710](https://github.com/aliromman/integer-mult-bounds/blob/ac185e4529a01156f7b574555b50827f8cb645bb/research/shared-hub-triple-word-obstruction/CREDITS.md) | +| 68 | A dyadic Hamilton quaternion matrix factors through its rational quaternion rank by integral Hurwitz reduction. Coefficient-height and movement costs for a supplier remain. | [PR 7753](https://github.com/CrocSwap/integer-mult-bounds/pull/7753) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/a9634c95a49e2ac10318abc35cfd981fc7a8e375/research/dyadic-quaternion-rank-factorization/README.md) · [Lineage 7745, 5803](https://github.com/aliromman/integer-mult-bounds/blob/a9634c95a49e2ac10318abc35cfd981fc7a8e375/research/dyadic-quaternion-rank-factorization/CREDITS.md) | +| 69 | At width four, pattern dimension three and matching one, the complete pools split into 386 conjugacy classes, 40 irreducible. Their words need a relation and tariff test. | [PR 7754](https://github.com/CrocSwap/integer-mult-bounds/pull/7754) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/d7da273405eeba7e01bc2332146573a5476092fa/research/width-four-projector-classification/README.md) · [Lineage 7749, 7707](https://github.com/aliromman/integer-mult-bounds/blob/d7da273405eeba7e01bc2332146573a5476092fa/research/width-four-projector-classification/CREDITS.md) | +| 76 | Seven calls with one repeated type on the shared-hub pool still force two triple blocks and no rank benefit. Longer words need more information. | [PR 7763](https://github.com/CrocSwap/integer-mult-bounds/pull/7763) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/3cb09504f8c1b6b728d8ce90b76ff938c06fbe5d/research/shared-hub-seven-call/README.md) · [Lineage 7742, 7707, 7711](https://github.com/aliromman/integer-mult-bounds/blob/3cb09504f8c1b6b728d8ce90b76ff938c06fbe5d/research/shared-hub-seven-call/CREDITS.md) | +| 77 | At any length on the fixed four-projector alphabet, at most three saturated relations force the fresh totals to cancel. A fourth relation changes the class. | [PR 7764](https://github.com/CrocSwap/integer-mult-bounds/pull/7764) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/030ecb77c63a4bcc0fb5286cee9e42e856975da3/research/one-edge-word-obstruction/README.md) · [Lineage 7749, 7707, 7711](https://github.com/aliromman/integer-mult-bounds/blob/030ecb77c63a4bcc0fb5286cee9e42e856975da3/research/one-edge-word-obstruction/CREDITS.md) | + +## Reversible samplers + +| Result | Established result, cost and boundary | Frozen record | +|---:|---|---| +| 70 | A fixed reversible circuit implements all 144 nonzero denominator-13 parameter pairs in 14,022 gates and 2,146 clean bits. Tape access and its caller are separate. | [PR 7755](https://github.com/CrocSwap/integer-mult-bounds/pull/7755) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/d6babec41a3d01928c3dbdb896ea2adf051ba305/research/runtime-paired-ddg-arithmetic/README.md) · [Lineage 7746, 6072, 6102](https://github.com/aliromman/integer-mult-bounds/blob/d6babec41a3d01928c3dbdb896ea2adf051ba305/research/runtime-paired-ddg-arithmetic/CREDITS.md) | +| 71 | Exactly 2,304 of 2,704 denominator-52 pairs enter the nonzero core within four coins with adjusted phase and parameters. Result 74 handles the other tails. | [PR 7757](https://github.com/CrocSwap/integer-mult-bounds/pull/7757) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/5640f56b63ea04afc8505d45a4df1ea35466a94e/research/inverse52-phase-entry/README.md) · [Lineage 6096, 7755, 7746](https://github.com/aliromman/integer-mult-bounds/blob/5640f56b63ea04afc8505d45a4df1ea35466a94e/research/inverse52-phase-entry/CREDITS.md) | +| 74 | A 1,277-gate, 382-clean-bit circuit handles 288 mixed and 96 endpoint tails preserving ordered coins. Result 85 installs it at the finite-window boundary. | [PR 7760](https://github.com/CrocSwap/integer-mult-bounds/pull/7760) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/2a79ceb497c8b76a44cd3a9f706becdf9bbb8b7a/research/nonunit52-tail-circuit/README.md) · [Lineage 7746, 6096, 7757](https://github.com/aliromman/integer-mult-bounds/blob/2a79ceb497c8b76a44cd3a9f706becdf9bbb8b7a/research/nonunit52-tail-circuit/CREDITS.md) | +| 85 | One reversible circuit serves every positive finite window and all 2,704 residue pairs with exact label, depth or censoring. The caller must handle the censored branch. | [PR 7773](https://github.com/CrocSwap/integer-mult-bounds/pull/7773) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/c1fe18fbca499b50c3bbfb12e11218ee9b6acf01/research/inverse52-window-circuit/README.md) · [Lineage 7746, 6096, 7757, 7760](https://github.com/aliromman/integer-mult-bounds/blob/c1fe18fbca499b50c3bbfb12e11218ee9b6acf01/research/inverse52-window-circuit/CREDITS.md) | +| 91 | Checkpointing preserves the full finite-window distribution. At L=64 work drops from 3,874 to 1,375 clean bits and gates rise to 238,965. Unbounded continuation and coin delivery remain. | [PR 7782](https://github.com/CrocSwap/integer-mult-bounds/pull/7782) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/7e0d1aa1558583fc06e5a4f14f27e2bb4ffbf3a9/research/inverse52-block-checkpoints/README.md) · [Lineage 7773, 7746, 6096, 6072](https://github.com/aliromman/integer-mult-bounds/blob/7e0d1aa1558583fc06e5a4f14f27e2bb4ffbf3a9/research/inverse52-block-checkpoints/CREDITS.md) | + +## Root-of-unity capacity + +| Result | Established result, cost and boundary | Frozen record | +|---:|---|---| +| 75 | For prime q≥5, q+3 rows over 3q^b roots cannot have q+2 orthogonal columns. This removes the proposed eight-row input at q=5. | [PR 7762](https://github.com/CrocSwap/integer-mult-bounds/pull/7762) · [Evidence](https://github.com/aliromman/integer-mult-bounds/blob/d33fe8562c86888307cbba4a2aa92490212c8d78/research/mixed-cycle-column-obstruction/README.md) · [Lineage 7744, 5865](https://github.com/aliromman/integer-mult-bounds/blob/d33fe8562c86888307cbba4a2aa92490212c8d78/research/mixed-cycle-column-obstruction/CREDITS.md) | +| 80 | For q+3 rows the column count obeys m²