Richard L. Graham

Open MPI: Goals, Concept, and Design of a Next Generation MPI Implementation

A large number of MPI implementations are currently available, each of which emphasize different aspects of high-performance computing or are intended to solve a specific research problem. The result is a myriad of incompatible MPI implementations, all of which require separate installation, and the combination of which present significant logistical challenges for end users. Building upon prior research, and influenced by experience gained from the code bases of the LAM/MPI, LA-MPI, and FT-MPI projects, Open MPI is an all-new, production-quality MPI-2 implementation that is fundamentally centered around component concepts. Open MPI provides a unique combination of novel features previously unavailable in an open-source, production-quality implementation of MPI. Its component architecture provides both a stable platform for third-party research as well as enabling the run-time composition of independent software add-ons. This paper presents a high-level overview the goals, design, and implementation of Open MPI.

Performance evaluation of the quadrics interconnection network

In this paper we present an in-depth description of the Quadrics interconnection network (QsNET) and an experimental performance evaluation on a 64-node AlphaServer cluster. We explore several performance dimensions and scaling properties of the network by using a collection of benchmarks, based on different traffic patterns. Experiments with permutation patterns and uniform traffic are conducted to illustrate the basic characteristics of the interconnect under conditions commonly created by parallel scientific applications. Moreover, the behavior of the QsNET under I/O traffic, and the influence of the placement of the I/O servers are analyzed. The effects of using dedicated I/O nodes or shared I/O nodes are also exposed. In addition, we evaluate how background I/O traffic interferes with other parallel applications running concurrently. The experimental results indicate that the QsNET provides excellent performance in most cases, with excellent contention resolution mechanisms. Some important guidelines for applications and I/O servers mapping on large-scale clusters are also given.

Open MPI: A High-Performance, Heterogeneous MPI

The growth in the number of generally available, distributed, heterogeneous computing systems places increasing importance on the development of user-friendly tools that enable application developers to efficiently use these resources. Open MPI provides support for several aspects of heterogeneity within a single, open-source MPI implementation. Through careful abstractions, heterogeneous support maintains efficient use of uniform computational platforms. We describe Open MPIs architecture for heterogeneous network and processor support. A key design features of this implementation is the transparency to the application developer while maintaining very high levels of performance. This is demonstrated with the results of several numerical experiments

A network-failure-tolerant message-passing system for terascale clusters

The Los Alamos Message Passing Interface (LA-MPI) is an end-to-end network-failure-tolerant message-passing system designed for terascale clusters. LA-MPI is a standard-compliant implementation of MPI designed to tolerate network-related failures including I/O bus errors, network card errors, and wire-transmission errors. This paper details the distinguishing features of LA-MPI, including support for concurrent use of multiple types of network interface, and reliable message transmission utilizing multiple network paths and routes between a given source and destination. In addition, performance measurements on production-grade platforms are presented.

Open MPI: a flexible high performance MPI

A large number of MPI implementations are currently available, each of which emphasize different aspects of high-performance computing or are intended to solve a specific research problem. The result is a myriad of incompatible MPI implementations, all of which require separate installation, and the combination of which present significant logistical challenges for end users. Building upon prior research, and influenced by experience gained from the code bases of the LAM/MPI, LA-MPI, FT-MPI, and PACX-MPI projects, Open MPI is an all-new, production-quality MPI-2 implementation that is fundamentally centered around component concepts. Open MPI provides a unique combination of novel features previously unavailable in an open-source, production-quality implementation of MPI. Its component architecture provides both a stable platform for third-party research as well as enabling the run-time composition of independent software add-ons. This paper presents a high-level overview the goals, design, and implementation of Open MPI, as well as performance results for its point-to-point implementation.

MPI Support for Multi-core Architectures: Optimized Shared Memory Collectives

With local core counts on the rise, taking advantage of shared-memory to optimize collective operations can improve performance. We study several on-host shared memory optimized algorithms for MPI_Bcast, MPI_Reduce, and MPI_Allreduce, using tree-based, and reduce-scatter algorithms. For small data operations with relatively large synchronization costs fan-in/fan-out algorithms generally perform best. For large messages data manipulation constitute the largest cost and reduce-scatter algorithms are best for reductions. These optimization improve performance by up to a factor of three. Memory and cache sharing effect require deliberate process layout and careful radix selection for tree-based methods.

Infiniband scalability in Open MPI

Infiniband is becoming an important interconnect technology in high performance computing. Efforts in large scale Infiniband deployments are raising scalability questions in the HPC community. Open MPI, a new open source implementation of the MPI standard targeted for production computing, provides several mechanisms to enhance Infiniband scalability. Initial comparisons with MVAPICH, the most widely used Infiniband MPI implementation, show similar performance but with much better scalability characteristics. Specifically, small message latency is improved by up to 10% in medium/large jobs and memory usage per host is reduced by as much as 300%. In addition, Open MPI provides predictable latency that is close to optimal without sacrificing bandwidth performance

Architecture of LA-MPI, a network-fault-tolerant MPI

Summary form only given. We discuss the unique architectural elements of the Los Alamos message passing interface (LA-MPI), a high-performance, network-fault-tolerant, thread-safe MPI library. LA-MPI is designed for use on terascale clusters which are inherently unreliable due to their sheer number of system components and trade-offs between cost and performance. We examine in detail the design concepts used to implement LA-MPI. These include reliability features, such as application-level checksumming, message retransmission, and automatic message rerouting. Other key performance enhancing features, such as concurrent message routing over multiple, diverse network adapters and protocols, and communication-specific optimizations (e.g., shared memory) are examined.

A case for standard non-blocking collective operations

In this paper we make the case for adding standard nonblocking collective operations to the MPI standard. The nonblocking point-to-point and blocking collective operations currently defined by MPI provide important performance and abstraction benefits. To allow these benefits to be simultaneously realized, we present an application programming interface for non-blocking collective operations in MPI. Microbenchmark and application-based performance results demonstrate that non-blocking collective operations offer not only improved convenience, but improved performance as well, when compared to manual use of threads with blocking collectives.

ConnectX-2 InfiniBand Management Queues: First Investigation of the New Support for Network Offloaded Collective Operations

This paper introduces the newly developed Infini- Band (IB) Management Queue capability, used by the Host Channel Adapter (HCA) to manage network task data flow dependancies, and progress the communications associated with such flows. These tasks include sends, receives, and the newly supported wait task, and are scheduled by the HCA based on a data dependency description provided by the user. This functionality is supported by the ConnectX-2 HCA, and provides the means for delegating collective communication management and progress to the HCA, also known as collective communication offload. This provides a means for overlapping collective communications managed by the HCA and computation on the Central Processing Unit (CPU), thus making it possible to reduce the impact of system noise on parallel applications using collective operations. This paper further describes how this new capability can be used to implement scalable Message Passing Interface (MPI) collective operations, describing the high level details of how this new capability is used to implement the MPI Barrier collective operation, focusing on the latency sensitive performance aspects of this new capability. This paper concludes with small scale bench- mark experiments comparing implementations of the barrier collective operation, using the new network offload capabilities, with established point-to-point based implementations of these same algorithms, which manage the data flow using the central processing unit. These early results demonstrate the promise this new capability provides to improve the scalability of high- performance applications using collective communications. The latency of the HCA based implementation of the barrier is similar to that of the best performing point-to-point based implementation managed by the central processing unit, starting to outperform these as the number of processes involved in the collective operation increases.