On Friday, 11 March 2016 09:09:14 UTC+8, Dragan Djuric wrote: > > This is already working well for the array programming APIs (it's easy to >> mix and match Clojure data structures, Vectorz Java-based arrays, GPU >> backed arrays in computations). >> > > While we could agree to some extent on the other parts of your post but > the GPU part is *NOT* true: I would like you to point me to a single > implementation anywhere (Clojure or other) that (easily or not) mixes and > matches arrays in RAM and arrays on the GPU backend. It simply does not > work that way. >
You misunderstand my point. Obviously, there may need to be some copying when you move between managed and unmanaged memory. But I'm not talking about that: the point is that this can happen "under the hood", without the user needing to do explicit conversions etc. All thanks to the protocol implementations, you can mix and match GPU, native and Java backed instances with the same API. core.matrix can trivially do stuff like (add! native-array java-array) for example. What's not to like about that? -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
