A new WG21 paper is available. A copy is linked below, and the paper will also appear in the next normal WG21 mailing. If you are not a committee member, please use the comments section below or the std-proposals forum for public discussion.
Document number: N3774
by Axel Naumann, Sandro Wenzel
In this paper we argue that vectorization is a very different concept to parallelization and needs to be supported explicitly by the language. The lack of C++ support for explicit vectorization costs factors (2 to 4 in real code) of performance on current commodity hardware. We demonstrate why we believe vectorization is a much needed language feature. The arguments presented in this paper are based on, and accompanied by, performance measurements of code used at CERN.