# AMD Strix Halo RDMA Cluster Setup Guide

This guide details how to configure a two-node **AMD Strix Halo** cluster linked via **Intel E810 (RoCE v2)** for distributed vLLM inference using Tensor Parallelism.

## Table of Contents

1. [TL;DR (Quick Start)](#1-tldr-quick-start)
2. [Concepts & Architecture](#2-concepts--architecture)
3. [Hardware Prerequisites](#3-hardware-prerequisites)
4. [Host Configuration (Fedora)](#4-host-configuration-fedora)
    *   [4.1 Install Packages](#41-install-packages)
    *   [4.2 Check Native Firmware](#42-check-native-firmware)
    *   [4.3 Network Configuration](#43-network-configuration)
    *   [4.4 BIOS & Kernel Configuration](#44-bios--kernel-configuration)
    *   [4.5 Firewall Rules](#45-firewall-rules)
5. [Toolbox Installation & Network Verification](#5-toolbox-installation--network-verification)
    *   [5.1 Prerequisites: Passwordless SSH](#51-prerequisites-passwordless-ssh)
    *   [5.2 Installation](#52-installation)
    *   [5.3 Verify RDMA Connection](#53-verify-rdma-connection)
6. [Running the Cluster](#6-running-the-cluster)
    *   [6.1 Setup & Verify](#61-setup--verify)
    *   [6.2 Launching vLLM](#62-launching-vllm)
7. [Troubleshooting](#7-troubleshooting)
8. [References & Acknowledgements](#8-references--acknowledgements)

---

## 1. TL;DR (Quick Start)

**On Both Nodes:**
1.  **Preparation**:
    *   **Install/Update Fedora 43** and the E810 NICs (Check firmware: `ethtool -i <iface>`).
    *   **BIOS/Kernel**: Set iGPU to 512MB and apply kernel params (`iommu=pt`, `pci=realloc`, etc.).
    *   **SSH**: Configure **passwordless SSH** between nodes.
2.  **Networking**: Assign static IPs (`192.168.100.1` & `.2`), set MTU 9000, and trust the interface in firewall.
3.  **Install Toolbox**: Run `./refresh_toolbox.sh` (this automatically installs the container with RDMA support and the custom `librccl.so` patch).
4.  **Run Cluster**:
    *   Run `start-vllm-cluster`.
    *   Select **"2. Start Ray Cluster"** (Follow prompts using the TUI).
    *   Select **"4. Launch VLLM Serve"** and choose your model. (Export `HF_TOKEN` first for gated models!)

**Key Note**: The `refresh_toolbox.sh` script detects your Infiniband/RDMA devices and automatically configures the container to expose them.

---

## 2. Concepts & Architecture

![concepts](concepts.png)

To fully utilize the Strix Halo cluster, it is helpful to understand the technologies involved:

*   **vLLM**: A high-performance inference engine. To run models larger than a single GPU (or APU) can handle, it splits the model using **Tensor Parallelism (TP)**.
*   **Ray**: A distributed computing framework. vLLM uses Ray to **orchestrate** the cluster, manage the "worker" processes on each node, and ensure they start up correctly. Ray handles the *control plane* (issuing commands).
*   **RCCL (ROCm Collective Communication Library)**: The AMD equivalent of NVIDIA's NCCL. This library handles the **data plane**—specifically, the extremely fast synchronization of tensor data between GPUs. When TP=2, the two nodes must exchange partial results after *every single layer* of the neural network. This happens thousands of times per second.
*   **RoCE v2 (RDMA over Converged Ethernet)**: The protocol that allows RCCL to write data directly from one Node's memory to the other Node's memory, bypassing the CPU and OS kernel.
    *   **Without RDMA**: Latency is ~70-100µs (TCP/IP overhead).
    *   **With RDMA**: Latency is ~5µs.
    *   **Why it matters**: For interactive token generation, high latency kills performance. RoCE makes the two nodes feel like a single machine.


---

## 3. Hardware Prerequisites

![cluster](cluster.png)


*   **Nodes**: 2x [Framework Desktop Mainboards](https://frame.work/gb/en/products/framework-desktop-mainboard-amd-ryzen-ai-max-300-series?v=FRAFMK0006) with AMD Ryzen AI MAX+ "Strix Halo", 128GB of Unified Memory.
*   **Network Cards**: [Intel Ethernet Controller E810-CQDA1](https://www.intel.com/content/www/us/en/products/sku/192558/intel-ethernet-network-adapter-e810cqda1/specifications.html) (or similar 100GbE QSFP28).
*   **Connection**: Direct Attach Copper (DAC) cable (e.g., [QSFPTEK 100G QSFP28 DAC](https://www.amazon.co.uk/dp/B09F32F7VK)). No switch required for 2 nodes.
*   **PCIe Note**: The Framework motherboard PCIe slot is physically **x4**, so a riser is required to plug in a 16x card (e.g., [CY PCI-E Express 4x to 16x Extender](https://www.amazon.co.uk/dp/B0837FZFJ6)). **Test Setup Note:** One of the boards in this setup has a modified PCIe slot (cut by Framework using an ultrasonic knife) to accept x16 cards directly. **This is not recommended for users.** Risers are the cheaper, safer, and easier solution. Performance is identical (~50Gbps bandwidth, ~5µs latency).


---

## 4. Host Configuration (Fedora)

Perform these steps on the **Host OS** (Fedora 43) of **both nodes**.

**Tested Host Configuration:**

| Node | Kernel | OS | IP (RDMA Interface) |
| :--- | :--- | :--- | :--- |
| **Node 1** | `6.18.5-200.fc43.x86_64` | Fedora Linux 43 | `192.168.100.1/30` |
| **Node 2** | `6.18.6-200.fc43.x86_64` | Fedora Linux 43 | `192.168.100.2/30` |

> **Note:** These specific kernel versions were verified to work. Fedora 43 is recommended.

### 4.1 Install Packages

Install the core RDMA userspace tools. You do **not** need proprietary Intel drivers; the in-kernel drivers work perfectly.
*   **Ethernet Driver:** `ice`
*   **RDMA Driver:** `irdma` (Unified driver for RoCE v2 & iWARP)

```bash
sudo dnf install rdma-core libibverbs-utils perftest
```

*   `rdma-core`: The userspace components for the RDMA subsystem (libraries, daemons, and configuration tools).
*   `libibverbs-utils`: Utilities for querying RDMA devices (e.g., `ibv_devinfo`).
*   `perftest`: A suite of benchmarks (e.g., `ib_write_bw`, `ib_send_lat`) to verify RDMA bandwidth and latency.

### 4.2 Check Native Firmware

Use `ethtool` to check the current firmware version of your Intel E810 card.

```bash
ethtool -i enp194s0np0
```

**Recommended Firmware:**
Ensure your firmware is at least as new as the version shown below (Firmware `4.91...`). If your firmware is older, please update it using the [Intel® Ethernet NVM Update Tool for E810 Series](https://www.intel.com/content/www/us/en/download/19624/non-volatile-memory-nvm-update-utility-for-intel-ethernet-network-adapter-e810-series-linux.html).

**Example Output:**
```text
driver: ice
version: 6.18.5-200.fc43.x86_64
firmware-version: 4.91 0x800214b5 1.3909.0
expansion-rom-version: 
bus-info: 0000:c2:00.0
supports-statistics: yes
supports-test: yes
supports-eeprom-access: yes
supports-register-dump: yes
supports-priv-flags: yes
```

### 4.3 Network Configuration

This guide assumes a subnet of `192.168.100.0/30`.

**Identify your interface**:
Run `ip link` to find your 100GbE card (e.g., `enp194s0np0`).

**Node 1 (Head - 192.168.100.1):**
```bash
# Bring link up
sudo ip link set enp194s0np0 up

# Assign IP
sudo ip addr add 192.168.100.1/30 dev enp194s0np0

# Set MTU (Jumbo Frames)
sudo nmcli connection modify "rdma0" ethernet.mtu 9000
sudo nmcli connection up "rdma0"
```

**Node 2 (Worker - 192.168.100.2):**
```bash
# Bring link up
sudo ip link set enp194s0np0 up

# Assign IP
sudo ip addr add 192.168.100.2/30 dev enp194s0np0

# Set MTU
sudo nmcli connection modify "rdma0" ethernet.mtu 9000
sudo nmcli connection up "rdma0"
```

**Verify Routing:**
Ensure the route exists on both:
```bash
sudo ip route add 192.168.100.0/30 dev enp194s0np0
```

**Verify Link:**
```bash
rdma link
# Output should show: state ACTIVE physical_state LINK_UP used_usec X ...
```

### 4.4 BIOS & Kernel Configuration

**1. BIOS Settings:**
Set the **iGPU Memory Allocation** to the **minimum possible (512MB)**. We will use the GTT (Graphics Translation Table) to dynamically allocate system memory as "Unified Memory" for the GPU.

**2. Kernel Parameters:**
Update GRUB to enable unified memory, optimize RDMA performance, and fix PCI resource allocation.

Edit `/etc/default/grub` and append to `GRUB_CMDLINE_LINUX`:
```text
iommu=pt pci=realloc pcie_aspm=off amdgpu.gttsize=126976 ttm.pages_limit=32505856
```

**Explanation of Parameters:**
*   `iommu=pt`: Sets IOMMU to "Pass-Through" mode. This is critical for performance, reducing overhead for both the RDMA NIC and the iGPU unified memory access.
*   `pci=realloc`: Reallocates PCI BARs. Often needed on consumer platforms to properly map large address spaces for devices like the E810 or Strix Halo.
*   `pcie_aspm=off`: Disables PCIe Active State Power Management. Prevents latency spikes and link negotiation issues on the 100GbE connection.
*   `amdgpu.gttsize=126976`: Caps the GPU GTT size to ~124GiB (126976MB). This defines how much system RAM the GPU can address as its own "VRAM".
*   `ttm.pages_limit=32505856`: Limits the Translation Table Manager to ~124GiB (in 4KB pages), matching the GTT size.

**3. Apply Changes:**
```bash
sudo grub2-mkconfig -o /boot/grub2/grub.cfg
sudo reboot
```

### 4.5 Firewall Rules

Applications like Ray and NCCL use random high ports. It is easiest to trust the internal RDMA interface completely.

```bash
# Assign the interface to the trusted zone permanently
sudo firewall-cmd --permanent --zone=trusted --add-interface=enp194s0np0

# Reload firewall
sudo firewall-cmd --reload
```

---

## 5. Toolbox Installation & Network Verification

### 5.1 Prerequisites: Passwordless SSH

The cluster management and verification scripts rely on SSH to execute commands on remote nodes. You must configure **passwordless SSH** between both nodes (root or sudo-enabled user).

*   **Guide:** [How to Set Up SSH Keys on Linux (DigitalOcean)](https://www.digitalocean.com/community/tutorials/how-to-set-up-ssh-keys-on-ubuntu-20-04)
*   **Quick Check:** Run `ssh <other-node-ip> date` from each node. It should print the date without asking for a password.

### 5.2 Installation

The toolbox container provided in this repo includes a **critical patch**: a custom-built `librccl.so` that enables `gfx1151` (Strix Halo) support for RDMA (https://github.com/kyuz0/rocm-systems/tree/gfx1151-rccl), which is currently missing in upstream ROCm packages. This library is automatically compiled using the [`build-rccl`](../.github/workflows/build-rccl.yml) GitHub Action in this repository, which generates the artifact that is then bundled into the Docker container.

To install the toolbox on **both nodes**, run:

```bash
./refresh_toolbox.sh
```

**What this does:**
1.  Pulls the latest `kyuz0/vllm-therock-gfx1151` image.
2.  Detects if `/dev/infiniband` exists on your host.
3.  Creates the toolbox with flags to expose:
    *   **iGPU Access**: `/dev/dri`, `/dev/kfd` (Required for ROCm)
    *   **RDMA Access**: `/dev/infiniband`, `--group-add rdma`
    *   **Memory Pinning**: `--ulimit memlock=-1` (Required for DMA)

### 5.3 Verify RDMA Connection

Before proceeding to run the cluster, verify that RDMA is active and providing low latency (~5µs vs ~70µs for Ethernet).

Run the provided verification script from the **Head Node**:

```bash
# Inside toolbox
/opt/compare_eth_vs_rdma.sh
```

**Expected Results:**
```text
Path                 Latency      Bandwidth   
------------------------------------------------
Ethernet (1G LAN)    0.074 ms     0.94 Gbps   
Ethernet (RoCE NIC)  0.068 ms     55.70 Gbps  
RDMA (RoCE)          5.23 us      50.64 Gbps  
```
*Note the massive latency drop (milliseconds to microseconds) for RDMA.*

---

## 6. Running the Cluster

A TUI utility, `start-vllm-cluster`, is provided to manage the Ray cluster and vLLM.

### 6.1 Setup & Verify

1.  **Enter the toolbox**:
    ```bash
    toolbox enter vllm
    ```
2.  **Run the Cluster Manager**:
    ```bash
    start-vllm-cluster
    ```
3.  **Configure IPs** (Option 1):
    *   Ensure Head is `192.168.100.1` and Worker is `192.168.100.2`.
4.  **Start Ray Cluster** (Option 2):
    *   **On Node 1**: Select **"Head"** when prompted.
    *   **On Node 2**: Select **"Worker"** when prompted.
    *   The script effectively runs:
        ```bash
        # Head
        export NCCL_SOCKET_IFNAME=<rdma_iface>
        ray start --head --node-ip-address=192.168.100.1 ...
        
        # Worker
        ray start --address=192.168.100.1:6379 ...
        ```
5.  **Check Status** (Option 3):
    *   Ensure you see **2 nodes** and adequate GPU resources (e.g., `2.0 GPU`).

### 6.2 Launching vLLM

Once the cluster is active (checked via Option 3):

1.  Select **"4. Launch VLLM Serve"** in the TUI.
2.  Choose a model (e.g., `Meta-Llama-3.1-8B-Instruct`).
3.  **Configuration Menu**:
    *   **Tensor Parallelism**: Set to `2` (one GPU per node).
    *   **Context Length**: Auto or custom (e.g., `131072`).
    *   **Erase vLLM Cache**: Select `YES` if you are restarting after a crash.
    *   **Force Eager Mode**: Select `YES`.
        *   *Why?* CUDA Graphs can be unstable on distributed APU clusters and cause deadlocks. Eager mode is safer, but you might be able to squeeze 1-3% more performance if you take a chance and disable it.
4.  **Launch**: Select "LAUNCH SERVER".

**Important Gotchas:**
*   **First Run Download**: When running a model for the first time, each node in the cluster must download the weights independently. This may take some time depending on your internet connection.
*   **Gated Models (e.g., Gemma)**:
    *   Models like `google/gemma-2-27b-it` are "gated" and require you to request access on Hugging Face.
    *   You must export your Hugging Face token before running the cluster script:
        ```bash
        export HF_TOKEN=your_token_here
        start-vllm-cluster
        ```
    *   If you don't provide a token or haven't accepted the license on Hugging Face, the download will fail.

---

## 7. Troubleshooting

### vLLM Deadlocks / Hangs
*   **Cause**: CUDA Graph capture can freeze on distributed APU nodes.
*   **Fix**: Enable **"Force Eager Mode"** in the start menu.

### Firmware
If you see link issues, ensure your Intel E810 firmware is up to date using the Intel standard tools.

---

## 8. References & Acknowledgements

*   **Reddit - Strix Halo Batching with Tensor Parallel**: [Thread by Hungry_Elk_3276](https://www.reddit.com/r/LocalLLaMA/comments/1p8nped/strix_halo_batching_with_tensor_parallel_and/)
    *   Special thanks to user **Hungry_Elk_3276** for their initial experiments with vLLM RDMA, which highlighted the missing `gfx1151` support in upstream RCCL.