rpu

Flagtree Third Party Backend - RPU Accelerator Support

Overview

Flagtree Third Party Backend for the Rhino RPU accelerator, providing the Triton compiler backend and the test suite for developing and deploying Triton kernels on RPU hardware.

The Rhino RPU is an AI accelerator from Huixi Intelligence (辉羲智能, rhino.auto), a company building high-performance, low-power compute platforms and chips for embodied AI and autonomous driving — such as its publicly announced 光至 R1 SoC.

中文版: README_cn.md

Prerequisites

• aarch64 Linux board with the RPU device node (/dev/rpu) and kernel module
• RPU LLVM toolchain (custom LLVM providing bin/clang)
• RPU launch runtime library rhino-launch-kernel (ships librhino_launch.so), plus the launch_kernel_runner CLI built against it (see "On-board launch_kernel check")
• Minimum 24 GB RAM and 60 GB free disk (clean build is ~18-25 minutes on a 24-core board)
• Python 3.10+, cmake >= 3.20, ninja, a recent pip

Environment Preparation

1. Pull Source Code

cd ~
git clone https://github.com/flagos-ai/FlagTree.git
cd FlagTree
git checkout main

2. Install RPU Driver and Runtime

Install the RPU kernel driver (provides /dev/rpu) and the rhino-launch-kernel runtime library (librhino_launch.so, used by the on-board launch_kernel test) per vendor instructions. Verify the device node:

ls /dev/rpu

3. Prepare RPU LLVM Toolchain

The RPU backend uses a custom LLVM as its .rpubin emitter. Point RPU_LLVM_ROOT at the toolchain install prefix (the directory containing bin/clang):

# Toolchain layout:
#   $RPU_LLVM_ROOT/bin/clang
#   $RPU_LLVM_ROOT/lib/...
export RPU_LLVM_ROOT=/opt/rpu/llvm

Contact the vendor to obtain the toolchain.

Build and Install

1. Configure Build Environment

export FLAGTREE_BACKEND=rpu
export MAX_JOBS=8                # tune to available RAM

2. Install Python Dependencies

cd ~/FlagTree/python
pip3 install -r requirements.txt

3. Build and Install Package

cd ~/FlagTree

# Initial build
pip3 install . --no-build-isolation -v

# Rebuild after source changes
pip3 install . --no-build-isolation --force-reinstall -v

The Triton MLIR LLVM is downloaded automatically from the public oaitriton blob the first time setup runs; no manual action needed.

Test Validation

Unit tests

The compile tests drive the real toolchain, so set RPU_LLVM_ROOT first (the directory containing bin/clang). If it is unset or its path is missing, the tests stop with a clear message naming the variable.

cd ~/FlagTree
export RPU_LLVM_ROOT=/opt/rpu/llvm
pytest -s third_party/rpu/python/test/unit

The suite covers the compile pipeline and per-op contracts.

On-board launch_kernel check

This needs the launch_kernel_runner CLI, a thin front-end over the rhino-launch-kernel runtime library. Build it together with the backend by enabling RPU_BUILD_LAUNCH_RUNNER (off by default; it needs the rhino-launch-kernel install) on the package build, then run the smoke test on a board with /dev/rpu:

TRITON_APPEND_CMAKE_ARGS="-DRPU_BUILD_LAUNCH_RUNNER=ON -DCMAKE_PREFIX_PATH=/path/to/rhino-launch-kernel/install" \
    pip3 install . --no-build-isolation --force-reinstall -v
export RPU_LK_RUNNER=$PWD/third_party/rpu/tools/launch_runner/launch_kernel_runner
python3 third_party/rpu/python/test/board/lk_board_smoke.py --require-board

It compiles a small kernel, dispatches it on the device, and compares the result to a numpy golden. Without --require-board it skips cleanly when the toolchain / runtime / device is unavailable. (tools/launch_runner can also be built standalone — see its README.)