SCIENCE
Appro Cluster Engine (ACE) Management Software Unwraps New Features for the Xtreme-X Supercomputer Reducing Cost, Simplifying Administration
Driven by the Latest Intel Xeon processors, the Appro Xtreme-X Supercomputer Offers More Configuration Options and Improved Cluster Management
The Appro Xtreme-X Supercomputer is a highly scalable architecture that groups high performance servers together into a unified, fully integrated Scalable Unit (SU) that can be provisioned and managed as a stand-alone supercomputer. The Appro Xtreme-X Supercomputer is based on the Quad-Core Intel Xeon Processor offering interconnect fabric options for single rail InfiniBand Network for price/performance or dual rail configuration for performance/availability computing.
As CPU performance and memory bandwidth improves, I/O is becoming more of a performance bottleneck. In an effort to address this bottleneck and improve both performance and reliability, Appro and Intel have been working together to bring significant improvements to MPI enhancing the performance of dual rail configurations. In addition, the new Intel MPI 4.0 provides improvements for traditional single rail implementations allowing each rail to be configured separately; one for MPI traffic and the other for Storage I/O. The performance gains can often be measured right out of the box with this new MPI software.
"These latest enhancements to the Appro Xtreme-X Supercomputer allow organizations to deploy solutions that are flexible enough to adapt to their workflows and budgets while maximizing performance and productivity. With Appro Cluster Engine software enabling even large-scale systems to get up and running quickly, Appro can deliver those advantages faster so customers such as Renault F1 Team Computational Aerodynamics Research Center, Computational Sciences at the University of Tsukuba can realize a return on their investments sooner," said John Lee, VP of Advanced Technology Solutions for Appro.
Greater Manageability, Control and Reliability with ACE Cluster Management Software
Appro ACE Management software offers plug and play for Xtreme-X Supercomputers, meaning all of the network and server hardware is automatically discovered and monitored. With its diskless booting of compute servers, ACE allows 10 to 10,000 compute nodes to boot quickly and simultaneously. Individual clusters can also be independently configured and provisioned to enable workload and energy management with improved levels of control. Each application can see an optimized cluster environment freeing the administrators to manage the global compute environment and energy consumption.
ACE provides deep insight into the entire infrastructure, having precise control and ongoing optimization for expanding physical and virtual Linux environments. ACE provides customers with powerful tools to manage the software stack, including workload manager, resource management tools, remote control and advanced power management functions. ACE allows network and server hardware monitoring so problems are quickly identified and fixed and it automatically restarts jobs after a compute node failure. It also provides a redundant hierarchical system management with all system configuration data kept on the fault tolerant data base and file system on the management servers.
The new additional features include: Fully stateless compute nodes and dynamic instant provisioning, revision control system enabling multiple cluster SW revisions and versions, support for full installation for Red Hat or CentOS distribution, host configuration synchronization, automatic assignment of cluster host names and IP addresses, automatic management of host database (/etc/hosts), support for efficient global and cluster-level configuration files, integrated support and configuration for Grid Engine for batch queuing, resource management system and scheduling. Also, ACE now supports either local or remote disks. This capability with diskless boot provides users with the best of both worlds. The reliability and speed of diskless boot ensures that the system is ready quickly, while the local disks provide options for faster distributed disk access. This combination provides the added performance and reliability of ACE and the application flexibility of local disks.
The Appro Xtreme-X Supercomputer is Intel Cluster Ready Certified. The Intel Cluster Ready certification solution uses diagnostic tool methodology to enable ISV applications to run across certified cluster platforms providing an outstanding out-of-the-box experience. By Q1 2010, ACE Management Software including all added features will also be fully certified to Intel Cluster Ready specifications to simplify the deployment, usage and management of clustered computer systems and to ensure hardware and software interoperability.
“Time to deployed performance is key for many of today’s HPC customers. The combination of innovations in Appro’s Xtreme-X and ACE management solutions and the confidence of Intel Cluster Ready certification will simplify the out of box experience, achieving quicker time to results and computing,” said Richard Dracott, Intel’s General Manager of High Performance Computing. “In addition, we are enhancing Intel MPI to exploit the increased I/O provided by dual-rail InfiniBand. This capability complements the power of the Intel Xeon 5500 series processor and delivers overall higher application performance with no required application changes.”