I haven't taken a look at the patch but I don't really know how it can be possible to speed things up without invoking UB (I will take a look at it today).

The obvious way to get the same performance as C is just to use a boost::container::vector instead of std::vector. Std::vector is safer but it is just not designed to support this use case "as efficient as possible" (it should still be pretty fast tho).

On Monday, August 10, 2015, Utkarsh Bhardwaj <hap_tork@hotmail.com> wrote:
Has this been incorporated in the newer versions of boost.mpi library (1.5+)?


_______________________________________________
Boost-mpi mailing list
Boost-mpi@lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-mpi


--
Dipl.-Ing. Gonzalo Brito Gadeschi
Institute of Aerodynamics and Chair of Fluid Mechanics
RWTH Aachen University
Wuellnerstraße 5a
D-52062 Aachen
Germany
Phone: ++49-(0)241-80-94821
Fax: ++49-(0)241-80-92257
E-mail:  g.brito@aia.rwth-aachen.de
Internet: www.aia.rwth-aachen.de