Skip to content

install.md

Latest commit

 

History

History
184 lines (132 loc) · 5.87 KB

install.md

File metadata and controls

184 lines (132 loc) · 5.87 KB
Raw

BioIR Package Installation

The following sections provide the steps to install the BioNeMo Inference Runtime (BioIR) package, along with the software and hardware requirements.

Requirements to Install the BioIR Wheel

The release wheel has the following software and hardware requirements:

  • Operating system: A Linux distribution that supports the latest NVIDIA drivers, with glibc 2.34 or newer. Release wheels are tagged manylinux_2_34, so pip reports no matching distribution on an older distribution such as Ubuntu 20.04 or RHEL 8.
  • CPU architecture: x86_64 (amd64) or aarch64 (arm64).
  • GPU: An NVIDIA GPU listed in the support matrix.
  • NVIDIA Driver: 580 or higher.
  • Python: Python 3.12. Release wheels are tagged cp312.
  • aarch64 build tools: A C/C++ compiler and Python 3.12 development headers. Some transitive dependencies build from source on aarch64.

Use the commands in Collecting System Information to verify these requirements. If the installed NVIDIA driver is older than version 580, refer to the GPU Stack guide.

GPU Requirements

H200, H100, A100, L40S, GB200, and GB300 are release-qualified with measured speed, peak memory, and accuracy in the benchmarks.

BioIR also runs on other NVIDIA GPUs, including Ampere, Ada Lovelace, Hopper, and Blackwell. The support matrix distinguishes backend compatibility from release qualification and lists the optimized kernels for each architecture.

Driver Requirements

Use NVIDIA driver 580 or newer. BioIR is built against CUDA 13.2, but the CUDA Toolkit does not need to be installed on the host. BioIR requires the driver-provided libcuda.so.1.

Drivers older than 580 are unsupported. The CUDA forward-compatibility shim caused hangs and crashes during inference testing on older drivers.

Check the installed GPU and driver with commands at Collecting System Information.

Choose an Installation Method

Choose one of the following installation methods based on your task:

  • Release wheel (recommended): Install BioIR into a Python environment to call its models, optimized modules, and prediction pipeline from your code. This method is the simplest way to start using BioIR.
  • Source build: Use the development container when changing BioIR or running its test suite. Follow the development workflow for Docker, NVIDIA Container Toolkit, compiler, and CUDA-header requirements.

Install the Release Wheel

BioIR is published on PyPI as bionemo-ir, with one wheel for each supported CPU architecture. pip selects the wheel that matches the host.

Install aarch64 Build Prerequisites

On Ubuntu 24.04 aarch64, install the compiler toolchain and Python headers used to build transitive dependencies:

sudo apt-get update
sudo apt-get install --yes build-essential python3.12-dev

The python3.12-dev package provides Python.h.

Create a Python Environment

The release wheel requires Python 3.12 (cp312). A dedicated Python environment is strongly recommended. It isolates BioIR and its dependencies from the system Python and other projects.

Choose one of the following methods to create and activate an environment:

Install uv, then run:

uv venv --python 3.12 --seed .venv
source .venv/bin/activate

The --seed flag installs pip in the environment for the shared install commands below.

[Install conda][conda], then create and activate a named environment: 
conda create --name bioir python=3.12
conda activate bioir
Install Python 3.12 on the host, then create and activate an environment with Python's built-in `venv` module: 
python3.12 -m venv .venv
source .venv/bin/activate

Keep the environment activated for the rest of this guide and for later work, including the Quickstart. Confirm that it uses Python 3.12, then upgrade pip:

python --version
python -m pip install --upgrade pip

Install the Package

Install BioIR into the active environment:

python -m pip install bionemo-ir

pip resolves PyTorch, the CUDA libraries, and the remaining dependencies from the same index. The wheel carries the GPU kernels precompiled, so the install builds no CUDA code and running it needs only the driver's libcuda.so.1.

Verify the install:

python -c "import bionemo_ir; print(bionemo_ir.__version__)"

Next Steps