Open MPI Publications

If you are looking for general-purpose Open MPI citations, please use the following (other, more specific papers are also available):

Open MPI Bibliography and Papers:

2012

• MPI Library and Low-Level Communication on the K computer
  Description of the MPI library and low-level communication on the K computer

• Java for High Performance Computing
  Presents an overview of the mpiJava bindings used in Open MPI.

• mpiJava 1.2: API Specification
  Presents an overview of the mpiJava API used in Open MPI.

• The Architecture of Open Source Applications, volume 2
  Chapter 15 presents an architectural overview of Open MPI.

2011

• Supercomputing 2011 slides
  Slides from the Open MPI State of the Union BOF at SC'11

• Locality-Aware Parallel Process Mapping for Multi-Core HPC Systems
  Locality-Aware Parallel Process Mapping for Multi-Core HPC Systems

• A Log-Scaling Fault Tolerant Agreement Algorithm for a Fault Tolerant MPI
  A Log-Scaling Fault Tolerant Agreement Algorithm for a Fault Tolerant MPI

• Building a Fault Tolerant MPI Application: A Ring Communication Example
  Building a Fault Tolerant MPI Application: A Ring Communication Example

• Preserving Collective Performance Across Process Failure for a Fault Tolerant MPI
  Preserving Collective Performance Across Process Failure for a Fault Tolerant MPI

2010

• Supercomputing 2010 slides
  Slides from various presentations about Open MPI or by the Open MPI team members.

• Checkpoint/Restart-Enabled Parallel Debugging
  Checkpoint/Restart-Enabled Parallel Debugging

• hwloc: a Generic Framework for Managing Hardware Affinities in HPC Applications
  Analysis, description, and results of using hardware affinity using the hwloc software package.

2009

• Supercomputing 2009 slides
  Slides from various presentations about Open MPI or by the Open MPI team members.

• State-of-the-Art in Parallel Computing with R
  This paper provides an overview of using MPI with R

• An Efficient Open MPI Transport System for Virtual Worker Nodes
  This paper discusses a shared memory Xen-based transport for Open MPI. It also provides a good technical description of Open MPI's BTL plugins.

• Interconnect agnostic checkpoint/restart in Open MPI
  Investigation of interconnect support for checkpoint/restart process fault tolerance in Open MPI.

2008

• Supercomputing 2008 slides
  Slides from various presentations about Open MPI or by the Open MPI team members.

• Evaluating Sparse Data Storage Techniques for MPI Groups and Communicators
  This paper evaluates various storage formats for groups and communicators for minimizing the memory footprint of MPI applications.

• X-SRQ - Improving Scalability and Performance of Multi-Core InfiniBand Clusters
  Using XRC - new transport protocol based on B-SRQ

2007

• An Extensible Framework for Distributed Testing of MPI Implementations
  An overview of the MPI Testing Tool (MTT).

• Investigations on InfiniBand: Efficient Network Buffer Utilization at Scale
  Using multiple queue pairs and pre-posted buffer sizes to get more efficient registered memory utilization.

• A Case for Standard Non-Blocking Collective Operations
  We draw the case for non-blocking collective operations in MPI and show several application benchmarks to support our claims.

• An Evaluation of Open MPI's Matching Transport Layer on the Cray XT
  This paper describes the implementation of Open MPI on the Cray XT.

• Analysis of Implementation Options for MPI-2 One-Sided
  An analysis of the implementation of MPI-2 one-sided functionality in Open MPI

• Memory Debugging of MPI-Parallel Applications in Open MPI
  Adding MPI semantics for memory debugging using valgrind

• Open MPI Community Meeting BOF
  Open MPI BOF at ISC 2007

• A practically constant-time MPI Broadcast Algorithm for large-scale InfiniBand Clusters with Multicast
  A reliable multicast-based broadcast algorithm for InfiniBand.

• The Design and Implementation of Checkpoint/Restart Process Fault Tolerance for Open MPI
  A high level description of the checkpoint/restart process fault tolerance infrastructure in Open MPI.

• Efficient Message Passing on Multi-Clusters: An IPv6 Extension to Open MPI
  General description and first benchmarks of the Open MPI IPv6 extension.

2006

• Supercomputing 2006 slides
  Slides from various presentations about Open MPI or by the Open MPI team members.

• Open MPI: A High-Performance, Heterogeneous MPI
  An overview of Open MPI's architecture, emphasizing the heterogeneous-enabled design.

• Self-Healing Network for Scalable Fault Tolerant Runtime Environments
  Describes fault-tolerant protocols for runtime environments.

• Approaches for Parallel Applications Fault Tolerance
  Invited talk about directions for application fault tolerance at scale

• High Performance RDMA Protocols in HPC
  An overview of Open MPI's RDMA-based point-to-point messaging layer (PML).

• Implementation and Usage of the PERUSE-Interface in Open MPI
  Describes the implementation and experience of the PERUSE specification in Open MPI.

• An Interface to Support the Identification of Dynamic MPI 2 Processes for Scalable Parallel Debugging
  Describes a proposed mechanism for attaching debuggers to MPI-2 dynamic processes.

• Scalable Fault Tolerant Protocol for Parallel Runtime Environments
  Robust runtime environment message passing for damaged and incomplete topologies.

• Optimizing a Conjugate Gradient Solver with Non Blocking Collective Operations
  Using a real-world application to analyze the applicability of non-blocking collective operations for MPI.

• MPI Collective Algorithm Selection and Quadtree Encoding
  Methods of selection of the "best" collective algorithms at run-time.

• A Checkpoint and Restart Service Specification for Open MPI
  An overview of the "CRS" framework and how it is used in Open MPI

• InfiniBand Scalability in Open MPI
  Analysis of scalability issues with InfiniBand and Open MPI

• Open MPI Developer's Workshop
  Slides from the Open MPI 2006 Developer's Workshop

2005

• Open MPI and OpenIB: Past, Present, and Future
  Talk presented at the OpenIB Workshop, 2006

• Implementing a Hardware-Based Barrier in Open MPI: A Case Study
  Adding hardware-aware barriers into Open MPI.

• Supercomputing 2005 slides
  Slides from various presentations about Open MPI or by the Open MPI team members.

• Implementation of Open MPI on Red Storm
  Presents an overview of Open MPI on the Red Storm platform.

• A Case for New MPI Fortran Bindings
  Rationale for new MPI Fortran bindings, including a specific propsal.

• Analysis of the Component Architectue Overhead in Open MPI
  Presents a detailed analysis of the [lack of] overhead generated Open MPI's component architecture.

• The Open Run-Time Enviroment (OpenRTE): A Transparent Multi-Cluster environment for High-Performance Computing
  A description of the run-time environment (ORTE) sub-project in Open MPI

• Open MPI: A Flexible High Performance MPI
  Overview of Open MPI and the OB1 point-to-point architecture.

• A Practical Approach to the Rating of Barrier Algorithms using the LogP Model and Open MPI
  Barrier measurements and performance analysis.

2004

• Open MPI: Goals, Concept, and Design of a Next Generation MPI Implementation
  Presents an overview of the Open MPI project.

• TEG: A High-Performance, Scalable, Multi-Network Point-to-Point Communications Methodology
  Overview and descrption of the point-to-point message passing design in Open MPI, focusing on the TEG PML component.

• Open MPI's TEG Point-to-Point Communications Methodology: Comparison to Existing Implementations
  Preliminary performance results using a pre-release version of Open MPI.

• The Component Architecture of Open MPI: Enabling Third-Party Collective Algorithms
  General description of the MPI Component Architecture, focusing on the details of the collective algorithm framework in Open MPI.