We compared Pymp and Joblib libraries on 2 majorly used datasets in the machine learning world. Those are Image and Stocks.
Below are the links for the datasets we used from Kaggle:
https://www.kaggle.com/jacksoncrow/stock-market-dataset
https://www.kaggle.com/c/dogs-vs-cats/data
Amount of Dataset used: 200 stocks files 2000 images of cats and dogs
We used a little bit of hyberparameter tuning to check the performance of both libraries on KNN algorithm. We choosed KNN because it is an non-paramteric machine learning algorithm and has the dataset grows the time complexity of the algorithm increases exponentially.
Note: The device used for testing is Quad-core Linux Environment. The performance of the libraries will be depending on the device configuaration.