crypto::generator_data< Curve > Class Template Reference

class that stores precomputed generators used for Pedersen commitments and Pedersen hashes More...

#include <generator_data.hpp>

Public Types
using	Group = typename Curve::Group
using	AffineElement = typename Curve::AffineElement
using	GeneratorList = std::vector< AffineElement >
using	GeneratorView = std::span< AffineElement const >

Public Member Functions
GeneratorView	get (const size_t num_generators, const size_t generator_offset=0, const std::string_view domain_separator=DEFAULT_DOMAIN_SEPARATOR) const

Static Public Member Functions
static constexpr std::array< AffineElement, DEFAULT_NUM_GENERATORS >	make_precomputed_generators ()
static generator_data *	get_default_generators ()

Static Public Attributes
static constexpr size_t	DEFAULT_NUM_GENERATORS = 8
static constexpr std::string_view	DEFAULT_DOMAIN_SEPARATOR = "DEFAULT_DOMAIN_SEPARATOR"
static constexpr std::array< AffineElement, DEFAULT_NUM_GENERATORS >	precomputed_generators
	Precompute a small number of generators at compile time. For small pedersen commitments + pedersen hashes, this prevents us from having to derive generators at runtime.

Detailed Description

template<typename Curve>
class crypto::generator_data< Curve >

class that stores precomputed generators used for Pedersen commitments and Pedersen hashes

We create distinct sets of generators via the use of a domain separator. This enables the use of context-specific commitments and hashes. For example, a circuit that generates commitments foo = commit({ a, b }) and bar = commit({c, d}) where foo and bar should not collide.

The goal of generator_data is twofold:

1. Prevent redundant computation of the same generators at runtime (i.e. store in a singleton object)
2. Compute a small number of default generators at compile-time, so that short processes that require a small number of generators do not have to execute the expensive g1::derive_generators method

We store generators in a key:value map, where the key is the domain separator and the value is the vector of associated generators. Pedersen methods take in a pointer to a generator_data object.

generator_data contains a static instantiation of the class: default_data. The intention is for default_data to be used as a singleton class. All Pedersen methods that require a *generator_data parameter (from now on referred to as "generator context") should default to using default_data.

Q: Why make the generator context an input parameter when it defaults to default_data? A: This is not thread-safe. Each process that uses a generator_data object may extend generator_data if more generators are required. i.e. either each process must use an independent generator_data object or the author must KNOW that generator_data will not be extended by any process

Template Parameters

Curve

Member Data Documentation

precomputed_generators

template<typename Curve >

constexpr std::array<AffineElement, DEFAULT_NUM_GENERATORS> crypto::generator_data< Curve >::precomputed_generators

inlinestaticconstexpr

Initial value:

=
        make_precomputed_generators()

Precompute a small number of generators at compile time. For small pedersen commitments + pedersen hashes, this prevents us from having to derive generators at runtime.

---

The documentation for this class was generated from the following file:

• src/barretenberg/crypto/generators/generator_data.hpp