Skip to content

Documentation updates for v0.48.0 #1770

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
matthewdouglas merged 6 commits into from
Sep 30, 2025
Merged

Documentation updates for v0.48.0 #1770

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
c8f642b
Update installation docs
matthewdouglas Sep 30, 2025
b2cf0d2
Update links
matthewdouglas Sep 30, 2025
a46b7e9
Fix cuda min glibc in doc
matthewdouglas Sep 30, 2025
a20ad5e
Update header levels
matthewdouglas Sep 30, 2025
bc64626
Update AMD section
matthewdouglas Sep 30, 2025
38b6ce9
typo
matthewdouglas Sep 30, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
222 changes: 101 additions & 121 deletions docs/source/installation.mdx
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,20 +1,34 @@
# Installation Guide

Welcome to the installation guide for the `bitsandbytes` library! This document provides step-by-step instructions to install `bitsandbytes` across various platforms and hardware configurations. The library primarily supports CUDA-based GPUs, but the team is actively working on enabling support for additional backends like CPU, AMD ROCm, Intel XPU, and Gaudi HPU.
Welcome to the installation guide for the `bitsandbytes` library! This document provides step-by-step instructions to install `bitsandbytes` across various platforms and hardware configurations.

We provide official support for NVIDIA GPUs, CPUs, Intel XPUs, and Intel Gaudi platforms. We also have experimental support for
additional platforms such as AMD ROCm.

## Table of Contents

- [CUDA](#cuda)
- [System Requirements](#requirements)
- [NVIDIA CUDA](#cuda)
- [Installation via PyPI](#cuda-pip)
- [Compile from Source](#cuda-compile)
- [Preview Wheels from `main`](#cuda-preview)
- [Multi-Backend Preview](#multi-backend)
- [Supported Backends](#multi-backend-supported-backends)
- [Pre-requisites](#multi-backend-pre-requisites)
- [Installation](#multi-backend-pip)
- [Compile from Source](#multi-backend-compile)
- [Intel XPU](#xpu)
- [Installation via PyPI](#xpu-pip)
- [Intel Gaudi](#gaudi)
- [Installation via PyPI](#gaudi-pip)
- [CPU](#cpu)
- [Installation via PyPI](#cpu-pip)
- [Compile from Source](#cpu-compile)
- [AMD ROCm (Preview)](#rocm-preview)
- [Preview Wheels](#preview-wheels)

## System Requirements[[requirements]]

These are the minimum requirements for `bitsandbytes` across all platforms. Please be aware that some compute platforms may impose more strict requirements.

## CUDA[[cuda]]
* Python >= 3.9
* PyTorch >= 2.3

## NVIDIA CUDA[[cuda]]

`bitsandbytes` is currently supported on NVIDIA GPUs with [Compute Capability](https://developer.nvidia.com/cuda-gpus) 6.0+.
The library can be built using CUDA Toolkit versions as old as **11.8**.
Expand All @@ -25,11 +39,13 @@ The library can be built using CUDA Toolkit versions as old as **11.8**.
| 8-bit optimizers/quantization | 6.0+ | Pascal (GTX 10X0 series, P100) or newer GPUs|
| NF4/FP4 quantization | 6.0+ | Pascal (GTX 10X0 series, P100) or newer GPUs|


> [!WARNING]
> Support for Maxwell GPUs is deprecated and will be removed in a future release.
> Maxwell support is not included in PyPI distributions from `v0.48.0` on and must be built from source.
> For the best results, a Turing generation device or newer is recommended.


### Installation via PyPI[[cuda-pip]]

This is the most straightforward and recommended installation option.
Expand All @@ -40,20 +56,22 @@ The currently distributed `bitsandbytes` packages are built with the following c
|--------------------|------------------|----------------------|--------------
| **Linux x86-64** | 11.8 - 12.6 | GCC 11.2 | sm60, sm70, sm75, sm80, sm86, sm89, sm90
| **Linux x86-64** | 12.8 - 12.9 | GCC 11.2 | sm70, sm75, sm80, sm86, sm89, sm90, sm100, sm120
| **Linux x86-64** | 13.0 | GCC 11.2 | sm75, sm80, sm86, sm89, sm90, sm100, sm120
| **Linux x86-64** | 13.0 | GCC 11.2 | sm75, sm80, sm86, sm89, sm90, sm100, sm110, sm120
| **Linux aarch64** | 11.8 - 12.6 | GCC 11.2 | sm75, sm80, sm90
| **Linux aarch64** | 12.8 - 13.0 | GCC 11.2 | sm75, sm80, sm90, sm100, sm120
| **Windows x86-64** | 11.8 - 12.6 | MSVC 19.43+ (VS2022) | sm50, sm60, sm75, sm80, sm86, sm89, sm90
| **Windows x86-64** | 12.8 - 12.9 | MSVC 19.43+ (VS2022) | sm70, sm75, sm80, sm86, sm89, sm90, sm100, sm120
| **Windows x86-64** | 13.0 | MSVC 19.43+ (VS2022) | sm75, sm80, sm86, sm89, sm90, sm100, sm120

Use `pip` or `uv` to install:
The Linux build has a minimum glibc version of 2.24.

Use `pip` or `uv` to install the latest release:

```bash
pip install bitsandbytes
```

### Compile from source[[cuda-compile]]
### Compile from Source[[cuda-compile]]

> [!TIP]
> Don't hesitate to compile from source! The process is pretty straight forward and resilient. This might be needed for older CUDA Toolkit versions or Linux distributions, or other less common configurations.
Expand Down Expand Up @@ -102,131 +120,102 @@ Big thanks to [wkpark](https://github.com/wkpark), [Jamezo97](https://github.com
</hfoption>
</hfoptions>

### Preview Wheels from `main`[[cuda-preview]]
## Intel XPU[[xpu]]

If you would like to use new features even before they are officially released and help us test them, feel free to install the wheel directly from our CI (*the wheel links will remain stable!*):
* A compatible PyTorch version with Intel XPU support is required. The current minimum is **PyTorch 2.6.0**. It is recommended to use the latest stable release. See [Getting Started on Intel GPU](https://docs.pytorch.org/docs/stable/notes/get_start_xpu.html) for guidance.

<hfoptions id="OS">
<hfoption id="Linux">
### Installation via PyPI[[xpu-pip]]

```bash
# Note: if you don't want to reinstall our dependencies, append the `--no-deps` flag!
This is the most straightforward and recommended installation option.

# x86_64 (most users)
pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-manylinux_2_24_x86_64.whl
The currently distributed `bitsandbytes` packages are built with the following configurations:

# ARM/aarch64
pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-manylinux_2_24_aarch64.whl
```
| **OS** | **oneAPI Toolkit** | **Kernel Implementation** |
|--------------------|------------------|----------------------|
| **Linux x86-64** | 2025.1.3 | SYCL + Triton |
| **Windows x86-64** | N/A | SYCL |

</hfoption>
<hfoption id="Windows">
The Linux build has a minimum glibc version of 2.34.

Use `pip` or `uv` to install the latest release:

```bash
# Note: if you don't want to reinstall our dependencies, append the `--no-deps` flag!
pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-win_amd64.whl
pip install bitsandbytes
```
</hfoption>
</hfoptions>

## Intel Gaudi[[gaudi]]

## Multi-Backend Preview[[multi-backend]]
* A compatible PyTorch version with Intel Gaudi support is required. The current minimum is **Gaudi v1.21** with **PyTorch 2.6.0**. It is recommended to use the latest stable release. See the Gaudi software [installation guide](https://docs.habana.ai/en/latest/Installation_Guide/index.html) for guidance.

> [!WARNING]
> This functionality existed as an early technical preview and is not recommended for production use. We are in the process of upstreaming improved support for AMD and Intel hardware into the main project.

We provide an early preview of support for AMD and Intel hardware as part of a development branch.
### Installation from PyPI[[gaudi-pip]]

### Supported Backends[[multi-backend-supported-backends]]

| **Backend** | **Supported Versions** | **Python versions** | **Architecture Support** | **Status** |
|-------------|------------------------|---------------------------|-------------------------|------------|
| **AMD ROCm** | 6.1+ | 3.10+ | minimum CDNA - `gfx90a`, RDNA - `gfx1100` | Alpha |
| **Intel CPU** | v2.4.0+ | 3.10+ | Intel CPU | Alpha |
| **Intel GPU** | v2.7.0+ | 3.10+ | Intel GPU | Experimental |
| **Ascend NPU** | 2.1.0+ (`torch_npu`) | 3.10+ | Ascend NPU | Experimental |

For each supported backend, follow the respective instructions below:

### Pre-requisites[[multi-backend-pre-requisites]]

To use this preview version of `bitsandbytes` with `transformers`, be sure to install:
Use `pip` or `uv` to install the latest release:

```bash
pip install "transformers>=4.45.1"
pip install bitsandbytes
```

<hfoptions id="backend">
<hfoption id="AMD ROCm">
## CPU[[cpu]]

> [!WARNING]
> Pre-compiled binaries are only built for ROCm versions `6.1.2`/`6.2.4`/`6.3.2` and `gfx90a`, `gfx942`, `gfx1100` GPU architectures. [Find the pip install instructions here](#multi-backend-pip).
>
> Other supported versions that don't come with pre-compiled binaries [can be compiled for with these instructions](#multi-backend-compile).
>
> **Windows is not supported for the ROCm backend**
### Installation from PyPI[[cpu-pip]]

> [!TIP]
> If you would like to install ROCm and PyTorch on bare metal, skip the Docker steps and refer to ROCm's official guides at [ROCm installation overview](https://rocm.docs.amd.com/projects/install-on-linux/en/latest/tutorial/install-overview.html#rocm-install-overview) and [Installing PyTorch for ROCm](https://rocm.docs.amd.com/projects/install-on-linux/en/latest/how-to/3rd-party/pytorch-install.html#using-wheels-package) (Step 3 of wheels build for quick installation). Special note: please make sure to get the respective ROCm-specific PyTorch wheel for the installed ROCm version, e.g. `https://download.pytorch.org/whl/nightly/rocm6.2/`!
This is the most straightforward and recommended installation option.

```bash
# Create a docker container with the ROCm image, which includes ROCm libraries
docker pull rocm/dev-ubuntu-22.04:6.3.4-complete
docker run -it --device=/dev/kfd --device=/dev/dri --group-add video rocm/dev-ubuntu-22.04:6.3.4-complete
apt-get update && apt-get install -y git && cd home
The currently distributed `bitsandbytes` packages are built with the following configurations:

# Install pytorch compatible with above ROCm version
pip install torch --index-url https://download.pytorch.org/whl/rocm6.3/
```
| **OS** | **Host Compiler** | Hardware Minimum
|--------------------|----------------------|----------------------|
| **Linux x86-64** | GCC 11.4 | AVX2 |
| **Linux aarch64** | GCC 11.4 | |
| **Windows x86-64** | MSVC 19.43+ (VS2022) | AVX2 |

</hfoption>
<hfoption id="Intel XPU">
The Linux build has a minimum glibc version of 2.24.

* A compatible PyTorch version with Intel XPU support is required. It is recommended to use the latest stable release. See [Getting Started on Intel GPU](https://docs.pytorch.org/docs/stable/notes/get_start_xpu.html) for guidance.
Use `pip` or `uv` to install the latest release:

</hfoption>
</hfoptions>
```bash
pip install bitsandbytes
```

### Installation
### Compile from Source[[cpu-compile]]

You can install the pre-built wheels for each backend, or compile from source for custom configurations.
To compile from source, simply install the package from source using `pip`. The package will be built for CPU only at this time.

#### Pre-built Wheel Installation (recommended)[[multi-backend-pip]]
```bash
git clone https://github.com/bitsandbytes-foundation/bitsandbytes.git && cd bitsandbytes/
pip install -e .
```

<hfoptions id="platform">
<hfoption id="Linux">
This wheel provides support for ROCm and Intel XPU platforms.
## AMD ROCm (Preview)[[rocm]]

```
# Note, if you don't want to reinstall our dependencies, append the `--no-deps` flag!
pip install --force-reinstall 'https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_multi-backend-refactor/bitsandbytes-0.44.1.dev0-py3-none-manylinux_2_24_x86_64.whl'
```
* A compatible PyTorch version with AMD ROCm support is required. It is recommended to use the latest stable release. See [PyTorch on ROCm](https://rocm.docs.amd.com/projects/install-on-linux/en/latest/install/3rd-party/pytorch-install.html) for guidance.
* ROCm support is currently only available in our preview wheels or when building from source.

</hfoption>
<hfoption id="Windows">
This wheel provides support for the Intel XPU platform.
### Preview Wheels from `main`[[rocm-preview]]

```bash
# Note, if you don't want to reinstall our dependencies, append the `--no-deps` flag!
pip install --force-reinstall 'https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_multi-backend-refactor/bitsandbytes-0.44.1.dev0-py3-none-win_amd64.whl'
```
The currently distributed preview `bitsandbytes` are built with the following configurations:

</hfoption>
</hfoptions>
| **OS** | **ROCm** | **Targets**
|--------------------|----------|---------------------------|
| **Linux x86-64** | 6.1.2 | gfx90a / gfx942 / gfx1100
| **Linux x86-64** | 6.2.4 | gfx90a / gfx942 / gfx1100
| **Linux x86-64** | 6.3.4 | gfx90a / gfx942 / gfx1100
| **Linux x86-64** | 6.4.4 | gfx90a / gfx942 / gfx1100
| **Linux x86-64** | 7.0.0 | gfx90a / gfx942 / gfx1100

#### Compile from Source[[multi-backend-compile]]
**Windows is not currently supported.**

<hfoptions id="backend">
<hfoption id="AMD ROCm">
Please see [Preview Wheels](#preview-wheels) for installation instructions.

#### AMD GPU
### Compile from Source[[rocm-compile]]

bitsandbytes is supported from ROCm 6.1 - ROCm 6.4.
bitsandbytes can be compiled from ROCm 6.1 - ROCm 7.0.

```bash
# Install bitsandbytes from source
# Clone bitsandbytes repo, ROCm backend is currently enabled on multi-backend-refactor branch
git clone -b multi-backend-refactor https://github.com/bitsandbytes-foundation/bitsandbytes.git && cd bitsandbytes/
# Clone bitsandbytes repo
git clone https://github.com/bitsandbytes-foundation/bitsandbytes.git && cd bitsandbytes/

# Compile & install
apt-get install -y build-essential cmake # install build tools dependencies, unless present
Expand All @@ -235,38 +224,29 @@ make
pip install -e . # `-e` for "editable" install, when developing BNB (otherwise leave that out)
```

</hfoption>
<hfoption id="Intel CPU + GPU">
## Preview Wheels[[preview-wheels]]

#### Intel CPU + GPU(XPU)
If you would like to use new features even before they are officially released and help us test them, feel free to install the wheel directly from our CI (*the wheel links will remain stable!*):

CPU needs to build CPU C++ codes, while XPU needs to build sycl codes.
Run `export bnb_device=xpu` if you are using xpu, run `export bnb_device=cpu` if you are using cpu.
```
git clone https://github.com/bitsandbytes-foundation/bitsandbytes.git && cd bitsandbytes/
cmake -DCOMPUTE_BACKEND=$bnb_device -S .
make
pip install -e .
```
<hfoptions id="OS">
<hfoption id="Linux">

```bash
# Note: if you don't want to reinstall our dependencies, append the `--no-deps` flag!

</hfoption>
<hfoption id="Ascend NPU">
# x86_64 (most users)
pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-manylinux_2_24_x86_64.whl

#### Ascend NPU
# ARM/aarch64
pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-manylinux_2_24_aarch64.whl
```

Please refer to [the official Ascend installations instructions](https://www.hiascend.com/document/detail/zh/Pytorch/60RC3/configandinstg/instg/insg_0001.html) for guidance on how to install the necessary `torch_npu` dependency.
</hfoption>
<hfoption id="Windows">

```bash
# Install bitsandbytes from source
# Clone bitsandbytes repo, Ascend NPU backend is currently enabled on multi-backend-refactor branch
git clone -b multi-backend-refactor https://github.com/bitsandbytes-foundation/bitsandbytes.git && cd bitsandbytes/

# Compile & install
apt-get install -y build-essential cmake # install build tools dependencies, unless present
cmake -DCOMPUTE_BACKEND=npu -S .
make
pip install -e . # `-e` for "editable" install, when developing BNB (otherwise leave that out)
# Note: if you don't want to reinstall our dependencies, append the `--no-deps` flag!
pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-win_amd64.whl
```
</hfoption>
</hfoptions>