The MPI design and its main implementations have proved surprisingly scalable. MPI has been robust, and been able to evolve, without fundamentally changing the model and specification. For this and many other reasons, MPI is currently the de-facto standard for HPC systems and applications. However, there is a need to re-examine of the Message-Passing model for extreme-scale systems that are characterized by asymptotically decreasing local memory and highly localized communication networks. Likewise, there is a need for exploring innovative and potentially disruptive concepts and algorithms beyond those taken by the MPI 3.1 standard. The aim of this workshop is to bring together developers and researchers to present and to discuss innovative algorithms, concepts, implementation and applications of the Message-Passing programming model. With the ExaMPI workshop series, we aim at creating a forum for open and potentially controversial discussions on the future of MPI in the exascale era and on the MPI requirements to support applications at exascale.