| |
| |
3rd Party Software, Applications and Tools used on Quadrics connected systems
Cluster Products & Associated Linux Distributions
Quadrics drivers kernel modules and libraries are integrated into in the following cluster products and Linux distributions:
HP XC Cluster: The XC Cluster, part of the HP Unified Cluster Portfolio, is a scalable, Linux-based production resource for high performance computing, providing a consolidated platform for multiple applications and users, and mixed workloads, within a single managed environment, enabling high system utilization. The industrial-strength features of XC Clusters incorporate comprehensive support, ease of administration, a flexible and adaptive system architecture, robust resource management tools, breadth of applications and development tools, and high-performance global I/O capability.
The XC System Software includes, a full Linux operating system, cluster management capabilities for installation and on-going system administration, robust resource management and scheduling, and integrated HP-MPI. For customers needing big, fast, reliable, and shared storage, XC Clusters are available with the HP StorageWorks Scalable File Share (HP SFS) based on Lustre technology.
Bull Advanced Server: Linux for large-scale servers, High-Performance Computing (HPC).
This software suite features a complete programming environment (compilers, libraries, debuggers). Most components of such environment are open and non proprietary tools. The use of optimized Open Source software allows customers to work in full confidence with this environment since they are developing expertise and knowledge that could be reused. Moreover, a comprehensive set of scientific and technical applications have been ported to Intel® Itanium® 2, giving a wide choice of software to users of Bull HPC solutions. www.bull.com
Rocks: Open Source High Performance Linux Cluster Solution, cluster configuration and management package to help customers bring their own clusters online quickly. Quadrics software release supports both Rocks 3.3 and Rocks 4.1. http://www.rocksclusters.org
CHAOS: Clustered High-Availability Operating System, cluster operating environment based on Red Hat Linux, http://www.llnl.gov/linux/chaos/, used in all Livermore Linux clusters and maintained by the Livermore Linux Project Team. Chaos includes a range of innovative features:
· Modified kernel - to support high performance hardware, the Lustre file system, and other Livermore requirements.
· New packages - support added for cluster monitoring, system installation, power/console management, parallel job launch, resource management, compilers, etc.
· Modified packages - a few Red Hat packages are modified to implement timely bug fixes or enable them to work with CHAOS packages.
Cluster Management
LSF: Platform Computing's LSF provides software for managing and accelerating batch workload processing Platform's LSF HPC product is widely used on Quadrics connected systems including HP XC and AlphaServer SC, and Bull NovaScale clusters. Version 6.1 of LSF HPC is compatible with current Quadrics software releases. http://www.platform.com/
PBSPro: PBSPro workload manager from Altair Engineering. PBS Pro goes beyond OpenPBS with a features including greater fault tolerance, enhanced reliability, superior cluster support, and a long list of new features and enhancements. http://www.altair.com/
SGE: Sun Grid Engine. With Sun Grid Engine enterprises can monitor and select the optimal usage of computer resources on most commercial operating systems and platforms. Sun Microsystems
http://www.sun.com/software/gridware/index.xml
SLURM: Open-source resource manager for large and small Linux clusters.
SLURM has been deployed on all LLNL Linux clusters having Quadrics Elan switches since the summer of 2003. This includes IA32 and IA64 clusters having over 1000 nodes. Fault-tolerance has been excellent. Parallel job performance has also been excellent. The throughput rate of simple 2000 task jobs across 1000 nodes is over 12 per minute or under 5 seconds per job. http://www.llnl.gov/linux/slurm/
ISV Applications Software
LS-Dyna: general purpose transient dynamic finite element program capable of simulating complex real world problems. It is optimised for shared and distributed memory Unix, Linux, and Windows based, platforms. http://www.lstc.com/
NWChem: Computational chemistry package that is designed to run on high-performance parallel supercomputers as well as conventional workstation clusters. It aims to be scalable both in its ability to treat large problems efficiently, and in its usage of available parallel computing resources. NWChem has been developed by the Molecular Sciences Software group of the Environmental Molecular Sciences Laboratory (EMSL) at the Pacific Northwest National Laboratory (PNNL). http://www.emsl.pnl.gov/docs/nwchem/nwchem.html
Fluent: Fluent, the world's largest provider of computational fluid dynamics (CFD) software, and Bull provider of NovaScale computer systems, have recently announced their intent to work together to provide high performance computing solutions for the CAE community. As a new member of Fluent's partnership program, Bull will leverage its expertise in HPC applications and high-end servers to ensure optimum performance of Fluent's software on the full range of NovaScale servers.
http://www.fluent.com/
Ansys: ANSYS and Bull have completed certification of Ansys products on Novascale Clusters. ANSYS Mechanical and ANSYS Multiphysics are general-purpose finite element analysis software packages. ANSYS, Inc. develops a complete range of CAE products, but is perhaps best known for ANSYS Mechanical & ANSYS Multiphysics (see http://www.ansys.com). Details of supported platforms can be found at (http://www.ansys.com/services/hardware-support-db.htm). Details of supported interconnects can be found at http://www.ansys.com/services/ss-interconnects.htm
Debuggers and Performance Analysis Tools
TotalView: Etnus TotalView is a complete, advanced parallel debugger, supports most commonly used parallel programming platforms and parallel programming models, such as MPI, Threads, OpenMP, PVM, SHMEM, Fork/exec, Hybrid. Includes support for QsNet on Intel and AMD CPUs running Linux as well as AlphaServer systems running Tru64 UNIX. http://www.etnus.com/TotalView/index.html
Intel Trace: Optimised MPI Performance Analysis, Intel Trace speeds up parallel application runs, locates hotspots and bottlenecks, and increases productivity. Intel Trace is also compatible with SHMEM support for Quadrics hardware. http://www.intel.com/cd/software/products/asmo-na/eng/cluster/tanalyzer/index.htm
Parallel Programming Languages
HP-UPC: HP Unified Parallel C (UPC) commercial release of a UPC compiler for the industry, HP-UPC implements the full UPC specification and provides application-level access to the low-latency Quadrics interconnect. It also performs caching and pre-fetching optimisations that allow programs written in a simple style to obtain high performance. http://h30097.www3.hp.com/upc/
Berkeley UPC: Unified Parallel C (UPC) is an extension of the C programming language designed for high performance computing on large-scale parallel machines. The language provides a uniform programming model for both shared and distributed memory hardware. The programmer is presented with a single shared, partitioned address space, where variables may be directly read and written by any processor, but each variable is physically associated with a single processor. Unified Parallel C (UPC) Berkeley UPC runs natively on Quadrics interconnect for maximal communication performance. http://upc.lbl.gov/
Parallel Programming Libraries
Quadrics actively promotes the porting an optimisation of industry standard parallel programming interfaces to QsNet. See www.quadrics.com/software for details of Quadrics MPI, Quadrics Shmem and our resource management system (RMS).
HP-MPI: HP-MPI for HP-UX, Linux and HP Tru64 UNIX is a high performance and production quality implementation of the Message-Passing Interface (MPI) standard for HP servers and workstations.
http://h21007.www2.hp.com/dspp/tech/tech_TechDocumentDetailPage_IDX/1,1701,1238,00.html
A wide range of applications are certified for HP MPI, including:
ABAQUS' ABAQUS, Finite Element Analysis
Accelrys' CASTEP, DMoI3, MesoDyn ONETEP, Material Sciences
ACUSIM AcuSolve, Computational Fluid Dynamics
ANSYS's ANSYS, Finite Element Analysis (DDS)
AVL' s Excite, Computer Aided Engineering
CD-Adapco's STAR-CD, Computational Fluid Dynamics
CDH's AMLS, Noise Vibrate Analysis
Exa's PowerFlow, Computational Fluid Dynamics
ESI's PAM-CRASH, Computer Aided Engineering
Fluent's FLUENT, Computational Fluid Dynamics
LSTC's LS DYNA, Three-dimensional Finite Element Analysis
Mecalog's Radioss, Finite Element Analysis
MSC Software's Nastran, Mechanical Aided Design
UGS's NX Nastran, Mechanical Aided Design
SCM's ADF, Computational Chemistry
Scali MPI: Connect, Scali MPI Library (2005 Q4) http://www.scali.com Applications certified for Scali MPI include Eclipse, Fluent, Magma, LS-Dyna, TurboWorx, Abaqus
Intel MPI: Intel® MPI Library provides a high degree of interoperability with Intel® tools and architectures. A wide range of applications are certified for Intel MPI, for more information see:
http://www.intel.com/cd/software/products/asmo-na/eng/219786.htm
ARMCI: The purpose of the Aggregate Remote Memory Copy (ARMCI) library is to provide a general-purpose, efficient, and widely portable remote memory access (RMA) operations (one-sided communication) optimised for contiguous and non-contiguous (strided, scatter/gather, I/O vector) data transfers. In addition, ARMCI includes a set of atomic and mutual exclusion operations. The development ARMCI is driven by the need to support the global-address space communication model in context of distributed regular or irregular distributed data structures, communication libraries, and compilers. Developed by the Environmental Molecular Sciences Laboratory (EMSL) at Pacific North West National Lab. http://www.emsl.pnl.gov/docs/parsoft/armci/
Global Arrays: The Global Arrays (GA) toolkit provides an efficient and portable 'shared-memory' programming interface for distributed-memory computers. Each process in a MIMD parallel program can asynchronously access logical blocks of physically distributed dense multi-dimensional arrays, without need for explicit cooperation by other processes. Unlike other shared-memory environments, the GA model exposes to the programmer the non-uniform memory access (NUMA) characteristics of the high performance computers and acknowledges that access to a remote portion of the shared data is slower than to the local portion. The locality information for the shared data is available, and a direct access to the local portions of shared data is provided. Developed by the Environmental Molecular Sciences Laboratory (EMSL) at Pacific North West National Lab. http://www.emsl.pnl.gov/docs/global/
File Systems
Quadrics actively promotes the optimisation of third party filesystems for QsNet systems. The Elan kernel comms layer provides a high bandwidth transport between client and server nodes. Alternatively filesystems may be layered on IP over Elan.
Lustre: Lustre is a scalable, secure, robust, highly-available cluster file system. It is designed, developed and maintained by Cluster File Systems, Inc.The central goal is the development of a next-generation cluster file system which can serve clusters with 10,000's of nodes, petabytes of storage, move 100's of GB/sec with state of the art security and management infrastructure. http://www.lustre.org/
HP StorageWorks Scalable File Share A high-bandwidth, scalable storage appliance for Linux® clusters HP StorageWorks Scalable File Share (HP SFS) solves the I/O bottleneck typically found on Linux clusters requiring scalable storage. Using SFS20 disk arrays, HP SFS provides industry-leading price/performance across many terabytes of highly reliable, high-bandwidth storage. HP StorageWorks Scalable File Share (HP SFS) is a powerful file server that gives users of Linux clusters scalable storage that is easy to use and easy to administer. http://h20311.www2.hp.com/HPC/cache/276636-0-0-225-121.html
PVFS2: Parallel I/O continues to be a topic of active development. Recent years have seen the creation of many new options. Even with these new choices, certain factors remain constant. Parallel applications need a fast I/O subsystem. Clusters need a parallel file system that can scale as the number of nodes increases to the thousands and tens of thousands. PVFS2 is our answer. http://www.pvfs.org/pvfs2. The PVFS project is conducted jointly between The Parallel Architecture Research Laboratory at Clemson University and The Mathematics and Computer Science Division at Argonne National Laboratory.
|
> Legal
| |
|