Removes source compilation of nixl dependency (#24874)

Signed-off-by: bbartels <benjamin@bartels.dev>
Signed-off-by: Benjamin Bartels <benjamin@bartels.dev>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Daniele <36171005+dtrifiro@users.noreply.github.com>

docker/Dockerfile

@@ -283,6 +283,10 @@ WORKDIR /vllm-workspace
 ENV DEBIAN_FRONTEND=noninteractive
 ARG TARGETPLATFORM
 
+ARG GDRCOPY_CUDA_VERSION=12.8
+# Keep in line with FINAL_BASE_IMAGE
+ARG GDRCOPY_OS_VERSION=Ubuntu22_04
+
 SHELL ["/bin/bash", "-c"]
 
 ARG DEADSNAKES_MIRROR_URL
@@ -441,13 +445,21 @@ COPY tools/install_deepgemm.sh /tmp/install_deepgemm.sh
 RUN --mount=type=cache,target=/root/.cache/uv \
     VLLM_DOCKER_BUILD_CONTEXT=1 /tmp/install_deepgemm.sh --cuda-version "${CUDA_VERSION}" ${DEEPGEMM_GIT_REF:+--ref "$DEEPGEMM_GIT_REF"} 
 
-# Install EP kernels(pplx-kernels and DeepEP), NixL
+COPY tools/install_gdrcopy.sh install_gdrcopy.sh
+RUN set -eux; \
+    case "${TARGETPLATFORM}" in \
+      linux/arm64) UUARCH="aarch64" ;; \
+      linux/amd64) UUARCH="x64" ;; \
+      *) echo "Unsupported TARGETPLATFORM: ${TARGETPLATFORM}" >&2; exit 1 ;; \
+    esac; \
+    ./install_gdrcopy.sh "${GDRCOPY_OS_VERSION}" "${GDRCOPY_CUDA_VERSION}" "${UUARCH}"; \
+    rm ./install_gdrcopy.sh
+
+# Install EP kernels(pplx-kernels and DeepEP)
 COPY tools/ep_kernels/install_python_libraries.sh install_python_libraries.sh
-COPY tools/install_nixl.sh install_nixl.sh
 ENV CUDA_HOME=/usr/local/cuda
 RUN export TORCH_CUDA_ARCH_LIST="${TORCH_CUDA_ARCH_LIST:-9.0a+PTX}" \
-    && bash install_python_libraries.sh \
-    && bash install_nixl.sh --force
+    && bash install_python_libraries.sh
 
 #################### vLLM installation IMAGE ####################
 

docs/serving/expert_parallel_deployment.md

@@ -10,7 +10,7 @@ Before using EP, you need to install the necessary dependencies. We are actively
 
 1. **Install DeepEP and pplx-kernels**: Set up host environment following vLLM's guide for EP kernels [here](gh-file:tools/ep_kernels).
 2. **Install DeepGEMM library**: Follow the [official instructions](https://github.com/deepseek-ai/DeepGEMM#installation).
-3. **For disaggregated serving**: Install UCX and NIXL following the [script](gh-file:tools/install_nixl.sh).
+3. **For disaggregated serving**: Install `gdrcopy` by running the [`install_gdrcopy.sh`](gh-file:tools/install_gdrcopy.sh) script (e.g., `install_gdrcopy.sh "${GDRCOPY_OS_VERSION}" "12.8" "x64"`). You can find available OS versions [here](https://developer.download.nvidia.com/compute/redist/gdrcopy/CUDA%2012.8/).
 
 ### Backend Selection Guide
 
@@ -191,7 +191,7 @@ For production deployments requiring strict SLA guarantees for time-to-first-tok
 
 ### Setup Steps
 
-1. **Install KV Connector**: Install NIXL using the [installation script](gh-file:tools/install_nixl.sh)
+1. **Install gdrcopy/ucx/nixl**: For maximum performance, run the [install_gdrcopy.sh](gh-file:tools/install_gdrcopy.sh) script to install `gdrcopy` (e.g., `install_gdrcopy.sh "${GDRCOPY_OS_VERSION}" "12.8" "x64"`). You can find available OS versions [here](https://developer.download.nvidia.com/compute/redist/gdrcopy/CUDA%2012.8/). If `gdrcopy` is not installed, things will still work with a plain `pip install nixl`, just with lower performance. `nixl` and `ucx` are installed as dependencies via pip.
 
 2. **Configure Both Instances**: Add this flag to both prefill and decode instances `--kv-transfer-config '{"kv_connector":"NixlConnector","kv_role":"kv_both"}`
 

requirements/kv_connectors.txt

-lmcache
+lmcache
\ No newline at end of file
+nixl >= 0.5.1 # Required for disaggregated prefill

tools/install_gdrcopy.sh

+#!/usr/bin/env bash
+set -euo pipefail
+
+# Usage: install_gdrcopy.sh <GDRCOPY_OS_VERSION> <GDRCOPY_CUDA_VERSION> <uuarch>
+# uuarch must be "x64" or "aarch64"
+# Optional: set GDRCOPY_VERSION to override the libgdrapi package version (default: 2.5.1-1)
+# Requires: curl, apt-get, root privileges
+if [[ $(id -u) -ne 0 ]]; then
+  echo "Must be run as root" >&2
+  
+  exit 1
+fi
+if [[ $# -ne 3 ]]; then
+  echo "Usage: $0 <GDRCOPY_OS_VERSION> <GDRCOPY_CUDA_VERSION> <uuarch(x64|aarch64)>" >&2
+  exit 1
+fi
+
+OS_VER="$1"
+CUDA_VER="$2"
+UUARCH_RAW="$3"
+
+# Normalize/validate arch
+case "${UUARCH_RAW,,}" in
+  aarch64|arm64)
+    URL_ARCH="aarch64"
+    DEB_ARCH="arm64"
+    ;;
+  x64|x86_64|amd64)
+    URL_ARCH="x64"
+    DEB_ARCH="amd64"
+    ;;
+  *)
+    echo "Unsupported uuarch: ${UUARCH_RAW}. Use 'x64' or 'aarch64'." >&2
+    exit 1
+    ;;
+esac
+
+OS_VER_LOWER="$(tr '[:upper:]' '[:lower:]' <<<"$OS_VER")"
+GDRCOPY_PKG_VER="${GDRCOPY_VERSION:-2.5.1-1}"
+
+DEB_NAME="libgdrapi_${GDRCOPY_PKG_VER}_${DEB_ARCH}.${OS_VER}.deb"
+BASE_URL="https://developer.download.nvidia.com/compute/redist/gdrcopy"
+URL="${BASE_URL}/CUDA%20${CUDA_VER}/${OS_VER_LOWER}/${URL_ARCH}/${DEB_NAME}"
+
+echo "Downloading: ${URL}"
+TMPDIR="$(mktemp -d)"
+trap 'rm -rf "${TMPDIR}"' EXIT
+
+curl -fSL "${URL}" -o "${TMPDIR}/${DEB_NAME}"
+
+export DEBIAN_FRONTEND=noninteractive
+apt-get update
+apt-get install -y "${TMPDIR}/${DEB_NAME}"
+apt-get clean
+rm -rf /var/lib/apt/lists/*
+
+echo "Installed ${DEB_NAME}"

tools/install_nixl.sh

-#!/bin/bash
-# Usage: ./install_nixl.sh [--force]
-
-FORCE=false
-if [ "$1" == "--force" ]; then
-    FORCE=true
-fi
-
-SUDO=false
-if command -v sudo >/dev/null 2>&1 && sudo -n true 2>/dev/null; then
-    SUDO=true
-fi
-
-ARCH=$(uname -m)
-
-ROOT_DIR="/usr/local"
-mkdir -p "$ROOT_DIR"
-GDR_HOME="$ROOT_DIR/gdrcopy"
-UCX_HOME="$ROOT_DIR/ucx"
-NIXL_HOME="$ROOT_DIR/nixl"
-CUDA_HOME=/usr/local/cuda
-
-export PATH="$GDR_HOME/bin:$UCX_HOME/bin:$NIXL_HOME/bin:$PATH"
-export LD_LIBRARY_PATH="$GDR_HOME/lib:$UCX_HOME/lib:$NIXL_HOME/lib/$ARCH-linux-gnu:$LD_LIBRARY_PATH"
-
-TEMP_DIR="nixl_installer"
-mkdir -p "$TEMP_DIR"
-cd "$TEMP_DIR"
-
-pip install meson ninja pybind11
-
-if [ ! -e "/dev/gdrdrv" ] || [ "$FORCE" = true ]; then
-    echo "Installing gdrcopy\n"
-    wget https://github.com/NVIDIA/gdrcopy/archive/refs/tags/v2.5.tar.gz
-    tar xzf v2.5.tar.gz; rm v2.5.tar.gz
-    cd gdrcopy-2.5
-    make prefix=$GDR_HOME CUDA=$CUDA_HOME all install
-    
-    if $SUDO; then
-        echo "Running insmod.sh with sudo"
-        sudo ./insmod.sh
-    else
-        echo "Skipping insmod.sh - sudo not available"
-        echo "Please run 'sudo ./gdrcopy-2.5/insmod.sh' manually if needed"
-    fi
-    
-    cd ..
-else
-    echo "Found /dev/gdrdrv. Skipping gdrcopy installation"
-fi
-
-if ! command -v ucx_info &> /dev/null || [ "$FORCE" = true ]; then
-    echo "Installing UCX"
-    wget https://github.com/openucx/ucx/releases/download/v1.18.0/ucx-1.18.0.tar.gz
-    tar xzf ucx-1.18.0.tar.gz; rm ucx-1.18.0.tar.gz
-    cd ucx-1.18.0
-    
-    # Checking Mellanox NICs
-    MLX_OPTS=""
-    if lspci | grep -i mellanox > /dev/null || command -v ibstat > /dev/null; then
-        echo "Mellanox NIC detected, adding Mellanox-specific options"
-        MLX_OPTS="--with-rdmacm \
-                  --with-mlx5-dv \
-                  --with-ib-hw-tm"
-    fi
-    
-    ./configure  --prefix=$UCX_HOME                \
-                --enable-shared                    \
-                --disable-static                   \
-                --disable-doxygen-doc              \
-                --enable-optimizations             \
-                --enable-cma                       \
-                --enable-devel-headers             \
-                --with-cuda=$CUDA_HOME             \
-                --with-dm                          \
-                --with-gdrcopy=$GDR_HOME           \
-                --with-verbs                       \
-                --enable-mt                        \
-                $MLX_OPTS
-    make -j
-    make -j install-strip
-    
-    if $SUDO; then
-        echo "Running ldconfig with sudo"
-        sudo ldconfig
-    else
-        echo "Skipping ldconfig - sudo not available"
-        echo "Please run 'sudo ldconfig' manually if needed"
-    fi
-
-    cd ..
-else
-    echo "Found existing UCX. Skipping UCX installation"  
-fi
-
-if ! command -v nixl_test &> /dev/null || [ "$FORCE" = true ]; then
-    echo "Installing NIXL"
-    wget https://github.com/ai-dynamo/nixl/archive/refs/tags/0.2.0.tar.gz
-    tar xzf 0.2.0.tar.gz; rm 0.2.0.tar.gz
-    cd nixl-0.2.0
-    meson setup build --prefix=$NIXL_HOME -Ducx_path=$UCX_HOME
-    cd build
-    ninja
-    ninja install
-
-    cd ../..
-else
-    echo "Found existing NIXL. Skipping NIXL installation"  
-fi