Mellanox Introduces HPC-X Scalable Software Toolkit
Mellanox Technologies today announced the availability of HPC-X Scalable Software Toolkit, a comprehensive software suite for high-performance computing environments, which provides enhancements to significantly increase the scalability and performance of message communications in the network.
HPC-X provides complete communication libraries to support MPI, SHMEM and PGAS programming languages, as well as performance accelerators that take advantage of Mellanox’s scalable interconnect solutions. HPC-X provides support for other interconnect solutions as well, based on the Ethernet and InfiniBand standards.
HPC-X Scalable Software Toolkit components and accelerators include:
- Mellanox ScalableMPI – Message Passing Interface based on Open MPI
- Mellanox ScalableSHMEM – One-sided communications library
- Mellanox ScaleableUPC –UPC parallel programming language library based on Berkeley UPC
- Mellanox MXM – Messaging Accelerator optimized for underlying hardware
- Mellanox FCA – Fabric Collectives Accelerator supporting MPI-3
- Mellanox IPM – Integrated Performance Monitoring and additional benchmarking tools to help users fine tune their HPC environment
“HPC-X enables you to rapidly deploy and deliver maximum application performance without the complexity and costs of licensed third-party tools and libraries,” said Scot Schultz, director of HPC and technical computing at Mellanox. “Users can now solve their most complex problems in reduced time and scale their solutions more efficiently.”
The benefits of HPC-X can easily be seen even at lower node counts compared to running with a standard version of Open MPI. OpenFOAM, a popular open-source computational fluid dynamics code used by both commercial and academic organizations, demonstrated a 58 percent productivity boost with HPC-X at just 16 nodes. LS-DYNA, a popular commercial general-purpose finite element program used by industry leaders in automotive and aerospace to analyze complex vehicle designs, demonstrated improved scalability and 76 percent improved productivity with HPC-X at only 32 nodes. The effects that HPC-X will demonstrate can be even greater with larger scale system deployments.
