Machine learning algorithms are becoming more sophisticated with time, their computational capabilities and resource consumption are also increasing. And with its escalating potential, machine learning enthusiasts constantly look for better and easier ways to run complex algorithms. Often, laptops and CPUs available for personal use have memory ranging from 4/8/12 GB and have comparatively lower computation capability. In such restricted environments, there is a trade-off between the accuracy of machine learning models and resource consumption. This barrier can be overcome by using cloud CPUs offered popularly by Google Colab and Kaggle. These popular platforms provide CPUs capable of training and evaluating machine learning algorithms. In this paper, we illustrate the performance of the CPUs offered by Google Colab and Kaggle and compare it with a local device (ASUS FX505GE) using KNN, Logistic Regression, and Random Forest algorithm. We could find that multithreading provides better performance in the case of machine learning algorithms. Random Forest takes comparatively longer for evaluation and training purposes because of the numerous calculations involved in its different phases. Even though KNN has no training phase, the impact of multi-processing cores is noticeable both in training and evaluation. However, Logistic Regression has the best performance with only a single core both in the training and evaluation period because of the simplicity of the model structure and requires fewer computational resources.