Course/apcomp297r

[kgray-hu]

Overview

The purpose of this PR is to create 2 IPython Jupyterlab based course enviornments for APCOMP297r, a capstone course in order to provide an interactive python cpu environment and GPU environment, this leverages the GPU/CPU setup for cs1090b.

Changes

• Added local/apcomp297r-cpu.yml.erb for CPU
• Added local/apcomp297r-gpu.yml.erb for GPU
• Both reference cs1090b-cpu and cs1090b-gpu mamba environments (reused from CS 1090b)

Notes

Key differences between APCOMP 297R GPU differs from CS 1090b GPU:

• Canvas ID, environment mode (hard-coded global vs. widget), core count (8 staging vs. 12 prod)

diff: apcomp297r-gpu.yml.erb vs. cs1090b-gpu.yml.erb

--- a/local/cs1090b-gpu.yml.erb
+++ b/local/apcomp297r-gpu.yml.erb
@@ -20,7 +20,7 @@
   # Cannot have multiple enabled groups if a spack installation in the course 
   # shared folder is in use. This is because the course installation uses the 
   # course shared folder, which is not accessible to users outside of that 
   # course.
-  "163602"
+  "162018" #APCOMP 297R
 ]
 
 def arrays_have_common_element(array1, array2)
@@ -44,14 +44,14 @@
 # a non-existant cluster.
 cluster: "<%= cluster %>"
 
-title: "Jupyter Lab - CS 1090B(GPU)"
+title: "Jupyter Lab - APCOMP 297R (GPU)"
 description: |
-  This app is configured for GPU access for CS 1090B.
+  This app is configured for GPU access for APCOMP 297R.
 
 cacheable: false
 
 attributes:
-  course: "163602"
+  course: "162018"
   spack: "mamba-gpu"
   mamba: "cs1090b-gpu"
   bc_queue: "gpu"
@@ -60,18 +60,7 @@
   # In this case, this form option toggles between using the global Spack 
   # installation and a spack installation in the course shared folder. If the 
   # course has no intention of managing their own Spack environments, this can 
   # be hard-coded to "global".
-  environment:
-    widget: "radio_button"
-    value: "course"
-    help: |
-      Choose the environment used to launch Jupyter Lab. The "Course"
-      environment is managed by the teaching staff and has the latest
-      dependencies, while the "Global" environment is managed at the system
-      level and provides the initial set of dependencies. Each session will have 1 GPU allocated, 12 CPU cores and 48GB of memory.
-    options:
-      - ["Course (course managed)", "course"]
-      - ["Global (system managed)", "global"]
+  environment: "global"
 
   # How many CPU cores to include in the slurm job submission
-  custom_num_cores: 12
+  custom_num_cores: 8  # For Stage Testing: 8, For Production: 12
   # How many GPU GRES resources to include in the job submission.
   custom_num_gpus: 1

References

APCOMP 297r Course Setup Outline

[jaguillette]

Glad to see reusing the existing environment for CS 1090b works well for this. I was able to validate in prod from a development app that the CPU app has the regular packages available, and that the GPU app has those packages and appropriate access to a GPU. Thanks!

[arthurian]

Looks good! 👍

[arthurian]

Nice idea to reuse the cs1090b environments.

[arthurian]

I was able to successfully launch the Jupyter Lab - APCOMP 297R (GPU) app in my sandbox apps from this branch and confirmed I was able to load tensorflow and see a GPU.

import tensorflow as tf
gpus = tf.config.list_physical_devices('GPU')
print(f"Detected GPUs: {gpus}")

Output:

Detected GPUs: [PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]