Update strong and weak scaling

[quocdang1998]

Refactor strong scaling and weak scaling script and extend the code for GPU scaling study.

To run strong scaling:

python3 generate_scaling_study --type=strong
pushd output/strong
source submit_jobs.sh
popd
python3 plot_strong.py output/strong

To run weak scaling:

python3 generate_scaling_study --type=weak
pushd output/weak
source submit_jobs.sh
popd
python3 plot_weak.py output/weak

Basic example:

[wdhawkins]

Don't delete the jupyter notebooks. This is our historical record of scaling performance. We need to be better about recording our scaling performance (both cpu and gpu). This might be something we should do quarterly... at least every time we release a new version. The typical process has been to conduct the scaling study and, when warranted, to update the jupyter notebooks to keep a record. I'm fine with the changes to the scripts... including plotting, but we still need to main a record in our notebooks.

[ragusa]

Don't delete the jupyter notebooks. This is our historical record of scaling performance. We need to be better about recording our scaling performance (both cpu and gpu). This might be something we should do quarterly... at least every time we release a new version. The typical process has been to conduct the scaling study and, when warranted, to update the jupyter notebooks to keep a record. I'm fine with the changes to the scripts... including plotting, but we still need to main a record in our notebooks.

@quocdang1998 I concur. We need to keep historical data somehow. For now, I guess it lives in the notebooks