|
 1.866.627.6951
 1.800.872.0040
(Support)
 Email
|
|
MultiCore Plus OpenMPI-PE
MultiCore Plus® OpenMPI-PE incorporates the open-source library Open MPI, enhanced for better throughput and latency, along with Mercury’s OFED provider software module for serial RapidIO®. OpenMPI-PE has been performance-enhanced by providing an implementation of open-standard MPI allocation functions to provide physically contiguous buffers, resulting in significantly improved application performance. OpenMPI-PE combines the productivity and ease-of- use of the MPI protocol, designed for data-intensive parallel computing, with the speed of an optimized fabric, designed for embedded computing.
Features
- Portability with additional high-speed RDMA-based fabrics (such as Ethernet)
- MPI integration with ICS for ultimate inner loop performance
- Alternative MPI implementations (open source, commercial) and other OFED-dependent software supported as an optional service
- Interoperation between PowerPC™-based and x86-based
Benefits
- Performance-enhanced MPI on serial RapidIO
- MPI portability provides code compatible across fabrics
- Zero-copy RDMA message-passing support with application buffers
- Enhanced MPI allocation provides contiguous memory to enable higher performance communication
- Optimized memory copy implementations added to Open MPI and OFED™ for higher performance shared memory communication
- RapidIO® throughput and low latency with minimal overhead
Specifications
System Requirements
Ensemble™ Series platforms with MultiCore Plus software
Linux OS
864x PPC Serial RapidIO
x86 Intel® Serial RapidIO Packaging
Included in MultiCore Plus Software Linux Pro Edition and Ultimate Edition
|
|
|