Add test for building without support for any DL frameworks (#974)

* Add test for building with no DL frameworks
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Use NGC container for HPC SDK

NGC CUDA container does not have Git.
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Install dependencies in NGC CUDA container
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Install CMake
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Debug sanity test
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Do sanity test out of repo
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Install ninja
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Remove ninja
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Build with 4 parallel processes
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Maximize parallelism in GitHub builds
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Reduce number of PyTorch parallel build jobs
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Update docs/installation.rst
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Signed-off-by: Phuong Nguyen <36155692+phu0ngng@users.noreply.github.com>

* Disable parallel build for PyTorch test
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

---------
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Signed-off-by: Phuong Nguyen <36155692+phu0ngng@users.noreply.github.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Co-authored-by: Phuong Nguyen <36155692+phu0ngng@users.noreply.github.com>

.github/workflows/build.yml

@@ -8,6 +8,28 @@ on:
   pull_request:
   workflow_dispatch:
 jobs:
+  core:
+    name: 'Core'
+    runs-on: ubuntu-latest
+    container:
+      image: nvcr.io/nvidia/cuda:12.5.0-devel-ubuntu22.04
+      options: --user root
+    steps:
+      - name: 'Dependencies'
+        run: |
+          apt-get update
+          apt-get install -y git python3.9 pip cmake ninja-build cudnn9-cuda-12
+      - name: 'Checkout'
+        uses: actions/checkout@v3
+        with:
+          submodules: recursive
+      - name: 'Build'
+        run: pip install . -v
+        env:
+          NVTE_FRAMEWORK: none
+      - name: 'Sanity check'
+        run: python3 -c "import transformer_engine"
+        working-directory: /
   pytorch:
     name: 'PyTorch'
     runs-on: ubuntu-latest

docs/installation.rst

@@ -63,7 +63,7 @@ Execute the following command to install the latest development build of Transfo
 
   pip install git+https://github.com/NVIDIA/TransformerEngine.git@main
 
-This will automatically detect if any supported deep learning frameworks are installed and build Transformer Engine support for them. To explicitly specify frameworks, set the environment variable `NVTE_FRAMEWORK` to a comma-separated list (e.g. `NVTE_FRAMEWORK=jax,pytorch`).
+This will automatically detect if any supported deep learning frameworks are installed and build Transformer Engine support for them. To explicitly specify frameworks, set the environment variable `NVTE_FRAMEWORK` to a comma-separated list (e.g. `NVTE_FRAMEWORK=jax,pytorch`). To only build the framework-agnostic C++ API, set `NVTE_FRAMEWORK=none`.
 
 In order to install a specific PR, execute after changing NNN to the PR number: