Add GPU watch

.pre-commit-config.yaml

@@ -8,7 +8,7 @@ repos:
     rev: v0.33.0
     hooks:
       - id: markdownlint
-        args: [--fix, --disable, MD013, --]
+        args: [--fix, --disable, MD013, MD024, --]
 
   # notebooks
   - repo: https://github.com/kynan/nbstripout

docs/n2v_hpc/start_jupyter.md

@@ -9,7 +9,7 @@ module load Jupyter-Notebook/6.4.3
 Then, we want to submit a job on a GPU node :fire::
 
 ```bash title="Submit GPU job"
-srun --time=04:00:00 --job-name=<your.name> -p neubias --gres=gpu:1 --qos=gpu  jupyter-notebook --no-browser --ip='0.0.0.0'
+srun --time=04:00:00 --job-name=<your.name> -p neubias --gres=gpu:1 --qos=gpu --mem=16G jupyter-notebook --no-browser --ip='0.0.0.0'
 ```
 
 You should now have a `jupyter` server running on a GPU node.

docs/n2v_hpc/watch_gpu.md

@@ -0,0 +1,32 @@
+---
+tags:
+  - help
+---
+# Watch GPU consumption
+
+In this section, we demonstrate how to watch the GPU consumption of your
+jupyter job on the compute node.
+
+## Check the compute node
+
+```bash title=""
+squeue | grep <your-job-name>
+```
+
+You should see something like `maestro-####`, write down the `####` numbers.
+
+## Connect to the compute node
+
+Open a new terminal and ssh into the compute node
+
+```bash
+ssh m####
+```
+
+where `####` are the numbers from the previous section
+
+## Watch GPU consumption
+
+```bash
+watch nvidia-smi
+```

mkdocs.yml

@@ -29,6 +29,7 @@ nav:
       - n2v_hpc/activate_env.md
       - n2v_hpc/data.md
       - n2v_hpc/start_jupyter.md
+      - n2v_hpc/watch_gpu.md
       - n2v_hpc/train.md
       - n2v_hpc/bonus.md
   - napari.md