High Performance Message Passing Interface with IBM Platform MPI
IBM Platform MPI is a high performance, production–quality implementation of the Message Passing Interface (MPI)—supporting the broadest range of industry standard platforms, interconnects and operating systems to help ensure that parallel applications can run on any platform.
Optimized application parallelization
The need for speed
Application parallelization, especially in scientific research or computational science, is necessary to run compute intensive applications. However, a sub-optimized environment jeopardizes competiveness and slows your speed to results. MPI is essential to accelerate applications. Developers want an MPI library that provides:
Accelerate applications with a high performance Message Passing Interface (MPI)
Broadest range of support
IBM Platform MPI is a high performance, production–quality implementation of the Message Passing Interface (MPI) to accelerate applications in distributed computing environments. It supports the broadest range of industry standard platforms, interconnects and operating systems helping ensure that parallel applications can run anywhere. Platform MPI helps clients:
Rather than have the application handle architectural differences in your infrastructure, Platform MPI manages them for you—no need to write multiple versions of the application to account for different interconnects. It also incorporates advanced CPU affinity features, dynamic selection of interface libraries, superior workload manager integrations and improved performance and scalability.
Accelerate time to result
Platform MPI designed to:
Reduce development effort & enable applications to use the latest technologies
Focus on portability
Platform MPI allows developers to build a single executable that transparently leverages the performance features of any type of interconnect, resulting in applications with optimal latency and bandwidth for each protocol. This reduces development effort and enables applications to use the latest technologies on Linux or Microsoft Windows without the need to re-compile and re-link applications. Automatic interconnect detection and selection can happen at runtime, while high availability features include interconnect fail-over and fault tolerance. Platform MPI helps take the risk out of mission critical high performance technical computing deployments by being pre-qualified and certified on a variety of platforms—continuing to be fully backward compatible with HP-MPI, and the applications it supports.
IBM Platform MPI – Standard Edition
IBM Platform MPI – Standard Edition is a fully featured version of Platform MPI, which is designed to deliver application performance at scale. It is a high performance, production-quality implementation of MPI designed to accelerate application performance in distributed computing environments.
IBM Platform MPI - Community Edition
IBM Platform MPI - Community Edition is ideal for smaller technical computing clients, as well as developers and business partners who serve these clients. It can be setup quickly, is easy-to-use, and includes self-managing capabilities. IBM Platform MPI - Community Edition embodies all of the core features of the more scalable Platform MPI - Standard Edition.
IBM Platform MPI - Community Edition is available for download and deployment at no charge. An optional yearly subscription is available for users who require technical support or greater scalability.
IBM Business Partners and Independent Solution Vendors (ISV) who want to embed or include Platform MPI as part of their solutions can distribute IBM Platform MPI - Community Edition at no-charge upon registering with IBM. Please contact email@example.com for details. An optional yearly support subscription is available for ISVs.
Community Edition(no support)
Community Edition(with support)
|OS Support||x86-64 Linux
|Scalability||4,096 ranks||8,192 ranks||>64k ranks|
|Additional Functionality||-||-||>HA Support
Cluster Test Tools
|Support||None||Subscription||>Perpetual or Subscription|
Connect with Platform Computing