From 13bd8a5dd1e89b01b95a95a00bc311762122e5ef Mon Sep 17 00:00:00 2001
From: Raghu Raja <raghunch@amazon.com>
Date: Mon, 26 Feb 2024 05:17:21 +0000
Subject: [PATCH] ci: Add Neuron to the distcheck matrix

Extend the PR CI test matrix to include a Neuron build. We don't really
need to install the Neuron runtime since we don't use NRT symbols
directly, but mimic'ing a typical deployment where that's available is
easy and also helps catch any NRT-specific version requirements we might
add.

Same runtime as before as the runners for the matrix get kicked off in
parallel.

Signed-off-by: Raghu Raja <raghunch@amazon.com>
---
 .github/workflows/distcheck.yaml | 56 +++++++++++++++++++++++++-------
 1 file changed, 44 insertions(+), 12 deletions(-)

diff --git a/.github/workflows/distcheck.yaml b/.github/workflows/distcheck.yaml
index d9f1bf203..9bdf6b317 100644
--- a/.github/workflows/distcheck.yaml
+++ b/.github/workflows/distcheck.yaml
@@ -16,40 +16,72 @@ jobs:
         cc:
           - gcc
           - clang
+        sdk:
+          - cuda
+          - neuron
       fail-fast: false
     steps:
       - name: Install Dependencies
         run: |
+          sudo apt-get update -y
           sudo apt-get install -y ${{ env.APT_PACKAGES }}
-      - name: Install CUDA
+      - name: Install CUDA SDK
+        if: matrix.sdk == 'cuda'
         run: |
           sudo apt-get install -y nvidia-cuda-toolkit
-      - uses: actions/checkout@v4
-      - name: Build and Distribution Checks
+      - name: Install Neuron SDK
+        if: matrix.sdk == 'neuron'
         run: |
-          set -x
+          # Configure Linux for Neuron repository updates
+          sudo tee /etc/apt/sources.list.d/neuron.list > /dev/null << EOF
+          deb https://apt.repos.neuron.amazonaws.com jammy main
+          EOF
+          wget -qO - https://apt.repos.neuron.amazonaws.com/GPG-PUB-KEY-AMAZON-AWS-NEURON.PUB | sudo apt-key add -
+          sudo apt update -y
 
+          # Install Neuron Runtime
+          sudo apt-get install aws-neuronx-runtime-lib -y
+
+      - uses: actions/checkout@v4
+      - name: Install Libfabric
+        run: |
           # We're just doing distchecks, so it is fine if we
-          # just grab the latest master.
+          # just grab the latest master and built a lean build.
           git clone --depth 1 https://github.com/ofiwg/libfabric.git
           pushd libfabric
           ./autogen.sh
-          ./configure --prefix=$PWD/install CC=${{ matrix.cc }}
+          ./configure --prefix=$PWD/install \
+                      --disable-sockets \
+                      --disable-udp \
+                      --disable-mrail \
+                      --disable-opx \
+                      CC=${{ matrix.cc }}
           make -j $(nproc)
           make install
           popd
 
+      - name: Build Plugin
+        run: |
+          set -x
+
           # actions/checkout@v4 would drop the plugin source in $PWD,
-          # so go ahead and build it
+          # so go ahead and build it.
           ./autogen.sh
-          ./configure --with-libfabric=$PWD/libfabric/install --with-cuda=/usr/local/cuda/ CC=${{ matrix.cc }}
+          if [ ${{ matrix.sdk }} == "cuda" ]
+          then
+            ./configure --with-libfabric=$PWD/libfabric/install \
+                        --with-cuda=/usr/local/cuda/ \
+                        CC=${{ matrix.cc }}
+          else
+            ./configure --with-libfabric=$PWD/libfabric/install \
+                        --enable-neuron \
+                        CC=${{ matrix.cc }}
+          fi
           make -j $(nproc)
 
-          # Run Unit tests
-          make check
+      - name: Run Dist Check
+        run: make distcheck
 
-          # Run dist tarball checks
-          make distcheck
       - name: Upload build logs
         if: failure()
         uses: actions/upload-artifact@v4