- We need to update the manuscript to account for the feature with more granular tracking - We also need to make sure there is an example of gpu-tracker in an HPC environment (we can use either the DGX or LCC) - Create a figure showing the details on howgpu_tracker was used to profile the submitted job. - Maybe create a new example using gpu_tracker on the command line for this purpose as well. - Include examples of submitting jobs on a SLURM scheduler using gpu_tracker. - Also, identify the examples that were run on a cluster. - Example job submission script can be included, with private details removed. - Each of the current use-case examples should include a figure showing how gpu_tracker was executed and the resulting output.
