It is no surprise that computational physics simulations require a lot of computing resources, the desire for bigger and faster computers has been the one constant in this field since its inception. This desire has been met historically through faster single chips until around 1990, then more cores and through parallel architectures until 2010, and in the last ten years through accelerators such as GPUs. Today if you look at the ten fastest computers in the world, seven of them have NVIDIA GPU accelerators. This talk will give a brief historical perspective, insights into what makes programming on GPUs different, and how you can get started making use of this hardware.
Zoom: https://unipd.zoom.us/j/82645715140?pwd=U1F6bEYxYVNIQnIvWU55MzVVVldjZz09
CISAS
