Yes, you’ll need a laptop (especially if you’re a hosteler and don’t have a desktop around you). The specifications need not be extremely high. A decent CPU is fine for your computational problems and good programmers usually don’t let the CPU brute force through their combinatorial explosions. I would recommend you to have a GPU because when you start learning mathematical computing, ML/DL (parts of Artificial Intelligence), simulation, etc. you require massive and robust parallel processing. Prefer Nvidia over AMD.
The laptop seriously doesn’t need to be too expensive. 1-8 GB RAM is perfectly fine. Don’t go for those fancy overkill. 500 gigs or 1TB storage is perfectly sufficient. Even less than that won’t harm you because you won’t have too heavy software. Even MATLAB won’t consume all that space.
The deep learning library of tensorflow natively supports Nvidia GPUs. You’ll use it later.
DO NOT install Ubuntu 18.04. Several 3rd party software packages aren’t build for it, including the CUDA drivers, GNU Octave, etc. Install Ubuntu 16.04.4 LTS only. And if possible, please avoid Windows.
Company of laptop doesn’t matter. You’re here to learn and code, not to show off. Macs are cool, but remember that either you purchase a proper MacBook Pro with proper specifications or you stick to a PC. You’ll realize later how important the PC performance is. Actually, a PC for the same price would outperform a Macbook.
We have a decent internet connection here in the campus and that’s practically all you guys need to learn anything. The library isn’t too big, but it has almost everything you might need.
You have 1.5 months before college begins so learn programming in C/C++. You can refer to several online sources and books like E. Balagurusamy.
Start reading about the recent trends in technology, have a brief idea about all the domains of technology. Read and compare the syllabus of various colleges. Read one page of Wikipedia on each topic mentioned in the syllabus before you enter the college so that you have a hint of what you’re going to learn and have a better planning accordingly.
And bachon, GPU is Graphics Card. The one you use for playing games. It renders your frames and resolves those tensors dynamically. It can do other processing operations too. But games mat khelna because you’ll miss out on a lot of things which are far more cool than gaming.