tachyon
tachyon copied to clipboard
Modular ZK(Zero Knowledge) backend accelerated by GPU
Overview
Tachyon is a Modular ZK Backend, powered by GPU.
Design Goals
- General Purpose: A versatile ZK library empowers developers to implement any proving scheme with minimal effort, often enhancing developer productivity. To create a general-purpose backend, aligning the code structure as closely as possible with the algebraic structure is paramount.
- Easy to Use: Achieving widespread adoption is essential for the success of any product. Consequently, one of the key focal points of the Tachyon project is to include offering packages for various programming languages and runtimes.
- Blazing Fast: Tachyon's foremost requirement is speed, and not just any speed, but blazing speed! This entails Tachyon delivering exceptional performance on both CPU and GPU platforms.
- GPU Interoperability: Tachyon's code is designed to be compatible with both CPU and GPU in the majority of scenarios.
Prerequisites
Bazel
Please follow the instructions here.
Ubuntu
> sudo apt install libgmp-dev libomp-dev
Macos
> brew install gmp
Getting started
Build
Build on Linux
> bazel build --config linux //...
Build on Macos arm64
> bazel build --config macos_arm64 //...
Build on Macos x64
> bazel build --config macos_x86_64 //...
NOTE:: MacOS v14.0.0 or later is recommended.
In certain versions of MacOS (prior to v13.5.1), a bug related to incorrect Bigint divide operations has been detected in the field generator when using the optimized build (-c opt).
The issue will be fixed as soon as possible.
Test
Test on Linux
> bazel test --config linux //...
Test on Macos arm64
> bazel test --config macos_arm64 //...
Test on Macos x64
> bazel test --config macos_x86_64 //...
Configuration
Polygon zkEVM backend prime field
NOTE:: Only x86_64 is supported.
-
--//:polygon_zkevm_backend: Enable goldilocks and zkevm-prover prime field backend.> bazel build --config ${os} --config avx512_${os} --//:polygon_zkevm_backend //...
Hardware acceleration
CUDA backend
-
--config cuda: Enable cuda backend.> bazel build --config ${os} --config cuda //...
ROCm backend
-
--config rocm: Enable rocm backend.> bazel build --config ${os} --config rocm //...
Matplotlib
Pyenv
If you are using pyenv, don't forget to add an option --enable-shared.
> CONFIGURE_OPTS=--enable-shared pyenv install <version>
Python dependencies
> pip install matplotlib
Frequently Asked Questions
Debugging on macOS
Please add this line to your .bazelrc.user.
build --spawn_strategy=local
Build on Ubuntu 20.04
Please update g++ version and try build again! The default g++-9 is not working.
> sudo apt install g++-10
> export CC=/usr/bin/gcc-10
> export CXX=/usr/bin/g++-10
> export GCC_HOST_COMPILER_PATH=/usr/bin/gcc-10
Build CUDA with rust toolchain
Use --sandbox_debug to see verbose messages from the sandbox and retain the sandbox build root for debugging
error: linking with `external/local_config_cuda/crosstool/clang/bin/crosstool_wrapper_driver_is_not_gcc` failed: exit status: 127
...
= note: /usr/bin/env: 'python': No such file or directory
Please make your python point to python interpreter to be run.
> sudo apt install python-is-python3
Plus, please include these lines to your .bazelc.user.
build --action_env=PATH=/usr/bin:/usr/local/bin
build --host_action_env=PATH=/usr/bin:/usr/local/bin