<@U65FK6QNB> GPU is the back-end implementation pr...

# datascience

[altavir]

@jimn GPU is the back-end implementation problem. User API does not depend on whether you use one core, multi core or GPU. We are discussing APIs so I need an example of actual mathematics. Operator overloading has nothing to do with it, I do not think that people really want to write those things in mathematical natural language.