Ultraplonk check_circuit #366

Rumata888 · 2023-04-19T12:57:39Z

Implements the ultra_circuit_constructor check_circuit method

The check_circuit method allows to check ultra circuits without using the proving system. The method tests:

Gate identities
Lookups
Permutations (sets, too!!)
The check involves computation of all finalized gates, but they are discarded right after the check. This should make iteratively debugging nasty ultra circuits much easier.

One of the major changes is the addition of the "in-the-head" circuit structure. It is being used in all methods that are called from check_circuit. The information about gate and witness changes is put there instead of the body of circuit_constructor, which allows us to discard this updated state. For this we use switched_* mechanic, where the location of actual selector and witness data is decided on the go.

One of the changes is that plookup tables have been moved to proof_system. The logic is that linking plonk to proof system caused gcc link issues and it's nature for the tables to be in the same place as the circuits.

Checklist:

codygunton · 2023-04-25T17:49:07Z

cpp/src/barretenberg/plonk/composer/ultra_composer.cpp

@@ -1845,6 +1845,7 @@ std::array<uint32_t, 2> UltraComposer::decompose_non_native_field_double_width_l
    const uint256_t value = get_variable(limb_idx);
    const uint256_t low = value & LIMB_MASK;
    const uint256_t hi = value >> DEFAULT_NON_NATIVE_FIELD_LIMB_BITS;
+    // WTF(kesha): What is this supposed to do? Unless uint256_t has failed, this should always work


Ha, harder to find this if it doesn't contain the word TODO somewhere...

codygunton · 2023-04-25T17:51:05Z

...rc/barretenberg/plonk/composer/splitting_tmp/composer_helper/ultra_plonk_composer_helper.cpp

@@ -50,7 +50,7 @@ void UltraPlonkComposerHelper<CircuitConstructor>::compute_witness(CircuitConstr

    // TODO(luke): subgroup size was already computed above but compute_witness_base computes it again. If we pass in
    // NUM_RANDOMIZED_GATES (as in the other split composers) the resulting sizes can differ. Reconcile this.
-    auto wire_polynomial_evaluations = compute_witness_base(circuit_constructor, total_num_gates, NUM_RANDOMIZED_GATES);
+    auto wire_polynomial_evaluations = compute_witness_base(circuit_constructor, total_num_gates, NUM_RESERVED_GATES);


Corresponding change in the definition of compute_witness_base? If it's really should be called NUM_RESERVED_GATES then we should change the name of the function argument, no? Your goal is just uniformity of naming, right, there's no difference?

But below we have num_randomized_gates = NUM_RESERVED_GATES. Are we just using the fact that presently we have NUM_RESERVED_GATES==NUM_RANDOMIZED_GATES? Both of these could be specified statically in the flavor, so it's probably best to keep the roles distinct.

codygunton

This is really impressive work and a huge win for us. I have some concerns about the complexity of its implementation though. Our code is already pretty complex (for newcomers, but even for people on adjacent teams), and I fear that this sets a concerning precedent (embeding a sort of sub-constructor in each constructor resulting in duplication; using hard-to-read macros), so I'm asking you to spend a bit of time thinking about potential simplifications and/or ways to reuse code, maybe avoid calling everything switched (also not a very intuitive term IMO), and possibly renaming some some of what you have done here becuase I really think this in_the_head motif is not intuitive and will trip many people up. Let's discuss more tomorrow and find a way to get this in soon.

codygunton · 2023-04-25T18:14:02Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.hpp

+        }
+    };
+    /**
+     * @brief Circuit-in-the-head is a structure we use to store all the information about the circuit which should be


I find this name pretty unintuitive and I suspect others will as well. You said you chose this name as an alternative to using the word "virtual" since C++ has a the keyword virtual. What do you think about something like TemporaryConstructor or ScratchConstructor?

Or, is it not possible to reduce duplication by doign something like this?

Derive UltraCircuitConstructorData : public CircuitConstructorBase<arithmetization::Ultra>

Derive UltraCircuitConstructor : public UltraCircuitConstructorData

Make a member UltraCircuitConstructorData scratch_data in UltraCircuitConstructor

Do you not like this suggestion? Feels cleaner to me, what do you think?

codygunton · 2023-04-25T18:48:07Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.hpp

+     * @brief Circuit-in-the-head is a structure we use to store all the information about the circuit which should be
+     * used during check_circuit method call, but needs to be discarded later
+     *
+     * @details In check_circuit method in UltraCircuitConstructor we want to check that the whole circuit works, but


Would you add another sentence like explaining further how we accomplish the goal? Maybe something like: 'This class lets us pause circuit construction and temporarily generate only the finalization gates needed for a partial circuit check. When the check is complete, we discard this data before continuing with main thread of circuit construction.'

codygunton · 2023-04-25T18:53:52Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.hpp

+            return stored_state;
+        }
+        /**
+         * @brief CHecks that the circuit states is the same as the stored circuit's one


codygunton · 2023-04-25T18:59:19Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.hpp

+         * @return true
+         * @return false
+         */
+        template <typename CircuitConstructor> bool is_same_state(const CircuitConstructor& circuit_constructor)


Why is this a template?

codygunton · 2023-04-25T19:08:31Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.hpp

+        }
+    }
+
+#define PARENS ()


Oh god, why is this all done with macros?

This is pretty hard to read. I'm spending a while trying to figure out how and where we specify that we're constructing 'in the head' or not.

codygunton · 2023-04-25T19:38:51Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.cpp

+ */
+void UltraCircuitConstructor::update_circuit_in_the_head()
+{
+    // Unfortunately we have to copy all variable structures


I guess this is for simplicity because you just want one global boolean switch?

codygunton · 2023-04-25T19:40:39Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.cpp

+    bool result = true;
+    // Put the circuit in the head
+    update_circuit_in_the_head();
+    in_the_head = true;


Do I have it right that this is the only place where we switch to in-the-head mode?

codygunton · 2023-04-25T19:43:55Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.hpp

+#define CHOOSE_VIRTUAL_OR_REAL_MULTIPLE(variable_prefix, switch_name, member_name, ...)                                \
+    FOR_EACH(ASSIGN_VARIABLE_TO_VIRTUAL_OR_REAL, variable_prefix, switch_name, member_name, __VA_ARGS__)
+#define ENABLE_ALL_IN_THE_HEAD_SWITCHES                                                                                \
+    CHOOSE_VIRTUAL_OR_REAL_MULTIPLE(switched_,                                                                         \


Like what does this do? Why is this called MULTIPLE?

codygunton · 2023-04-25T19:53:38Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.cpp

@@ -2368,4 +2443,943 @@ void UltraCircuitConstructor::process_RAM_arrays(const size_t gate_offset_from_p
    }
 }

+/**
+ * @brief Various methods relating to circuit evaluation


I think this will eventually confuse doxygen, can you write the comment withou tags? This is just like a big divider, right?

codygunton · 2023-04-25T20:02:26Z

cpp/src/barretenberg/proof_system/circuit_constructors/ultra_circuit_constructor.cpp

+ */
+
+/**
+ * @brief Arithmetic gate-related methods


I think we need to have some tests that connect the check circuit functions that verify they are correct. Am I missing that this is already implicit somwhere? I think presently we are relying on someone thorougly auditing the code below and making sure to keep it updated with any changes that may happen (hopefully don't happen, but may) with how circuits are Ultra constructed. Do you have any thoughts about how/when to do that?

Namely, for each type of gate, there should be a unit test of that gate creation function that runs check_circuit and checks that returns true if and only if the circuit assignment is valid. The hope is that this is sufficient, though I feel we need an argument that composition is then valid, and we probably need to mix in funny things like genperm range constraints and assert_equals. And of course we should still test some higher-level circuits thoroughly in the full construction-verification flow.

codygunton · 2023-04-27T18:19:31Z

This is awesome, thanks.

* Added dynamic array abstraction into standard library (AztecProtocol#112) Implements RAM/ROM stuff and dynamic arrays as well as separated all fixed_base operation in standard plonk into a separate file, so that it is no longer part of composer * fix: Store lagrange forms of selector polys w/ Ultra (AztecProtocol#255) * store lagrange forms of selector polynomials when serializing pk for Ultra * added comment to ultra_selector_properties * feat(ts): allow passing srs via env functions (AztecProtocol#260) * feat(ts): switch to node-modules linker * feat(ts): add new env for SRS objects * feat(ts): test srs bindings * fix: proper uint8_t include * feat(ts): revert unneeded changes * feat(ts): revert unneeded changes * feat(ts): unify writeMemory arg order * Update barretenberg_wasm.ts * feat(ts): fix srs comments * Update data_store.hpp --------- Co-authored-by: Adam Domurad <adam@aztecprotocol.com> * Lde/transcript (AztecProtocol#248) * adding adrians new transcript classes * tests added for transcript and new manifest concept --------- Co-authored-by: codygunton <codygunton@gmail.com> * fix(build): git add -f .yalc (AztecProtocol#265) * feat(ts): switch to node-modules linker * feat(ts): add new env for SRS objects * feat(ts): test srs bindings * fix: proper uint8_t include * feat(ts): revert unneeded changes * feat(ts): revert unneeded changes * feat(ts): unify writeMemory arg order * Update barretenberg_wasm.ts * feat(ts): fix srs comments * Fix deps * Fix comments * fix(build): git add -f .yalc * Merge --------- Co-authored-by: Adam Domurad <adam@aztecprotocol.com> * chore: modularize bb (AztecProtocol#271) * chore: modularize ts * chore: reformat * Adding foundation to bb.js (AztecProtocol#274) * cleaning up bb.js deps * update bb structure to use workspaces * remove foundation .yarn/cache * chore: don't bundle .yalc * Update readme * chore: modularize bb (AztecProtocol#271) * chore: modularize ts * chore: reformat * merge * remove yalc * Unbundle tsbuildinfo --------- Co-authored-by: ludamad <domuradical@gmail.com> Co-authored-by: ludamad <adam.domurad@gmail.com> * Fix build of ts package (AztecProtocol#276) * Splitting turbo composer (AztecProtocol#266) * Turbo Circuit Constructor working * Turbo!! And also fixed some of the fuzzer compilation issues * Luke: Addressing my own comments and adding minor TODOs where necessary --------- Co-authored-by: ledwards2225 <l.edwards.d@gmail.com> * Cg/move to shared (AztecProtocol#294) * Move circuit constructors to shared. * Move helper lib and perm helper. * Move tmp composers and helpers for plonk. * Fix namespace and red herring comment. * Remove pointless namespace declaration. * Fix more namespaces. * Split flavor * Rename tests to avoid ambiguity. * Remove redundant macro defs. * Fix comment formatting. * StandardArithmetization is not shared with plonk. * Lde/split gemini (AztecProtocol#256) * adding adrians new transcript classes * building with some failing tests * tests passing * tests added for transcript and new manifest concept * improvements to the manifest concept * prover now operating on split gemini fuctionality * make shplonk test independent of Gemini * gemini and kzg tests updated; reduce prove removed from gemini * general cleanup * woops, fix gcc build * minor rebase fix * make gemini method return fold polys per Adrians suggestion * fix bad move * Lde/lookup grand product (AztecProtocol#286) * moving perm grand product to prover lib and fleshing out lookup grand product * cleaning up perm grand product test * lookup grand product test in place * cleaning up lookup grand prod test and adding sorted list accum method and test * rename prover tests to prover library tests * general cleanup * improve naming for gamma and beta constants * rabse fix * Cg/arithmetization (AztecProtocol#296) * Move gate data to better location. * Add basic arithmetization class. * CircuitConstructor takes Arithmetization. * Remove FooSelector enums from split composers. * feat: Working UltraPlonk for Noir (AztecProtocol#299) * Make dsl composer agnostic. * change SYSTEM_COMPOSER under stdlib::types to ultra composer type * use ultra logic constraints * in process of debugging, move to using ultra logic constraints * add get_total_circuit_size method * acir format tests showing failures with range constraints of different bit sizes * remove unnecessary comment * (fix) Temporarily add a redundant add-gate for variables that need range constraint < 8 bits. * rename functions * Implement get_solidity_verifier function * Fix no longer available properties * remove constraint system * logic gate changes using plookup * logic gate debugging * test for logic gates passing * last debug things XOR and AND returnign correct results, XOR still failing * cleanup * pedersen_plookup * plookup funcs * add to header * fixed error in pedersen hash when RHS is a circuit constant * added ACIR test for XOR gate pedersen hash test now checks y coordinate * temp disable wasm-opt * Making everything compile with any composer & add a cmake flag to switch on turbo * enable wasm-opt for asyncify but disable optimizations * remove using in header * fixed work queue bug with wasm wasm code path was not correctly storing fft outputs in proving key * added bitwise logic operations into stdlib stdlib method is utility method to provide Composer-agnostic interface due to the use of plookup tables if enabled * updated acir_format to use new stdlib logic class Updated ReadMe to include wasm example that supports gtest filtering * reenable tests * linting fixes * disable binaryen with comment * write instead of read * remove random * WIP * cleanup the debug logging * restore the randomness * only add a zero/one test instead of replacing * remove unused change * changes to make solgen work correctly in bindings * fix join_split_tests.test_deposit_construct_proof * working serialized proving key size and circuit change test for ultra (AztecProtocol#307) * USE_TURBO for join_split * Empty-Commit * Don't default one function; tweak comments. * Empty-Commit --------- Co-authored-by: Suyash Bagad <suyashnbagad1997@gmail.com> Co-authored-by: vezenovm <mvezenov@gmail.com> Co-authored-by: Maxim Vezenov <maximvezenov@Maxims-MacBook-Pro.local> Co-authored-by: zac-williamson <blorktronics@gmail.com> Co-authored-by: kevaundray <kevtheappdev@gmail.com> Co-authored-by: codygunton <codygunton@gmail.com> * Add debugging CMake preset & update code-workspace (AztecProtocol#308) * Add debugging CMake preset & update code-workspace --------- Co-authored-by: Blaine Bublitz <blaine.bublitz@gmail.com> * Lde/ultra composer (AztecProtocol#302) * duplicate ultra composer with tests passing * instantiating a circuit constructor in composer but not using it yet * directory updates after rebase plus finalize circuit function added * WiP almost have composer helper proving key computation building * WiP still debugging linker error * linker issue seemingly resolved * create prover building and running with new composer * proving key polys match old composer for simple circuit * circuit with no lookups is verifying * all composer tests passing with split ultra composer * kill poly store debug code * cleanup * fix arithmetization rebase issues * WiP new test * fix bad circuit size bug * cleanup * fix(nix): Use wasi-sdk 12 to provide barretenberg-wasm in overlay (AztecProtocol#315) * fix(nix): Use wasi-sdk 12 to provide barretenberg-wasm in overlay * chore: Remove the wasm stuff from main package * chore(nix): Switch the default llvm to 11 * chore(nix): Add transcript00 to the overlay chore(nix): Cleanup for nix flake check * Use hash for each platform * avoid symlinks * try wasi-sdk that someone wrote on github * fix hash for linux * try to ignore libstdc++ * need the whole name * try to include std lib instead of ignore * cleanup and nix flake check * chore(ci): Check the nix flake in CI * run default build instead of llvm12 * Prep: move composer type, proving key and verification key. (AztecProtocol#303) * Move composer type from plonk to bonk. * Move pk & vk into plonk. * bonk ~> proof_system; nest plonk and honk in it. * proof_system independent of plonk. * fix(dsl): Use info instead of std::cout to log (AztecProtocol#323) * fix(dsl): Use info instead of std::cout to log * Empty-Commit --------- Co-authored-by: Maxim Vezenov <mvezenov@gmail.com> * fix(nix): Disable ASM & ADX when building in Nix (AztecProtocol#327) * fix(nix): Disable ASM & ADX when building in Nix * Empty-Commit --------- Co-authored-by: kevaundray <kevtheappdev@gmail.com> * Aztec3 Specific Work in Barretenberg (AztecProtocol#142) * Split Pedersen Hash & Commitment Gadgets (AztecProtocol#95) * [SQUASHED] Pedersen refactor into hash and commitment. Use lookup pedersen for merkle tree, fixed-base pedersen for commitments. --------- Co-authored-by: Suyash Bagad <suyashnbagad1997@gmail.com> Port `copy_as_new_witness`. Port `must_imply`. `operator++`. Port changes from `common`. Port `ecc/groups`. * [CPM] add missing dependencies to libbarretenberg.a (AztecProtocol#154) --------- * Increase Pedersen Generator indices and subindices. (AztecProtocol#169) * Remove a3 specific types. (AztecProtocol#252) * Address Luke's Comments on `aztec3 -> master` (AztecProtocol#263) * Add must_imply tests. * Added a test for `field_t::copy_as_new_witness` * add test for `conditional_assign` * Added `infinity` test. * Add `add_affine_test`. * Tests for Array Object in `stdlib` (AztecProtocol#262) * basic array tests. * Add `composer_type` while hashing/compressing a vkey. * Add `contains_recursive_proof` to Recursive VK (AztecProtocol#268) * feat: debug utility for serialization (AztecProtocol#290) * feat: enable asan config * `array_push` for Generic Type (AztecProtocol#291) * Add Indexed Merkle Tree (AztecProtocol#281) * remove ts (consulted with Adam and we're good to go). (AztecProtocol#292) * Add cout for verification_key struct (AztecProtocol#295) * compute tree (AztecProtocol#298) * [SQUASHED] fixing `push_array_to_array` method. (AztecProtocol#304) * feat(memory_tree|a3): add sibling path calculations (AztecProtocol#301) * feat(memory_tree): frontier paths * fix array and resolve merge conflicts (AztecProtocol#305) * Mc/hash vk (AztecProtocol#306) * Increase number of sub-generators to 128. * Build a3crypto.wasm (AztecProtocol#311) * More Tests on A3 `stdlib` methods (AztecProtocol#316) * test: more vk tests to compare circuit/native/vk_data (AztecProtocol#310) * Mc/hash vk (AztecProtocol#306) * inc num_generators_per_hash_index to 128. (AztecProtocol#309) * fix. (AztecProtocol#318) * Added test for `compute_tree_native`. (AztecProtocol#319) * Install instructions for apt on ubuntu (AztecProtocol#312) * Fix address compilation. (AztecProtocol#329) --------- Co-authored-by: David Banks <47112877+dbanks12@users.noreply.github.com> Co-authored-by: Michael Connor <mike@aztecprotocol.com> Co-authored-by: dbanks12 <david@aztecprotocol.com> Co-authored-by: Santiago Palladino <spalladino@gmail.com> Co-authored-by: ludamad <adam.domurad@gmail.com> Co-authored-by: Maddiaa <47148561+cheethas@users.noreply.github.com> Co-authored-by: Santiago Palladino <santiago@aztecprotocol.com> Co-authored-by: ludamad <domuradical@gmail.com> Co-authored-by: cheethas <urmasurda@gmail.com> * Split shplonk in prep for work queue (AztecProtocol#321) * Consolidate permutation mapping computation into one method (AztecProtocol#330) * Lde/reinstate work queue (AztecProtocol#324) * make MSM size in work queue more flexible * new work queue hooked up everywhere excluding shplonk * improve interface and remove commitment key from prover * move old work queue to plonk namespace * fix(cmake): Remove leveldb dependency that was accidentally re-added (AztecProtocol#335) * fix(cmake): Remove leveldb dep d that was accidentally re-added * Empty-Commit --------- Co-authored-by: kevaundray <kevtheappdev@gmail.com> * change to get_num_gates() inside get_total_circuit_size() (AztecProtocol#332) * Mm/ensure all stdlib_primitives_tests are run using all four composers * UltraHonk Composer (Split) (AztecProtocol#339) * add split UltraHonk composer and checks for consistency with UltraPlonk * adding issue number to some TODOs * fix: Revert generator changes that cause memory OOB access (AztecProtocol#338) * fix: Revert generator changes that cause memory OOB access * Empty-Commit * Fix cci (temporarily). * comment out one more test. --------- Co-authored-by: kevaundray <kevtheappdev@gmail.com> Co-authored-by: Suyash Bagad <suyashnbagad1997@gmail.com> * doc: Document more thoroughly why fields don't 0-init (AztecProtocol#349) * Update field.hpp * Update field.hpp * Update field.hpp * 32-Byte Keccak256 challenges for UltraPlonK (AztecProtocol#350) * Add WithKeccak variants. * Update SYSTEM_COMPOSER dependents. * Ultra Honk arithmetic and grand product relations (AztecProtocol#351) * add width 4 perm grand prod construction relation * make grand prod construction use id polys, relation correctness passing * reorganize relation testing suites * primary ultra arithmetic relation with passing tests * secondary arith relation and grand prod init relation plus tests * add modified consistency check for selectors (AztecProtocol#354) * No `SYSTEM_COMPOSER` (AztecProtocol#352) * Get rid of system composer. * Remove USE_TURBO * Lde/lookup grand prod relation (AztecProtocol#359) * lookup grand product relation tests passing * ignore final entry in table polys for consistency chec, same reason as for selectors * adding eta and lookup gp delta to relation params * incorporate lookup gp init relation into tests * correcting the degree of lookup relation * fixed bug where range constraining connected witnesses threw an error (AztecProtocol#369) * fixed bug where range constraining connected witnesses threw an error * can now apply multiple overlapping ranges to the same witness (or copies of the witness) --------- Co-authored-by: codygunton <codygunton@gmail.com> * Add Mutex to global initialisation of generators (AztecProtocol#371) * Mutex lock initialization call * add comment on mutex --------- Co-authored-by: kevaundray <kevtheappdev@gmail.com> * chore: Add cachix action for uploading binary cache artifacts (AztecProtocol#373) * chore: Add cachix action for uploading binary cache artifacts * Only run the nix action on master * run on my branch and remove todos * Remove running on my branch * Zw/recursion constraint reduction (AztecProtocol#377) * removed blake3s hash from ultraplonk recursive prover * UltraComposer will now not create duplicate non-native field multiplication constraints * Propagate new stuff to Honk and splitting_tmp. * Clean up and add comments. --------- Co-authored-by: codygunton <codygunton@gmail.com> * add constraints to ensure that the result of the logic constraint is indeed produced from the left and right operands of the functions and improve testings. * fixed error where applying copy constraints to range-constrained indices would create an unbalanced set membership check * improve documentation and fix issue in non-ultra variants of the composer * [SQUASHED] ecdsa key recovery with test(s). (AztecProtocol#346) init commit. Recover pubkey works! Make gcc happy. Make gcc happy (again) gcc fix. don't use y = 0 as error condition. instead use (x = 0, y = 0) as failure return. * feat(nullifier_tree): make empty nullifier tree leaves hash be 0 (AztecProtocol#360) (AztecProtocol#382) * feat(nullifier_tree): make empty nullifier tree leaves be 0 * fix: add append zero behaviour and test * fix: explicit type name * clean: update class semantics --------- Co-authored-by: Maddiaa <47148561+cheethas@users.noreply.github.com> Co-authored-by: cheethas <urmasurda@gmail.com> Co-authored-by: cheethas <addaboy@gmail.com> * Solidity Ultra Verifier with tests (AztecProtocol#363) * initial setup * feat: add test setup * fix: revert unneeded cmakelist changes + typos * fix: solidity helper binaries in docker * chore: Use `-assert` variant instead of plain * chore: config.yml missing whitespace * fix: alpine base dockerfile * chore: add sol to build_manifest * Dockerfile: add string utils * fix: use different base, add bash * chore: remove stale misc * circle-ci fiddling * more circle fiddling * fiddle * Circle-ci fiddling for verifiers (AztecProtocol#365) * build_manifest update * fiddling * fiddling * skip ensure_repo * skipping more stuff * fiddling * get docker version * force change in key_gen * fiddle * docker version in cond_spot * fiddle * update path * fiddle * build in init * fiddle * fiddle * fiddle * fiddle * alpine * package naming * add apt-repository * add apt rep key * lld-15 * fiddle with docker * chore: cleanups * chore: remove log --------- Co-authored-by: cheethas <addaboy@gmail.com> * fix: throw -> throw_or_abort in sol gen (AztecProtocol#388) * fix: throw -> throw_or_abort in sol gen * toggle nix build * fix: toggle nix build * chore: revert toggle --------- Co-authored-by: cheethas <addaboy@gmail.com> * feat!: replace `MerkleMembershipConstraint` with`ComputeMerkleRootConstraint` (AztecProtocol#385) * feat: replace `MerkleMembershipConstraint` with`ComputeMerkleRootConstraint` * Update acir_format.cpp * Ultraplonk check_circuit (AztecProtocol#366) * Add check_circuit with mid-construction introspection * updating IPA to use transcript (AztecProtocol#367) * hack: introduce BARRETENBERG_CRYPTO_GENERATOR_PARAMETERS_HACK (AztecProtocol#409) * hack: introduce BARRETENBERG_CRYPTO_GENERATOR_PARAMETERS_HACK * doc: concise * Update generator_data.cpp * Update generator_data.cpp * Add Keccak constraints to acir_format (AztecProtocol#393) * exp: fix alpine versioning (AztecProtocol#415) Co-authored-by: cheethas <addaboy@gmail.com> * Small change that was left out of construct_addition_chains fix (AztecProtocol#404) Fixes the intermittent construct_addition_chains bugs (the previous fix was incomplete) and cleans the test up a bit * Pending bb work for aztec3 (AztecProtocol#368) * update js vk (because we now use UP for merkle hashing) * Helpers for ECDSA in A3 (AztecProtocol#364) * Add `stdlib_keccak` in cmake. Correct an assertion in `to_byte_array` in bigfield. * Add `random_element` to affine element. * negate y conditionally. * Change pedersen hash c_bind to use `pedersen_hash::lookup`. * c_binds and other ECDSA related fixes (AztecProtocol#407) * Add v to stdlib ecdsa. * create an engine if its empty. * Add ecdsa c_bind. * print v as a uint32. * Add secp256k1 cbind. add c_bind.hpp Change hpp to h. remove hpp. * Add ecdsa in cmakelists. remove stdlib_ecdsa from build. * chore: align BARRETENBERG_CRYPTO_GENERATOR_PARAMETERS_HACK usage (AztecProtocol#411) * Update join_split test * Tweaks to comments * Add comment for the assertion in bigfield. * Expanded on ecdsa comment. --------- Co-authored-by: ludamad <adam.domurad@gmail.com> Co-authored-by: ludamad <domuradical@gmail.com> Co-authored-by: codygunton <codygunton@gmail.com> * feat: CI to test aztec circuits with current commit of bberg (AztecProtocol#418) * More generators for aztec3. * update js vk (because we now use UP for merkle hashing) * Helpers for ECDSA in A3 (AztecProtocol#364) * Add `stdlib_keccak` in cmake. Correct an assertion in `to_byte_array` in bigfield. * Add `random_element` to affine element. * negate y conditionally. * Change pedersen hash c_bind to use `pedersen_hash::lookup`. * c_binds and other ECDSA related fixes (AztecProtocol#407) * Add v to stdlib ecdsa. * create an engine if its empty. * Add ecdsa c_bind. * print v as a uint32. * Add secp256k1 cbind. add c_bind.hpp Change hpp to h. remove hpp. * Add ecdsa in cmakelists. remove stdlib_ecdsa from build. * hack: (aztec3) introduce barretenberg crypto generator parameters hack (AztecProtocol#408) * hack: introduce BARRETENBERG_CRYPTO_GENERATOR_PARAMETERS_HACK * doc: concise * chore: align BARRETENBERG_CRYPTO_GENERATOR_PARAMETERS_HACK usage * CI to test aztec circuits with current commit of bberg * build manifest * chore: align BARRETENBERG_CRYPTO_GENERATOR_PARAMETERS_HACK usage (AztecProtocol#411) * try other branch of aztec packages * ci rename script * Update join_split test * bump aztec version and merge in aztec3-temporary fixes * aztec commit switched to branch * bump aztec commit and document * typo README.md * Update README.md --------- Co-authored-by: Suyash Bagad <suyashnbagad1997@gmail.com> Co-authored-by: Suyash Bagad <suyash@aztecprotocol.com> Co-authored-by: ludamad <adam.domurad@gmail.com> Co-authored-by: ludamad <domuradical@gmail.com> * chore: use build-system submodule (AztecProtocol#419) * attempt to remove old CI files and replace with build-system submodule * project and version files for CCI * slack orb context * slack bberg channel (AztecProtocol#422) * Cg/flavor (AztecProtocol#326) - Introducing the flavor classes (mainly honk, splash of plonk) --------- Co-authored-by: ledwards2225 <l.edwards.d@gmail.com> * Add external benchmarks (AztecProtocol#401) Adds an external_bench file with benchmarks we use for external benchmarking projects * Reduce occurence of using namespace syntax in header files (AztecProtocol#387) Co-authored-by: maramihali <mara@aztecprotocol.com> * ensure all operand sizes are tested (AztecProtocol#432) Co-authored-by: maramihali <mara@aztecprotocol.com> * Add ECDSA test for ACIR and fix (AztecProtocol#435) --------- Co-authored-by: zac-williamson <blorktronics@gmail.com> * chore(ci): Add Noir CI that runs with a github label (AztecProtocol#430) * Adds prehashed message variant of EcDSA (AztecProtocol#437) * verification takes a pre-hashed message : Note: if len(hash) > 32 bytes, then bigfield will fail * use hashed_message when generating signature * modify acir structure and function to now use prehashed variant * message -> hashed_message * Ultra Honk (AztecProtocol#412) * DSL: Add valid dummy data for ecdsa constraints when verifier is creating circuit (AztecProtocol#438) * Add way to make verifiers data valid by replacing zeroes with valid public keys and signatures Co-authored-by: Zachary James Williamson <zac-williamson@users.noreply.github.com> * Update cpp/src/barretenberg/dsl/acir_format/ecdsa_secp256k1.test.cpp * replace templates with concrete methods * add comment * PR review * add comments * change to use boolean flag, so dummy_ecdsa method lives in ecdsa * ad true as default --------- Co-authored-by: Zachary James Williamson <zac-williamson@users.noreply.github.com> * chore(ci): Add explicit ref for Noir checkout (AztecProtocol#440) * feat!: add support for ROM and RAM ACVM opcodes (AztecProtocol#417) * *WIP* do not push * Generate constraints for dynamic memory * fix unit test: add missing block_constraint * add unit test for dynamic memory * missed one block constraint in ecdsa unit test * trying a rebase * remove comments * fix the build (AztecProtocol#442) * msgpack: initial support for friendly binary serialization format (AztecProtocol#374) * Regenerate pedersen lookup tables if they're empty * re-init generator tables if they're empty. * feat(nullifier_tree): make empty nullifier tree leaves hash be 0 (AztecProtocol#360) * feat(nullifier_tree): make empty nullifier tree leaves be 0 * fix: add append zero behaviour and test * fix: explicit type name * clean: update class semantics --------- Co-authored-by: cheethas <urmasurda@gmail.com> Co-authored-by: cheethas <addaboy@gmail.com> * More generators for aztec3. * update js vk (because we now use UP for merkle hashing) * Helpers for ECDSA in A3 (AztecProtocol#364) * Add `stdlib_keccak` in cmake. Correct an assertion in `to_byte_array` in bigfield. * Add `random_element` to affine element. * negate y conditionally. * feat(nullifier_tree): make empty nullifier tree leaves hash be 0 (AztecProtocol#360) * feat(nullifier_tree): make empty nullifier tree leaves be 0 * fix: add append zero behaviour and test * fix: explicit type name * clean: update class semantics --------- Co-authored-by: cheethas <urmasurda@gmail.com> Co-authored-by: cheethas <addaboy@gmail.com> * Change pedersen hash c_bind to use `pedersen_hash::lookup`. * feat: add msgpack-c submodule * Give up on msgpack c_master * Working hacky msgpack test * Interim work * Interim work * Getting rid of memory hacks * fix: memory leaks * Start of demoing cbinds * Align with other methods * chore: Remove need to return from msgpack method * Iterate example * fix: Hack around generator issues * feat: iterate on msgpack in bb * fix: fork msgpack for greater checks * Refactor * cleanup * Update turbo_circuit_constructor.cpp * chore: continued cleanup * chore: continued cleanup * chore: continued cleanup * Refactor * Refactor * fix: ci * feat(wasm): hacks to make work in a fno-exceptions wasm environment * feat(wasm): bump msgpack-c * feat(msgpack): first 'complex' object bound * More wasm fixes. Was breaking throw() declaration * Fix field serialization * refactoring * Update CMakeLists.txt * Remove // TODO redundant with msgpack * Refactor to use macro * Refactor to use macro * fix printing bug * fix: fieldd msgpack endianness fix * fix: remove shared ptr reference * doc * Add static checking for MSGPACK usage * Revert log.hpp change * Update struct_map_impl.hpp * Revert * remote_build fix * Keep trying to init submodules * Keep trying to init submodules * Bump * Add missing init_submodules * Msgpack test fix * Msgpack test fix * Msgpack test fix * Msgpack test fix * Update polynomial_store.test.cpp * Merge master * Update msgpack error * Better abort distinguishing * fix: join split VK hash * Serialization updates * Fix circuits build * Try to make circuits test work again * Try to make circuits test work again * Try to make circuits test work again * fix: initialization warning * fix: prefer default constructor for field, related cleanup * Grand rename * chore: remove unused funcs * Revert fields constructor change for now * chore: Revert .circleci changes * chore: Revert foundation removal * Revert .gitmodules * Update affine_element.hpp * Update element.hpp * Revert header optimizations * Revert init line * Update polynomial_store.test.cpp * Revert header optimization * Update raw_pointer.hpp * Update raw_pointer.hpp * Update func_traits.hpp documentation * Document msgpack methods in field_impl.hpp * Update msgpack.hpp * Update cbind.hpp * Update msgpack.hpp * Update msgpack.hpp * Update schema_impl.hpp * Update g1.hpp --------- Co-authored-by: Suyash Bagad <suyashnbagad1997@gmail.com> Co-authored-by: Maddiaa <47148561+cheethas@users.noreply.github.com> Co-authored-by: cheethas <urmasurda@gmail.com> Co-authored-by: cheethas <addaboy@gmail.com> Co-authored-by: Suyash Bagad <suyash@aztecprotocol.com> * Zw/noir recursion 2 (AztecProtocol#414) * removed redundant `reduce` operations after negating biggroup elements simplified hash input structure when hashing transcripts cached partial non native field multiplications reverted how native transcript computes hash buffers pedersen_plookup can be configured to skip the hash_single range check under limited conditions fixed the range check in pedersen_plookup::hash_single pedersen_plookup::hash_single now validates the low and high scalar slice values match the original scalar bigfield::operator- now correctly uses the UltraPlonk code path if able to added biggroup::multiple_montgomery_ladder to reduce required field multiplications added biggroup::quadruple_and_add to reduce required field multiplications biggroup_nafs now directly calls the Composer range constraint methods to avoid creating redundant arithmetic gates when using the PlookupComposer biggroup plookup ROM tables now track the maximum size of any field element recovered from the table (i.e. the maximum of the input maximum sizes) biggroup batch tables prefer to create size-6 lookup tables if doing so reduces the number of individual tables required for a given MSM recursion::transcript no longer performs redundant range constraints when adding buffer elements recursion::transcript correctly checks that, when slicing field elements , the slice values are correct over the integers (i.e. slice_sum != original + p) recursion::verification_key now optimally packs key data into minimum required number of field elements before hashing recursion::verifier proof and key data is now correctly extracted from the transcript/key instead of being generated directly as witnesses. cleaned up code + comments code tidy, added more comments cleaned up how aggregation object handles public inputs native verification_key::compress matches circuit output fixed compile errors + failing tests compiler error join_split.test.cpp passing Note: not changing any upstream .js verification keys. I don't think we need to as bberg is now decoupled from aztec connect * compiler fix * more compiler fix * attempt to fix .js and .sol tests * revert keccak transcript to original functionality * added hash_index back into verification_key::compress fixed composer bug where `decompose_into_default_range` was sometimes not range-constraining last limb removed commented-out code added more descriptive comments to PedersenPreimageBuilder * changed join-split vkey * temporarily point to branch of aztec that updates aggregation state usage until fix is in aztec master * revert .aztec-packages-commit * header brittleness fix * compiler fix * compiler fix w. aggregation object * reverting changes to `assign_object_to_proof_outputs` to preserve backwards-compatibility with a3-packages * more backwards compatibility fixes * wip --------- Co-authored-by: dbanks12 <david@aztecprotocol.com> Co-authored-by: David Banks <47112877+dbanks12@users.noreply.github.com> * Chore: bundle msgpack to fix nix-build (AztecProtocol#450) * Revert msgpack submodule * Bundle msgpack to avoid issues with submodules * variable-length keccak (AztecProtocol#441) * updated stdlib::keccak to be able to hash variable-length inputs (where input size not known at circuit-compile time, only a maximum possible input size) * compile error * compile fils * compiler fix * more fix * compiler fix * compile fix * Update cpp/src/barretenberg/stdlib/hash/keccak/keccak.test.cpp Co-authored-by: kevaundray <kevtheappdev@gmail.com> * Update cpp/src/barretenberg/stdlib/hash/keccak/keccak.test.cpp Co-authored-by: kevaundray <kevtheappdev@gmail.com> * Update cpp/src/barretenberg/stdlib/hash/keccak/keccak.test.cpp Co-authored-by: kevaundray <kevtheappdev@gmail.com> * Update cpp/src/barretenberg/stdlib/hash/keccak/keccak.test.cpp Co-authored-by: kevaundray <kevtheappdev@gmail.com> * Update cpp/src/barretenberg/stdlib/primitives/field/field.test.cpp Co-authored-by: kevaundray <kevtheappdev@gmail.com> * improved readability of stdlib test * replaced magic numbers in keccak with constants + comments --------- Co-authored-by: kevaundray <kevtheappdev@gmail.com> * chore: disable circuits tests in master (AztecProtocol#454) * fix: msgpack error (AztecProtocol#456) * Add missing `hash_index` while compressing vk. (AztecProtocol#457) * Add missing `hash_index` while compressing vk. * comment back vk tests with hash index > 0. * Adam/fix allow explicit field init (AztecProtocol#460) * fix: msgpack error * fix: allow explicit field init * fix: msgpack variant_impl.hpp (AztecProtocol#462) Previous version accidentally created a packer<packer<Stream>> * fix: bbmalloc linker error (AztecProtocol#459) * format msgpack serialization and excldue msgpack-c from clang-format (AztecProtocol#467) * patch: temporarily remove broken solidity ci (AztecProtocol#470) * Sumcheck improvements (AztecProtocol#455) * convert partially evaluated polynomials from vectors to Polynomials and rename * rename fold method to partially_evaluate * static constexpr barycentric arrays * change purported evaluations to claimed evaluations * specify relations in Flavor * Fixed a bug in biggroup tests (AztecProtocol#478) * DSL: Add KeccakVar opcode (AztecProtocol#476) * add initial KeccakVar code * add result field * add keccak_var_constraints to fields * Multi-constraint Relations (AztecProtocol#444) Allow for correct and efficient batching over identities in the Sumcheck relation --------- Co-authored-by: Zachary James Williamson <blorktronics@gmail.com> Co-authored-by: Maxim Vezenov <mvezenov@gmail.com> Co-authored-by: Adam Domurad <adam@aztecprotocol.com> Co-authored-by: ledwards2225 <98505400+ledwards2225@users.noreply.github.com> Co-authored-by: codygunton <codygunton@gmail.com> Co-authored-by: spypsy <spypsy@users.noreply.github.com> Co-authored-by: Santiago Palladino <spalladino@gmail.com> Co-authored-by: Innokentii Sennovskii <isennovskiy@gmail.com> Co-authored-by: ledwards2225 <l.edwards.d@gmail.com> Co-authored-by: Blaine Bublitz <blaine.bublitz@gmail.com> Co-authored-by: Suyash Bagad <suyashnbagad1997@gmail.com> Co-authored-by: Maxim Vezenov <maximvezenov@Maxims-MacBook-Pro.local> Co-authored-by: kevaundray <kevtheappdev@gmail.com> Co-authored-by: Suyash Bagad <suyash@aztecprotocol.com> Co-authored-by: David Banks <47112877+dbanks12@users.noreply.github.com> Co-authored-by: Michael Connor <mike@aztecprotocol.com> Co-authored-by: dbanks12 <david@aztecprotocol.com> Co-authored-by: Maddiaa <47148561+cheethas@users.noreply.github.com> Co-authored-by: Santiago Palladino <santiago@aztecprotocol.com> Co-authored-by: cheethas <urmasurda@gmail.com> Co-authored-by: maramihali <mara@aztecprotocol.com> Co-authored-by: Max Hora <maxhora.ua@gmail.com> Co-authored-by: cheethas <addaboy@gmail.com> Co-authored-by: Lasse Herskind <16536249+LHerskind@users.noreply.github.com> Co-authored-by: Tom French <15848336+TomAFrench@users.noreply.github.com> Co-authored-by: guipublic <47281315+guipublic@users.noreply.github.com> Co-authored-by: maramihali <maramihali@google.com> Co-authored-by: Zachary James Williamson <zac-williamson@users.noreply.github.com> Co-authored-by: Maddiaa <47148561+Maddiaa0@users.noreply.github.com>

* Add check_circuit with mid-construction introspection

Rumata888 force-pushed the is/check_circuit branch from f9634a7 to ae7f89b Compare April 21, 2023 15:55

Rumata888 requested a review from ledwards2225 April 21, 2023 22:45

Rumata888 marked this pull request as ready for review April 21, 2023 22:45

codygunton mentioned this pull request Apr 24, 2023

Port UltraComposer changes to UltraPlonkComposer and UltraHonkComposer #378

Closed

Rumata888 added 9 commits April 25, 2023 13:49

FInished aux

9579fe9

In-the-head stuff is added

993ef5b

Check circuit is woests and comments left

b52c434

Added comments and state checks

aec6ac6

Move plookup tables into proof_system

2b11ed6

All the standard tests are working with check_circuit

8fe3463

Added a showcase for check_circuit

77a6e6f

Moved members from proof_system into the UltraCircuitConstructor

4f5ac86

Fixed non-native caching

26abe0e

Rumata888 force-pushed the is/check_circuit branch from a1b6748 to 26abe0e Compare April 25, 2023 14:43

codygunton requested review from codygunton and removed request for ledwards2225 April 25, 2023 15:17

codygunton reviewed Apr 25, 2023

View reviewed changes

codygunton requested changes Apr 25, 2023

View reviewed changes

codygunton assigned Rumata888 Apr 25, 2023

Rumata888 added 2 commits April 26, 2023 22:56

Changed check_circuit according to Cody's suggestion

88e682d

Some cosmetic changes

3ffca90

codygunton approved these changes Apr 27, 2023

View reviewed changes

codygunton merged commit 7b55988 into master Apr 27, 2023

codygunton deleted the is/check_circuit branch April 27, 2023 20:16

ludamad pushed a commit to AztecProtocol/aztec-packages that referenced this pull request Jul 22, 2023

Ultraplonk check_circuit (AztecProtocol/barretenberg#366)

113292d

* Add check_circuit with mid-construction introspection

ludamad pushed a commit to AztecProtocol/aztec-packages that referenced this pull request Jul 24, 2023

Ultraplonk check_circuit (AztecProtocol/barretenberg#366)

c94730d

* Add check_circuit with mid-construction introspection

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Ultraplonk check_circuit #366

Ultraplonk check_circuit #366

Rumata888 commented Apr 19, 2023 •

edited

Loading

codygunton Apr 25, 2023

codygunton Apr 25, 2023 •

edited

Loading

codygunton Apr 25, 2023

codygunton left a comment

codygunton Apr 25, 2023

codygunton Apr 25, 2023

codygunton Apr 25, 2023

codygunton Apr 25, 2023

codygunton Apr 25, 2023

codygunton Apr 25, 2023

codygunton Apr 25, 2023

codygunton Apr 25, 2023 •

edited

Loading

codygunton Apr 25, 2023

codygunton Apr 25, 2023

codygunton Apr 25, 2023 •

edited

Loading

codygunton Apr 25, 2023

codygunton Apr 25, 2023 •

edited

Loading

codygunton commented Apr 27, 2023

Ultraplonk check_circuit #366

Ultraplonk check_circuit #366

Conversation

Rumata888 commented Apr 19, 2023 • edited Loading

Implements the ultra_circuit_constructor check_circuit method

Checklist:

Choose a reason for hiding this comment

codygunton Apr 25, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codygunton left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codygunton Apr 25, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codygunton Apr 25, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codygunton Apr 25, 2023 • edited Loading

Choose a reason for hiding this comment

codygunton commented Apr 27, 2023

Rumata888 commented Apr 19, 2023 •

edited

Loading

codygunton Apr 25, 2023 •

edited

Loading

codygunton Apr 25, 2023 •

edited

Loading

codygunton Apr 25, 2023 •

edited

Loading

codygunton Apr 25, 2023 •

edited

Loading