According to studies, buyers expect that the purchased laptop will last at least 4-5 years. For comparison, a smartphone is chosen with an eye on 2-3 years. Therefore, the choice of the best laptop for machine learning should be treated with particular care. In this article, we will analyze the main points that are worth paying attention to.
Ultrabook or laptop?
Perhaps a couple of years ago, when choosing a laptop for students, the choice would certainly have been mainly from the category of light and portable models with long battery life. But the complexity and variety of tasks is growing, and if the functionality of a typewriter and a device for viewing videos is clearly cramped for you, you should pay attention to productive models with a discrete GPU. The world has become absolutely mobile, and access to productivity is now required not only at night, to play games, but at any time of the day and anywhere. Now you need to be able to study and work from anywhere. If only there was fast internet. Coronavirus remote work has only spurred this trend.
The low performance of ultrabooks and the often inability to upgrade are not the only problems with ultrabooks. In principle, they do not have a power reserve that will allow them to remain relevant after a couple of years. Lack of power will affect any task.
The best laptop for school in 2021 should be versatile: have desktop-like power, be comfortable to use, and have good battery life. All this is about laptops on RTX 30-series video cards.
In their face, you get a powerful computer with a high margin of safety, the performance of which will be relevant even at the end of its life. At the same time, the thickness of many models does not exceed 20 mm. And these laptops are quiet and have extended battery life.
Accelerating AI, Machine Learning and Data Science
With thousands of cores capable of working simultaneously, the GPU is many times more efficient than the CPU in many tasks. Tasks related to artificial intelligence and neural networks are one of the areas in which RTX GPUs reveal their potential to 100%.
English speakers can watch an interesting video in which the image recognition model is trained, first using only the excellent AMD Ryzen 5900X processor, and then with the GeForce RTX 3070. The difference is enormous. Where the CPU spends 4.16 seconds, the RTX 3070 spends just 0.16 seconds. In this case, we are talking about one step. Accordingly, where the CPU spends an hour, the video card can handle it in a couple of minutes.