I am a Machine Learning Engineer and have to use NVIDIA GPU accelerators (CUDA implementation) for training learning algorithms and speeding up the process.
I have a Dell G3 15 with two disk drives (1: 256GB SSD and 2: 1TB HDD) and plan to install Ubuntu Desktop 18.04.3LTS but I currently have Windows 10 installed on my SSD as well.
Will Ubuntu help in the extracting every last millisecond? Or rephrasing, should I consider installing Ubuntu (knowing the work I will have to do) to begin with? As evidently, Linux is preferred over Windows for pretty much every programming task.
If yes, how should I manage the disk drives' storage and the OSs to get the best performance? Meaning, should I install Ubuntu on SSD and Windows on HDD? Or both can be installed on the same SSD (with or without creating partitions and which would be better)?
Edit: Though the question may seem familiar to other question on here, this is specifically in the context of Machine Learning and the importance/need/usage of Ubuntu for Programming practices instead of Windows.