Running Ray GPU Example
Describes how to run the Ray GPU example in HPE Ezmeral Unified Analytics Software.
Prerequisites
- Sign in to HPE Ezmeral Unified Analytics Software.
- Verify that the installed Ray client and server versions match. To verify,
complete the following steps in the terminal:
- To switch to Ray's environment,
run:
source /opt/conda/etc/profile.d/conda.sh && conda activate ray
- To verify that the Ray client and server versions match, run
:
ray --version
- To switch to Ray's environment,
run:
- Verify that the GPU support is enabled in your Ray cluster. See Enabling GPU Support During HPE Ezmeral Unified Analytics Software Installation or Enabling GPU Support and Configuring Resources After HPE Ezmeral Unified Analytics Software Installation.
About this task
In this tutorial, you will run the sample Ray GPU example and analyze logs to ensure that Unified Analytics is running the GPU-accelerated jobs.
You will complete the following steps:
Procedure
-
Create a notebook server using the
jupyter-tensorflow-cuda-full
image with at least 3 CPUs and 4 Gi of memory in Kubeflow. See Creating GPU-Enabled Notebook Servers. - In your notebook environment, activate the Ray-specific Python kernel.
-
To ensure optimal performance, use dedicated directories containing only the
essential files needed for that job submission as a working directory.
For example, if you do not see the
Ray-GPU
folder in the<username>
directory, copy the folder from theshared/ezua-tutorials/current-release/Data-Science/Ray/Ray-GPU
directory into the<username>
directory. Theshared
directory is accessible to all users. Editing or running examples from theshared
directory is not advised. The<username>
directory is specific to you and cannot be accessed by other users. -
Open the
ray-gpu-executor.ipynb
file in the<username>/Ray-GPU
directory. -
Select the first cell of the
ray-gpu-executor.ipynb
notebook and click Run the selected cells and advance (play icon). Continue until you run all cells.