beclab · fnalways · Mar 3, 2026 · fnalways · Mar 3, 2026
diff --git a/docs/manual/get-started/gpu-requirements.md b/docs/manual/get-started/gpu-requirements.md
@@ -0,0 +1,38 @@
+---
+search: false
+---
+
+A GPU is not required to install Olares, but is necessary for most AI applications. Only NVIDIA GPUs are supported.
+
+- **Architecture**: Turing or newer (GTX 16xx, RTX 20xx, 30xx, 40xx, 50xx series, and later).
+:::info
+GPUs with older architectures are not recognized by Olares. AI applications that require GPU access will not run.
+:::
+- **VRAM**: At least 8 GB recommended. Even on a supported GPU, limited VRAM will prevent many AI applications from running.
+
+:::details Not sure if your GPU is supported?
+Run the following command and check the code name prefix in the output:
+
+```bash
+lspci | grep -i nvidia
+```
+
+Example output:
+
+```
+3b:00.0 VGA compatible controller: NVIDIA Corporation AD102 [GeForce RTX 4090] (rev a1)
+```
+
+The code name prefix maps to GPU architecture as follows:
+
+| Code name prefix | Architecture | Supported |
+|:---|:---|:---:|
+| GB | Blackwell | ✓ |
+| AD | Ada Lovelace | ✓ |
+| GA | Ampere | ✓ |
+| TU | Turing | ✓ |
+| GP | Pascal | ✗ |
+| GM | Maxwell | ✗ |
+
+You can also verify against the [full compatible GPU list](https://github.com/NVIDIA/open-gpu-kernel-modules?tab=readme-ov-file#compatible-gpus) from the NVIDIA open driver repository.
+:::
diff --git a/docs/manual/get-started/install-linux-docker.md b/docs/manual/get-started/install-linux-docker.md
@@ -5,9 +5,7 @@ description: Learn how to deploy Olares on a Linux server using Docker Compose.
 # Install Olares on Linux using Docker Compose
 You can use Docker to install and run Olares in a containerized environment. This guide walks you through setting up Olares with Docker, preparing the installation environment, completing the activation process, and managing the container lifecycle.
 
-:::tip Recommendation for production use
-For best performance and stability, we recommend [installing Olares on Linux via script](/manual/get-started/install-olares.md).
-:::
+<!--@include: ./reusables.md{44,51}-->
 
 ## System requirements
 
@@ -31,11 +29,7 @@ While these specific versions are confirmed to work, the process may still work
 
 ### Optional hardware
 
-A GPU is not required to install Olares, but is necessary for AI applications.
-
-- **GPU (NVIDIA only)**:
-  - **Architecture**: Turing or newer (e.g., GTX 16 series, RTX 20 series).
-  - **Verification**: Run `lspci | grep -i nvidia` and check the [compatible GPU table](https://github.com/NVIDIA/open-gpu-kernel-modules?tab=readme-ov-file#compatible-gpus).
+<!--@include: ./gpu-requirements.md{5,}-->
 
 ## Before you begin
 Before you begin, ensure the following:
@@ -75,8 +69,8 @@ cd ~/olares-config
     sudo dpkg -i --force-all /tmp/keyring.deb
 
     sudo apt update
-    sudo apt install nvidia-kernel-open-570
-    sudo apt install nvidia-driver-570
+    sudo apt install nvidia-kernel-open-590
+    sudo apt install nvidia-driver-590
     ````
 
 2. Install the NVIDIA Container Toolkit to enable Docker to access your GPU. 
@@ -114,7 +108,7 @@ cd ~/olares-config
 
     ```
     +-----------------------------------------------------------------------------------------+
-    | NVIDIA-SMI 570.169                Driver Version: 570.169        CUDA Version: 12.8     |
+    | NVIDIA-SMI 590.44                 Driver Version: 590.44         CUDA Version: 13.1     |
     |-----------------------------------------+------------------------+----------------------+
     | GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
     | Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |

diff --git a/docs/manual/get-started/install-linux-iso.md b/docs/manual/get-started/install-linux-iso.md
@@ -9,27 +9,24 @@ This guide explains how to install Olares on a physical machine using the offici
 
 <!--@include: ./reusables.md{44,51}-->
 
-## Prerequisites
-
-### Required
-- **Host requirements**:
-  - **CPU**: Minimum 4 cores with **x86-64 architecture** (Intel or AMD). ARM-based processors are not currently supported for this method.
-  - **Memory**: At least 8 GB of available RAM.
-  - **Storage**: Minimum 150 GB SSD.
-  - **Network**: The device must be connected to a wired LAN.
-  ::: warning SSD required
-  The installation will likely fail if an HDD (mechanical hard drive) is used instead of an SSD.
-  :::
+## System requirements
+
+Make sure your device meets the following requirements.
 
-- **USB flash drive**: **8 GB** capacity or larger.
+### Required specifications
 
-### Optional
+- **CPU**: At least 4 cores. Intel or AMD x86-64 architecture required. ARM is not currently supported.
+- **RAM**: At least 8 GB of available memory.
+- **Storage**: At least 150 GB of available SSD storage.
+  :::warning SSD required
+  The installation will fail if an HDD (mechanical hard drive) is used instead of an SSD.
+  :::
+- **Network**: Wired LAN connection required.
+- **USB flash drive**: 8 GB capacity or larger.
 
-A GPU is not required to install Olares, but is necessary for AI applications.
+### Optional hardware
 
-- **GPU (NVIDIA only)**:
-  - **Architecture**: Turing or newer (e.g., GTX 16 series, RTX 20 series).
-  - **Verification**: Run `lspci | grep -i nvidia` and check the [compatible GPU table](https://github.com/NVIDIA/open-gpu-kernel-modules?tab=readme-ov-file#compatible-gpus).
+<!--@include: ./gpu-requirements.md{5,}-->
 
 ## Create a bootable USB drive
 

diff --git a/docs/manual/get-started/install-linux-script.md b/docs/manual/get-started/install-linux-script.md
@@ -29,11 +29,7 @@ While these specific versions are confirmed to work, the process may still work
 
 ### Optional hardware
 
-A GPU is not required to install Olares, but is necessary for AI applications.
-
-- **GPU (NVIDIA only)**:
-  - **Architecture**: Turing or newer (e.g., GTX 16 series, RTX 20 series).
-  - **Verification**: Run `lspci | grep -i nvidia` and check the [compatible GPU table](https://github.com/NVIDIA/open-gpu-kernel-modules?tab=readme-ov-file#compatible-gpus).
+<!--@include: ./gpu-requirements.md{5,}-->
 
 ## Install Olares
 

diff --git a/docs/manual/get-started/install-lxc.md b/docs/manual/get-started/install-lxc.md
@@ -4,24 +4,27 @@ description: Instructions for installing Olares on Linux Containers (LXC) includ
 # Install Olares on LXC
 LXC (Linux Containers) is a lightweight virtualization method that runs applications in isolated containers. When used on PVE, it enables an efficient way to deploy Olares without the overhead of a full virtual machine.
 
-:::warning Not recommended for production use
+:::warning Not for production use
 Currently, Olares on LXC has certain limitations. We recommend using it only for development or testing purposes.
 :::
 
 <!--@include: ./reusables.md{44,51}-->
 
 ## System requirements
+
 Make sure your device meets the following requirements.
+
 ### Required specifications
-- CPU: At least 4 cores
-- RAM: At least 8 GB of available memory
-- Storage: At least 150 GB of available SSD storage.
-    ::: warning SSD required
-    The installation will likely fail if an HDD (mechanical hard drive) is used instead of an SSD.
-    :::
-- Supported systems:
-    - PVE 8.2.2
-    - Linux container: Debian 12 (for existing LXC containers on PVE)
+
+- **CPU**: At least 4 cores.
+- **RAM**: At least 8 GB of available memory.
+- **Storage**: At least 150 GB of available SSD storage.
+  :::warning SSD required
+  The installation will fail if an HDD (mechanical hard drive) is used instead of an SSD.
+  :::
+- **Supported systems**:
+  - PVE 8.2.2
+  - Linux container: Debian 12 (for existing LXC containers on PVE)
 
 
 :::info Version compatibility
@@ -30,12 +33,11 @@ While the specific versions are confirmed to work, the process may still work on
 
 ### Optional hardware
 
-A GPU is not required to install Olares, but is necessary for AI applications.
+<!--@include: ./gpu-requirements.md{5,}-->
 
-- GPU (NVIDIA only):
-   - Architecture: Turing or newer (e.g., GTX 16 series, RTX 20 series, and later).
-   - Verification: Run `lspci | grep -i nvidia` in the PVE host shell to confirm the card is detected.
-   - Setup: To utilize the GPU, you must configure LXC device passthrough. Please refer to [Configure GPU passthrough in PVE](/manual/best-practices/install-olares-gpu-passthrough.md#configure-gpu-passthrough-in-pve) for detailed instructions.
+:::tip LXC device passthrough required
+To use the GPU within Olares on LXC, you must configure device passthrough first. Refer to [Configure GPU passthrough in PVE](/manual/best-practices/install-olares-gpu-passthrough.md#configure-gpu-passthrough-in-pve) for instructions.
+:::
 
 ## Prerequisites
 

diff --git a/docs/manual/get-started/install-mac-docker.md b/docs/manual/get-started/install-mac-docker.md
@@ -16,18 +16,15 @@ We recommend using it only for development or testing purposes.
 <!--@include: ./reusables.md{44,51}-->
 
 ## System requirements
+
 Make sure your device meets the following requirements.
 
-- Architecture: AMD64 or ARM64
-- CPU: At least 4 cores
-- RAM: At least 8 GB of available memory
-- Storage: At least 150 GB of available SSD storage. 
-::: warning SSD required
-The installation will fail if an HDD (mechanical hard drive) is used instead of an SSD.
-:::
-::: info GPU limitation
-Olares GPU acceleration currently supports NVIDIA GPUs only. Consequently, GPU resources cannot be managed or utilized for AI workloads on macOS devices.
-:::
+### Required specifications
+
+- **Architecture**: x86-64 or ARM64.
+- **CPU**: At least 4 cores.
+- **RAM**: At least 8 GB of available memory.
+- **Storage**: At least 150 GB of available storage.
 
 ## Before you begin
 Before you begin, ensure the following:
@@ -47,7 +44,7 @@ If you are unsure which chip your Mac is using, go to the Apple menu and select
 <tabs>
 <template #Apple-Silicon>
 
-1. Download `olaresd-proxy` via the link: https://dc3p1870nn3cj.cloudfront.net/olaresd-proxy-v0.1.0-darwin-arm64.tar.gz .
+1. Download [`olaresd-proxy`](https://dc3p1870nn3cj.cloudfront.net/olaresd-proxy-v0.1.0-darwin-arm64.tar.gz).
 2. Unzip the file, then start `olaresd-proxy`.
    :::info Keep `olaresd-proxy` running in the background
    During Olares installation and activation, keep `olaresd-proxy` running in the background.

diff --git a/docs/manual/get-started/install-mac-script.md b/docs/manual/get-started/install-mac-script.md
@@ -14,19 +14,17 @@ We recommend using it only for development or testing purposes.
 
 <!--@include: ./reusables.md{44,51}-->
 
-## System compatibility
+## System requirements
+
 Make sure your Mac meets the following requirements.
-- Architecture: X86-64 or ARM64.
-- RAM: At least 8 GB of available memory.
-- Storage: At least 150 GB of available SSD storage.
-    ::: warning SSD required
-    The installation will likely fail if an HDD (mechanical hard drive) is used instead of an SSD.
-    :::
-- MacOS: Monterey (12) or later.
 
-::: info GPU limitation
-Olares GPU acceleration currently supports NVIDIA GPUs only. Consequently, GPU resources cannot be managed or utilized for AI workloads on macOS devices.
-:::
+### Required specifications
+
+- **Architecture**: x86-64 or ARM64.
+- **CPU**: At least 4 cores.
+- **RAM**: At least 8 GB of available memory.
+- **Storage**: At least 150 GB of available storage.
+- **macOS**: Monterey 12 or later.
 
 ## Before you begin
 Ensure you have the following installed:

diff --git a/docs/manual/get-started/install-olares.md b/docs/manual/get-started/install-olares.md
@@ -7,9 +7,10 @@ outline: [2,4]
 
 This page provides an overview of supported installation methods for Olares.
 
-Before installation, make sure you have:
-- Created your [Olares ID](create-olares-id.md).
-- Verified your operating system and hardware meet the minimum requirements as described in the specific guide.
+## Before you begin
+
+- Create your [Olares ID](create-olares-id.md) if you haven't already.
+- Check that your operating system and hardware meet the minimum requirements described in the specific installation guide.
 
 ## Choosing the right method
 
@@ -19,16 +20,13 @@ Olares supports multiple platforms and deployment methods. Choose the installati
 
 Linux (Ubuntu or Debian) is the recommended platform for running Olares, as it offers the best performance and stability in production environments.
 
-- [**ISO image**](install-linux-iso.md): Perform a fresh installation on a physical machine, auto<br/> configuring the Linux host environment, container runtime,<br/> drivers, and core dependencies.
-- [**One-line script**](install-linux-script.md): Quick install on existing Linux systems.                                   
-
-:::tip Recommendation
-The ISO installation method ensures maximum compatibility, performance, and system-level optimization.
-:::
+- [**ISO image**](install-linux-iso.md) (Recommended): Perform a fresh installation on a physical machine, auto-configuring the Linux host environment, container runtime, drivers, and core dependencies.
+- [**One-line script**](install-linux-script.md): Quick install on existing Linux systems.
+- [**Docker image**](install-linux-docker.md): Run Olares in a containerized environment on Linux.
 
 ### Alternative installation methods
 
-These methods are suitable for **development**, **testing**, or **lightweight environments**. 
+These methods are suitable for development, testing, or lightweight environments. 
 
 #### Windows
 - [**One-line script**](install-windows-script.md): Install Olares in Windows Subsystem for Linux 2 (WSL 2).

diff --git a/docs/manual/get-started/install-pve-iso.md b/docs/manual/get-started/install-pve-iso.md
@@ -4,31 +4,33 @@ description: Guide to installing Olares on Proxmox VE (PVE) using ISO image with
 # Install Olares on PVE with ISO image
 You can install Olares directly on Proxmox Virtual Environment (PVE) using an ISO image. This guide walks you through downloading the Olares ISO, configuring PVE environment, completing the installation, and getting your Olares up and running.
 
-:::warning Not recommended for production use
+:::warning Not for production use
 Currently, Olares on PVE has certain limitations. We recommend using it only for development or testing purposes.
 :::
 
 <!--@include: ./reusables.md{44,51}-->
 
 ## System requirements
+
 Make sure your device meets the following requirements.
+
 ### Required specifications
-- CPU: Minimum 4 cores with x86-64 architecture (Intel or AMD). ARM-based processors are not currently supported for this method.
-- RAM: At least 8 GB of available memory
-- Storage: At least 200 GB of available SSD storage. 
-    ::: warning SSD required
-    The installation will likely fail if an HDD (mechanical hard drive) is used instead of an SSD.
-    :::
-- Supported Systems: PVE 8.2.2
+
+- **CPU**: At least 4 cores. Intel or AMD x86-64 architecture required. ARM is not currently supported.
+- **RAM**: At least 8 GB of available memory.
+- **Storage**: At least 200 GB of available SSD storage.
+  :::warning SSD required
+  The installation will likely fail if an HDD (mechanical hard drive) is used instead of an SSD.
+  :::
+- **Supported systems**: PVE 8.2.2
 
 ### Optional hardware
 
-A GPU is not required to install Olares, but is necessary for AI applications.
+<!--@include: ./gpu-requirements.md{5,}-->
 
-- GPU (NVIDIA only):
-   - Architecture: Turing or newer (e.g., GTX 16 series, RTX 20 series, and later).
-   - Verification: Run `lspci | grep -i nvidia` in the PVE host shell to confirm the card is detected.
-   - Setup: To utilize the GPU, you must configure PCI passthrough. Please refer to [Configure GPU passthrough in PVE](/manual/best-practices/install-olares-gpu-passthrough.md#configure-gpu-passthrough-in-pve) for detailed instructions.
+:::tip PCI passthrough required
+To use the GPU within Olares on PVE, you must configure PCI passthrough first. Refer to [Configure GPU passthrough in PVE](/manual/best-practices/install-olares-gpu-passthrough.md#configure-gpu-passthrough-in-pve) for instructions.
+:::
 
 ## Download Olares ISO image
 Click [here](https://cdn.olares.com/olares-latest-amd64.iso) to download the official Olares ISO image.

diff --git a/docs/manual/get-started/install-pve-script.md b/docs/manual/get-started/install-pve-script.md
@@ -4,36 +4,37 @@ description: Guide to installing Olares on Proxmox VE (PVE) with system requirem
 # Install Olares on PVE via the script
 Proxmox Virtual Environment (PVE) is an open-source virtualization platform based on Debian Linux. This guide explains how to install Olares in a PVE environment using the provided installation script.
 
-:::warning Not recommended for production use
+:::warning Not for production use
 Currently, Olares on PVE has certain limitations. We recommend using it only for development or testing purposes.
 :::
 
 <!--@include: ./reusables.md{44,51}-->
 
 ## System requirements
+
 Make sure your device meets the following requirements.
 
 ### Required specifications
-- CPU: At least 4 cores
-- RAM: At least 8 GB of available memory
-- Storage: At least 150 GB of available SSD storage. 
-    ::: warning SSD required
-    The installation will likely fail if an HDD (mechanical hard drive) is used instead of an SSD.
-    :::
-- Supported Systems: PVE 8.2.2
+
+- **CPU**: At least 4 cores.
+- **RAM**: At least 8 GB of available memory.
+- **Storage**: At least 150 GB of available SSD storage.
+  :::warning SSD required
+  The installation will fail if an HDD (mechanical hard drive) is used instead of an SSD.
+  :::
+- **Supported systems**: PVE 8.2.2
 
 :::info Version compatibility
 While the specific version is confirmed to work, the process may still work on other versions. Adjustments may be necessary depending on your environment. If you meet any issues with these platforms, feel free to raise an issue on [GitHub](https://github.com/beclab/Olares/issues/new).
 :::
 
 ### Optional hardware
 
-A GPU is not required to install Olares, but is necessary for AI applications.
+<!--@include: ./gpu-requirements.md{5,}-->
 
-- GPU (NVIDIA only):
-   - Architecture: Turing or newer (e.g., GTX 16 series, RTX 20 series, and later).
-   - Verification: Run `lspci | grep -i nvidia` in the PVE host shell to confirm the card is detected.
-   - Setup: To utilize the GPU, you must configure PCI passthrough. Please refer to [Configure GPU passthrough in PVE](/manual/best-practices/install-olares-gpu-passthrough.md#configure-gpu-passthrough-in-pve) for detailed instructions.
+:::tip PCI passthrough required
+To use the GPU within Olares on PVE, you must configure PCI passthrough first. Refer to [Configure GPU passthrough in PVE](/manual/best-practices/install-olares-gpu-passthrough.md#configure-gpu-passthrough-in-pve) for instructions.
+:::
 
 ## Install on PVE