Requirements

Requirements#

Hardware#

  • NVIDIA GPU with CUDA support.

Dependencies#

You can prepare a docker environment with the following dependencies. For more information, see Quick Start.

Required dependencies#

Dependency

Minimum Version

Notes

CMake

3.18

Build system

C++ compiler

C++20 support

GCC 11+, Clang 14+

Python

3.12

JIT codegen + test runner

flagtree

0.5.0

Triton TLE support

SQLite3

—

Tuning database

CUDA Toolkit

12.x

cudart, cuFFT (for test adaptor/benchmarks)

libtriton_jit

submodule

Triton JIT compiler (deps/libtriton_jit)

PyYAML

—

Test runner (pip install pyyaml)

Optional dependencies#

Dependency

Purpose

Google Test

C++ unit tests (auto-fetched via FetchContent when FLAGFFT_BUILD_TESTS=ON)

Ninja

Faster build backend (cmake -G Ninja)

pytest

Python codegen tests

Submodule#

git submodule update --init --recursive

This pulls in deps/libtriton_jit, which provides the Triton JIT compiler and nlohmann_json. For more information, see Quick Start.