search

eyalroz / cuda-api-wrappers

Thin, unified, C++-flavored wrappers for the CUDA APIs
BSD 3-Clause "New" or "Revised" License
785 stars 80 forks source link

Use range-based copy, and fill - like in the C++ standard library #11

Open eyalroz opened 7 years ago

eyalroz commented 7 years ago

We currently have copy(destination, source, length); perhaps we should change that to the same style / calling convention as the standard libary?

eyalroz commented 1 year ago

Well, I guess I have to answer this with a _no, since pairs-of-iterators are not really relevant for us. They relate to closely to the metaphor (and implementation...) of the serial loop implementing the iteration. On GPUs there is no such loop. If anything, we'll have containerish/memory-area APIs, which are somewhat range-like. But not iterator pairs.