libff

C++ library for Finite Fields and Elliptic Curves

(forked from https://github.com/scipr-lab/libff)

libff is a C++ library for finite fields and elliptic curves. The library is developed by SCIPR Lab and contributors (see AUTHORS file) and is released under the MIT License (see LICENSE file).

Table of contents

• Directory structure
• Elliptic curve choices
• Build guide

Directory structure

The directory structure is as follows:

• libff: C++ source code, containing the following modules:
  • algebra: fields and elliptic curve groups
  • common: miscellaneous utilities
• depends: dependency libraries

Elliptic curve choices

The libff library currently provides the following options:

• edwards: an instantiation based on an Edwards curve, providing 80 bits of security.

• bn128: an instantiation based on a Barreto-Naehrig curve, providing 128 bits of security. The underlying curve implementation is [ate-pairing], which has incorporated our patch that changes the BN curve to one suitable for SNARK applications.

  • This implementation uses dynamically-generated machine code for the curve arithmetic. Some modern systems disallow execution of code on the heap, and will thus block this implementation.

    For example, on Fedora 20 at its default settings, you will get the error zmInit ERR:can't protect when running this code. To solve this, run sudo setsebool -P allow_execheap 1 to allow execution, or use make CURVE=ALT_BN128 instead.

• alt_bn128: an alternative to bn128, somewhat slower but avoids dynamic code generation.

• mnt: An implementation of the MNT cycle (MNT4-MNT6) with small fields.

• bls12-377: An implementation of the BLS12_377 curve as introduced in the Zexe paper.

• bw6_761: An implementation of the BW6_761 curve as introduced in Optimized and secure pairing-friendly elliptic curves suitable for one layer proof composition

Note that bn128 requires an x86-64 CPU while the other curve choices should be architecture-independent.

Build guide

The library has the following dependencies:

• Boost
• CMake
• GMP
• libprocps

Furthermore, Doxygen is used to generate the documentation.

The library has been tested on Linux, but it is compatible with Windows and Mac OS X.

Installation

On Ubuntu 20.04 LTS:

sudo apt update -y
sudo apt install \
    build-essential \
    git \
    libboost-all-dev \
    cmake \
    libgmp3-dev \
    libssl-dev \
    libprocps-dev \
    pkg-config

Note: To install Doxygen, run sudo apt install doxygen graphviz.

Fetch dependencies from their GitHub repos:

git submodule init && git submodule update

Compilation

To compile, starting at the project root directory, create the build directory and Makefile:

mkdir build && cd build && cmake ..

Optionally, you can specify the install location by providing the desired install path prefix:

cmake .. -DCMAKE_INSTALL_PREFIX=/install/path

Then, to compile and install the library, run this within the build directory:

make
make install

This will install libff.a into /install/path/lib; so your application should be linked using -L/install/path/lib -lff. It also installs the requisite headers into /install/path/include; so your application should be compiled using -I/install/path/include.

Testing

To execute the tests for this library, run:

make check

Generate the Doxygen documentation

To generate the documentation, run:

cd build
cmake .. -DGEN_DOC=ON && make docs

Profile

To compile the multi-exponentiation profiler in this library, run:

make profile

The resulting profiler is named multiexp_profile and can be found in the libff folder under the build directory.