From 03e52ecba3b60b04b552d82809043e5642509005 Mon Sep 17 00:00:00 2001
From: Frank Lee <somerlee.9@gmail.com>
Date: Fri, 10 Jun 2022 17:38:15 +0800
Subject: [PATCH] [workflow] added regular 8 GPU testing (#1099)

* [workflow] added regular 8 GPU testing

* polish workflow
---
 .github/workflows/build_gpu_8.yml     | 37 +++++++++++++++++++++++++++
 .github/workflows/release_nightly.yml | 14 ++--------
 2 files changed, 39 insertions(+), 12 deletions(-)
 create mode 100644 .github/workflows/build_gpu_8.yml

diff --git a/.github/workflows/build_gpu_8.yml b/.github/workflows/build_gpu_8.yml
new file mode 100644
index 000000000..410bb2167
--- /dev/null
+++ b/.github/workflows/build_gpu_8.yml
@@ -0,0 +1,37 @@
+name: Build on 8 GPUs
+
+on:
+  schedule:
+    # run at 00:00 of every Sunday 
+    - cron:  '0 0 * * *'
+  workflow_dispatch:
+
+jobs:
+  build:
+    name: Build and Test Colossal-AI
+    if: |
+        github.event.pull_request.base.repo.full_name == 'hpcaitech/ColossalAI'
+    runs-on: [self-hosted, 8-gpu]
+    container:
+      image: hpcaitech/pytorch-cuda:1.10.1-11.3.0
+      options: --gpus all --rm -v /data/scratch/cifar-10:/data/scratch/cifar-10
+    timeout-minutes: 40
+    steps:
+      - uses: actions/checkout@v2
+        with:
+          ssh-key: ${{ secrets.SSH_KEY_FOR_CI }}
+      - name: Install Colossal-AI      
+        run: |
+          [ ! -z "$(ls -A /github/home/cuda_ext_cache/)" ] && cp -r /github/home/cuda_ext_cache/* /__w/ColossalAI/ColossalAI/
+          pip install -r requirements/requirements.txt
+          pip install -v -e .
+          cp -r /__w/ColossalAI/ColossalAI/build /github/home/cuda_ext_cache/
+          cp /__w/ColossalAI/ColossalAI/*.so /github/home/cuda_ext_cache/
+          pip install -r requirements/requirements-test.txt
+      - name: Unit Testing
+        run: |
+          gpu_used=$(nvidia-smi -i 0 --query-gpu=memory.used --format=csv,noheader,nounits)
+          [ "$gpu_used" -gt "100" ] && PYTHONPATH=$PWD pytest tests
+        env:
+          DATA: /data/scratch/cifar-10
+          
\ No newline at end of file
diff --git a/.github/workflows/release_nightly.yml b/.github/workflows/release_nightly.yml
index 9014b1290..0ef942841 100644
--- a/.github/workflows/release_nightly.yml
+++ b/.github/workflows/release_nightly.yml
@@ -5,16 +5,7 @@ on:
     # run at 00:00 of every Sunday 
     - cron:  '0 0 * * 6'
   workflow_dispatch:
-    inputs:
-      cuda_version:
-        type: choice
-        description: CUDA Version
-        default: "all"
-        required: true
-        options:
-        - "all"
-        - "11.3"
-        - "10.2"
+  
 jobs:
   matrix_preparation:
     name: Prepare Container List
@@ -24,8 +15,7 @@ jobs:
     steps:
     - id: set-matrix
       run: |
-        [ "${{github.event.inputs.cuda_version}}" != "all" ] && matrix="[\"hpcaitech/cuda-conda:${{github.event.inputs.cuda_version}}\"]"
-        [ "${{github.event.inputs.cuda_version}}" == "all" || "${{github.event.inputs.cuda_version}}" == "" ] && matrix="[\"hpcaitech/cuda-conda:11.3\", \"hpcaitech/cuda-conda:10.2\"]"
+        matrix="[\"hpcaitech/cuda-conda:11.3\", \"hpcaitech/cuda-conda:10.2\"]"
         echo $matrix
         echo "::set-output name=matrix::{\"container\":$(echo $matrix)}"