Connect to iGPU Lab

IMPORTANT: If you are NOT on campus, before continuing with these steps, please make sure that you have successfully logged into https://rds.syr.edu/ and have an active session.

To use the Interactive GPU Learning Lab (IGL):

  1. Open a web browser, preferably Edge or Chrome, and then navigate to maxlabs.maxwell.syr.edu
  2. Log in using your SUID. Once logged in you will be presented with your RemoteLab Pool options.
  3. Select the pool named “ARCLAB -T1000 GPU Workstation Cluster”.

After selecting the ARCLAB pool, you will be presented with the ability to make a reservation for one of the GPU Workstations. To do so, find and click the “make my reservation” button. After clicking the button, you will be taken to another screen showing that you have reserved a workstation. Here, click the button “connect to remote lab” to download the RDP File.

Please note, some browsers will automatically download this RDP file directly into your Downloads folder on your computer. Check there for the file if the file isn’t readily presented to you.

Upon opening the RDP file, you should see that it already has your username completed for you. You just need to put in your password and click OK to login. At this point, you have successfully reserved and connected to the IGL!

Please feel free to contact us directly at: ICTResearch@syr.edu if you experience any difficulties or have any questions about this resource.

Conda Environments

We’ve pre-created (3) three pre-configured conda environments that should allow you to execute some general python scripts/notebooks you may come across the internet with from your research. These are “GPU-Enabled” and should allow you to quickly toil with some python concepts prior to having to create an actual environment.

Getting Started: Pre-Configured Conda

These three environments are related to their versions, so for Python Version 3.7, the environment is named: py3.7.  For Versions 3.8 and 3.9 (the other two environments), those would be named: py3.8 and py3.9 respectively.

Consistent across them all, you will find some version of the following packages installed (version determined at install-time):

  • NumPy
    • Pandas
    • Scipy
    • Matplotlib
    • SciKit-Learn
    • PyTorch
    • Scrapy
    • Statsmodels
    • Seaborn
    • Plotly
    • Jupyter
  • Additionally installed is TensorFlow version 2.9
    • This includes TensorFlow-GPU and Keras

You can activate any of these (3) three environments from the base environment by running the following where “X” is the Minor Version.

conda activate py3.X

When you have completed your efforts/work with your script/notebook – or if you want to switch between conda environments – you will need to deactivate the current environment first, by running the following:

conda deactivate

Note, you do not need to specify the environment here.