Joel Falcou wrote:
Ok, I'm really curious to here what you think is worthy to be a library and what is only worth a mere patch ?
Why not modify Boost.uBlas so that it uses vectorization, then use Boost.uBlas as a basis to write other vectorization-aware libraries? Introducing a new library that does it then rewriting Boost.uBlas in terms of that library seems to be more work; especially since both libraries would have a very similar interface.
Have you ever tried optimizing *any* code using vector intrinsics on any kind of platform ?
I have used the GCC vector intrinsics, yes. I didn't find it that hard. At least much easier than writing inline assembly. Reading that sentence, however, I'm not too sure we're talking about the same thing. Are you saying those intrinsics aren't enough to generate optimal code?