Legate Numpy: Accelerated and Distributed Array Computing |SciPy 2020| Michael Bauer
Legate NumPy is a drop-in replacement for NumPy that allows users to transparently distribute and accelerate NumPy programs on supercomputers with hundreds of GPUs simply by changing a single Python import statement. This feat is accomplished by using Legion, a more traditional HPC programming model and runtime system. Legate NumPy translates NumPy programs down to the Legion programming model and then leverages the Legion runtime to run on distributed and heterogeneous machines. In this talk, we’ll describe the basic architecture of Legate NumPy and how it translates the NumPy programming model to Legion. We’ll also describe how users can get started building their own Python libraries on top of Legion with Legate and how the Legion programming model enables these libraries to naturally compose.
*****************
/ enthought
/ enthought
/ enthought
Sur cette page du site, vous pouvez voir la vidéo en ligne Legate Numpy: Accelerated and Distributed Array Computing |SciPy 2020| Michael Bauer durée heure minute seconde en bonne qualité , qui a été Téléchargé par l'utilisateur Enthought 05 juillet 2020, Partagez le lien avec vos amis et connaissances, sur youtube cette vidéo a déjà été regardée 2,221 fois et il a aimé 42 téléspectateurs. Bon visionnage!
