-
Notifications
You must be signed in to change notification settings - Fork 110
150 lines (134 loc) · 6.11 KB
/
linux-gpu-x64-build.yml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
name: "Linux CUDA x64 Build"
on:
workflow_dispatch:
push:
branches:
- main
- rel-*
pull_request:
concurrency:
group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
cancel-in-progress: true
env:
ORT_NIGHTLY_REST_API: "https://feeds.dev.azure.com/aiinfra/PublicPackages/_apis/packaging/Feeds/ORT-Nightly/packages?packageNameQuery=Microsoft.ML.OnnxRuntime.Gpu.Linux&api-version=6.0-preview.1"
ORT_PACKAGE_NAME: Microsoft.ML.OnnxRuntime.Gpu.Linux
ORT_NIGHTLY_SOURCE: "https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/ORT-Nightly/nuget/v3/index.json"
DOTNET_INSTALL_DIR: "${{ github.workspace }}/dotnet"
jobs:
linux-cuda-x64-build:
env :
PYTHON_EXECUTABLE: "/opt/python/cp38-cp38/bin/python3.8"
runs-on: [ "self-hosted", "1ES.Pool=onnxruntime-genai-Ubuntu2204-A10" ]
steps:
- name: Checkout OnnxRuntime GenAI repo
uses: actions/checkout@v4
with:
submodules: true
# We are using the same manylinux repo as the one used in the packaging build
- name: Checkout ManyLinux repo
uses: actions/checkout@v4
with:
repository: pypa/manylinux
ref: 5eda9aded5462201e6310105728d33016e637ea7
clean: true
path: manylinux
submodules: true
- uses: actions/setup-dotnet@v4
with:
dotnet-version: '8.0.x'
- name: Get the Latest OnnxRuntime Nightly Version
shell: pwsh
run: |
$resp = Invoke-RestMethod "${{ env.ORT_NIGHTLY_REST_API }}"
$ORT_NIGHTLY_VERSION = $resp.value[0].versions[0].normalizedVersion
Write-Host "$ORT_NIGHTLY_VERSION"
"ORT_NIGHTLY_VERSION=$ORT_NIGHTLY_VERSION" | Out-File -FilePath $env:GITHUB_ENV -Append
- name: Download OnnxRuntime Nightly
run: |
dotnet new console
dotnet add package ${{ env.ORT_PACKAGE_NAME }} --version ${{ env.ORT_NIGHTLY_VERSION }} --source ${{ env.ORT_NIGHTLY_SOURCE }} --package-directory .
dotnet build
continue-on-error: true
- name: list files
shell: bash
run: |
ls -l
ls -R ${{ env.ORT_PACKAGE_NAME }}
continue-on-error: true
- name: Extract OnnxRuntime library and header files
run: |
set -e -x
mkdir -p ort/lib
mv microsoft.ml.onnxruntime.gpu.linux/${{ env.ORT_NIGHTLY_VERSION }}/buildTransitive/native/include ort/
mv microsoft.ml.onnxruntime.gpu.linux/${{ env.ORT_NIGHTLY_VERSION }}/runtimes/linux-x64/native/* ort/lib/
cp ort/lib/libonnxruntime.so ort/lib/libonnxruntime.so.1
- name: Get Docker Image
run: |
set -e -x
az login --identity --username 63b63039-6328-442f-954b-5a64d124e5b4
az acr login --name onnxruntimebuildcache --subscription 00c06639-6ee4-454e-8058-8d8b1703bd87
python3 tools/ci_build/get_docker_image.py --dockerfile tools/ci_build/github/linux/docker/manylinux/Dockerfile.manylinux2_28_cuda_12.2 \
--context tools/ci_build/github/linux/docker/manylinux \
--docker-build-args "--build-arg BUILD_UID=$( id -u )" \
--container-registry onnxruntimebuildcache \
--manylinux-src manylinux \
--multiple_repos \
--repository onnxruntimecudabuildx64
- name: Config with Cmake in Docker
run: |
set -e -x
docker run \
--gpus all \
--rm \
--volume $GITHUB_WORKSPACE:/ort_genai_src \
-w /ort_genai_src onnxruntimecudabuildx64 \
bash -c " \
/usr/bin/cmake --preset linux_gcc_cuda_release \
-DMANYLINUX=ON \
-DPYTHON_EXECUTABLE=${{ env.PYTHON_EXECUTABLE }} "
- name: Build with Cmake in Docker
run: |
set -e -x
docker run \
--gpus all \
--rm \
--volume $GITHUB_WORKSPACE:/ort_genai_src \
-w /ort_genai_src onnxruntimecudabuildx64 \
bash -c " \
/usr/bin/cmake --build --preset linux_gcc_cuda_release"
- name: Use Dummy HuggingFace Token
run: |
echo "HF_TOKEN=12345" >> $GITHUB_ENV
- name: Install the onnxruntime-genai Python wheel and run python test
run: |
echo "Installing the onnxruntime-genai Python wheel and running the Python tests"
docker run \
--gpus all \
--rm \
--volume /data/ortgenai_pytorch_models:/data/ortgenai_pytorch_models \
--volume $GITHUB_WORKSPACE:/ort_genai_src \
-e HF_TOKEN=$HF_TOKEN \
-w /ort_genai_src onnxruntimecudabuildx64 bash -c " \
${{ env.PYTHON_EXECUTABLE }} -m pip install -r test/python/requirements.txt --user && \
${{ env.PYTHON_EXECUTABLE }} -m pip install -r test/python/requirements-cuda.txt --user && \
${{ env.PYTHON_EXECUTABLE }} -m pip install /ort_genai_src/build/cuda/wheel/onnxruntime_genai*manylinux*.whl --no-deps --user && \
${{ env.PYTHON_EXECUTABLE }} test/python/test_onnxruntime_genai.py --cwd test/python --test_models test/test_models --e2e"
# TODO: Enable this by adding dotnet to the docker image
# - name: Build the C# API and Run the C# Tests
# run: |
# echo "Building the C# API and running the C# tests"
# docker run \
# --gpus all \
# --rm \
# --volume $GITHUB_WORKSPACE:/ort_genai_src \
# -w /ort_genai_src/test/csharp onnxruntimecudabuildx64 bash -c " \
# dotnet test /p:NativeBuildOutputDir='/ort_genai_src/build/cuda/'"
- name: Docker -- Run unit tests
run: |
echo "Running docker image onnxruntimecudabuildx64"
docker run \
--gpus all \
--rm \
--volume /data/ortgenai_pytorch_models:/data/ortgenai_pytorch_models \
--volume $GITHUB_WORKSPACE:/ort_genai_src \
-w /ort_genai_src onnxruntimecudabuildx64 bash -c "ORTGENAI_LOG_ORT_LIB=1 LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/ort_genai_src/build/cuda/ /ort_genai_src/build/cuda/test/unit_tests"