feat: sumcheck with disabled rows #10068
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,179 @@ | ||
| #pragma once | ||
| #include "barretenberg/common/compiler_hints.hpp" | ||
| #include "barretenberg/common/op_count.hpp" | ||
| #include "barretenberg/common/thread.hpp" | ||
| #include "barretenberg/stdlib/primitives/bool/bool.hpp" | ||
|
|
||
| #include <cstddef> | ||
| #include <vector> | ||
| namespace bb { | ||
| /** | ||
| * @struct RowDisablingPolynomial | ||
| * @brief Polynomial for Sumcheck with disabled Rows | ||
| * | ||
| * \f$ n = 2^d \f$ circuit size | ||
| * \f$ L_i \f$ multilinear Lagrange in \f$ d \f$ variables, \f$ i = 0,\ldots, n-1 \f$. | ||
| * | ||
| * Assume we are given a "valid" execution trace at rows \f$ 0,\ldots, n-5 \f$, i.e., | ||
| * \f[ | ||
| * \sum_{\mathbb{H} \setminus \{n-1, n-2, n-3, n-4\}} H = 0. | ||
| * \f] | ||
| * | ||
| * We want to pad the witness polynomials with random field elements in rows \f$ n-1, n-2, n-3 \f$. | ||
| * Since the commitment to the shift must coincide with the commitment to its unshifted counterpart, | ||
| * we have to reserve \f$ 4 \f$ rows at the end to be able to. | ||
| * To achieve this, we multiply the Honk relation \f$ H \f$ by the polynomial | ||
| * \f[ | ||
| * 1 - L = 1 - L_{n-1} - L_{n-2} - L_{n-3} - L_{n-4}. | ||
| * \f] | ||
| * that vanishes at the last \f$ 4 \f$ rows and is equal to \f$ 1 \f$ everywhere else on the hypercube. | ||
| * | ||
| * We consider the sumcheck protocol for the modified relation | ||
| * \f[ | ||
| * \sum_{\mathbb{H}} (1 - L) H = \sum_{\mathbb{H}} H - \sum_{\mathbb{H}} L \cdot H. | ||
| * \f] | ||
| * | ||
| * Note that the target sum remains \f$ 0 \f$ because the contributions from the last rows are multiplied by \f$ 0 \f$. | ||
| * | ||
| * Recall: | ||
| * - \f$ n-1 = 2^d - 1 = (1,1, \ldots, 1) \f$ | ||
| * - \f$ n-2 = (0,1,\ldots,1) \f$ | ||
| * - \f$ n-3 = (1,0,\ldots,1) \f$ | ||
| * - \f$ n-4 = (0,0,\ldots,1) \f$ | ||
| * | ||
| * ### Round 0: | ||
| * \f[ | ||
| * \begin{aligned} | ||
| * S' &= | ||
| * S_{H,0} - \Big(L_{n-1}(X, 1, \ldots, 1) + L_{n-2}(X, 1,\ldots,1)\Big) H(X,1,\ldots, 1) \\ | ||
| * &\quad - \Big(L_{n-3}(X, 0,1,\ldots,1) + L_{n-4}(X,0,1,\ldots,1)\Big) H(X,0,1,\ldots,1) | ||
| * \end{aligned} | ||
| * \f] | ||
| * | ||
| * We do not modify the algorithm computing \f$ S_{H,0} \f$. Simply add a method that computes the contribution from the | ||
| * edges \f$ (0,1,\ldots,1) \f$ and \f$ (1,\ldots,1) \in \mathbb{H}^{d-1} \f$. | ||
| * | ||
| * First, compute the coefficients in the Lagrange basis of the factor coming from the Lagranges: | ||
| * \f[ | ||
| * \begin{aligned} | ||
| * L_{n-1}(X,\vec{1}) + L_{n-2}(X,\vec{1}) &= X + (1 - X) = 1 \\ | ||
| * L_{n-3}(X,0,1,\ldots,1) + L_{n-4}(X,0,1,\ldots,1) &= 1 | ||
| * \end{aligned} | ||
| * \f] | ||
| * | ||
| * \f[ | ||
| * S'_0 = S_{H,0} - H(X,1,\ldots,1) - H(X,0,1,\ldots,1) | ||
| * \f] | ||
| * | ||
| * ### Round 1: | ||
| * \f[ | ||
| * \begin{aligned} | ||
| * L_{n-1}(u_0,X,\vec{1}) + L_{n-2}(u_0,X,\vec{1}) &= | ||
| * u_0 X + (1 - u_0) X = X \\ | ||
| * L_{n-3}(u_0,X,\vec{1}) + L_{n-4}(u_0,X,\vec{1}) &= | ||
| * u_0 (1 - X) + (1 - u_0)(1 - X) = (1 - X) | ||
| * \end{aligned} | ||
| * \f] | ||
| * | ||
| * \f[ | ||
| * S'_1 = S_{H,1} - H(X,1,\ldots,1) | ||
| * \f] | ||
| * | ||
| * ### Round 2: | ||
| * \f[ | ||
| * S'_2 = S_{H,2} - X \cdot H(u_0,u_1,X,1,\ldots,1) | ||
| * \f] | ||
| * | ||
| * ### Rounds i > 1: | ||
| * We can compute the restricted sumcheck univariates \f$ S' \f$ in each round as follows: | ||
| * \f[ | ||
| * \begin{aligned} | ||
| * S' = S_{H,i} - | ||
| * \Big(L_{n-1}(u_0, \ldots, u_{i-1}, X, 1, \ldots, 1) + L_{n-2}(u_0, \ldots, u_{i-1}, X, 1,\ldots,1) \\ | ||
| * + L_{n-3}(u_0, \ldots, u_{i-1}, X, 1,\ldots,1) + L_{n-4}(u_0, \ldots, u_{i-1}, X, 1,\ldots,1)\Big) \\ | ||
| * \times H(u_0, \ldots, u_{i-1}, X, 1,\ldots,1) | ||
| * \end{aligned} | ||
| * \f] | ||
| * | ||
| * Compute the factor coming from the Lagranges: | ||
| * \f[ | ||
| * \begin{aligned} | ||
| * &\left(u_0 \cdots u_{i-1} + (1 - u_0) u_1 \cdots u_{i-1} + u_0 (1 - u_1) u_2 \cdots u_{i-1} + (1 - u_0)(1 - u_1) u_2 | ||
| * \cdots u_{i-1}\right) X \\ | ||
| * &= \left(u_0 u_1 + (1 - u_0) u_1 + u_0 (1 - u_1) + (1 - u_0)(1 - u_1)\right) u_2 \cdots u_{i-1} X \\ | ||
| * &= 0 \cdot (1 - X) + 1 \cdot u_2 \cdots u_{i-1} \cdot X | ||
| * \end{aligned} | ||
| * \f] | ||
| * | ||
| * This way, we get: | ||
| * \f[ | ||
| * S_{H,i}(X) = S_{H,i} - u_2 \cdots u_{i-1} \cdot X \cdot H(u_0, \ldots, u_{i-1}, X, 1, \ldots, 1). | ||
| * \f] | ||
| * | ||
| * ## The algorithm: | ||
| * | ||
| * Let \f$ D \f$ be the max partial degree of \f$ H \f$. | ||
| * | ||
| * 1. Compute \f$ S_{H,i} \f$ without any modifications as a polynomial of degree \f$ D \f$. Extend it to degree \f$ D + | ||
| * 1 \f$, because it is the max partial degree of \f$ L \cdot H \f$. | ||
| * | ||
| * 2. If \f$ i = 0 \f$, compute \f$ H(X,1,1,\ldots,1) + H(X,0,1,\ldots,1) \f$ as a univariate of degree \f$ D \f$, else | ||
| * compute \f$ H(u_0, \ldots, u_{i-1}, X, 1, \ldots, 1) \f$ as a univariate of degree \f$ D \f$. Extend to degree \f$ D | ||
| * + 1 \f$. | ||
| * | ||
| * 3. Compute the extension of \f$ L^{(i)} = L(u_0, \ldots, u_{i-1}, X, 1, \ldots, 1) \f$ to the degree \f$ D + 1 \f$ | ||
| * polynomial. | ||
| * | ||
| * 4. Compute the coefficients of the product \f$ L^{(i)} \cdot H^{(i)} \f$. | ||
| * | ||
| * 5. Compute the coefficients of \f$ S_{H,i} - L^{(i)} \cdot H^{(i)} \f$ (degree \f$ D + 1 \f$ univariate). | ||
| * | ||
| * The verifier needs to evaluate \f$ 1 - L(u_0, \ldots, u_{d-1}) \f$, which is equal to \f$ 0 \f$ if \f$ d < 2 \f$, and | ||
| * is equal to \f$ 1- u_2 \cdots u_{d-1} \f$ otherwise. | ||
| */ | ||
|
|
||
| template <typename FF> struct RowDisablingPolynomial { | ||
| // initialized as a constant linear polynomial = 1 | ||
| FF eval_at_0{ 1 }; | ||
| FF eval_at_1{ 1 }; | ||
|
|
||
| RowDisablingPolynomial() = default; | ||
| /** | ||
| * @brief Compute the evaluations of L^{(i)} at 0 and 1. | ||
| * | ||
| * @details In every round, the contribution from the Honk relation computed at | ||
| * disabled rows has to be mutiplied by \f$ L^{(i)} \f$, which is a linear combination of Lagrange polynomials | ||
| * defined above. | ||
| * | ||
| * @param round_challenge Sumcheck round challenge | ||
| * @param round_idx Sumcheck round index | ||
| */ | ||
| void update_evaluations(FF round_challenge, size_t round_idx) | ||
|
There was a problem hiding this comment. add function docs please for this and below |
||
| { | ||
| if (round_idx == 1) { | ||
| eval_at_0 = FF{ 0 }; | ||
| } | ||
| if (round_idx >= 2) { | ||
| eval_at_1 *= round_challenge; | ||
| } | ||
| } | ||
| /** | ||
| * @brief Compute the evaluation of \f$ 1 - L \f$ at the sumcheck challenge | ||
| * | ||
| * @param multivariate_challenge | ||
| * @param log_circuit_size | ||
| * @return FF | ||
| */ | ||
| static FF evaluate_at_challenge(std::vector<FF> multivariate_challenge, const size_t log_circuit_size) | ||
| { | ||
| FF evaluation_at_multivariate_challenge{ 1 }; | ||
|
|
||
| for (size_t idx = 2; idx < log_circuit_size; idx++) { | ||
| evaluation_at_multivariate_challenge *= multivariate_challenge[idx]; | ||
| } | ||
|
|
||
| return FF{ 1 } - evaluation_at_multivariate_challenge; | ||
| } | ||
| }; | ||
|
|
||
| } // namespace bb | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it's like that someone modifying the
table_offsetis not going to be exactly aware of this additional MASKING_OFFSET and all the places it has to be modified so I wonder if subtractions can be unified a bit.