|
Sun Fire Visual Grid System Architecture White Paper
Sun Fire Visual Grid systems represent a new approach that combines the
resources and capabilities of multiple Sun Fire servers into a
tightly-integrated and unified high-performance visualization system.
These flexible systems scale in all dimensions-processors, memory,
storage, and displays - to deliver the capacity needed for the most
challenging computational problems with some of the industry's best
dynamic visual quality.
Sun Fire Superclusters for High Performance and Technical Computing
The Sun Fire Supercluster solution platform is a high performance and technical computing environment built around highly scalable server and network architectures. As a ready-to-deploy powerplant for the most demanding government, education, and research requirements, the Sun Fire Supercluster solution platform couples Sun's best HPC clustering technologies with the scalability and reliability of our high-end Sun Fire servers.
Sun Fire V880z Visualization Server Technical White Paper
White paper discusses how the Sun XVR-4000 graphics accelerator provides
high-performance graphics and new levels of visual quality for
multi-display environments. Packaged in compact and powerful Sun Fire
V880z visualization and Sun Fire Visual systems, the Sun XVR-4000
graphics accelerator brings visual clarity, high-resolution, and
interactive visualization solutions to workgroups at costs that are a
fraction of previous systems - helping organizations get the most from
their valuable technical resources.
Ultra-High Performance Communication
with MPI & the Sun Fire Link Interconnect
This paper presents a new low-latency system area network that provides the ultra-high
bandwidth needed to fuse a collection of large SMP servers into a capability
cluster. The network adapter exports a remote shared memory (RSM) model that
supports low latency kernel bypass messaging. The Sun MPI library uses the
RSM interface to implement a highly efficient memory-to-memory messaging protocol
in which the library directly manages buffers and data structures in remote memory.
This allows flexible allocation of buffer space to active connections, while
avoiding resource contention that could otherwise increase latencies. We discuss
the characteristics of the interconnect, describe the MPI protocols, and measure
the performance of a number of MPI benchmarks. Our results include MPI inter-node
bandwidths of almost 3 Gigabytes per second and MPI ping-pong latencies as low as
3.7 microseconds.
Sun Grid Engine Portal: Making Information and Applications Accessible
While Grid Computing aggregates resources, a mechanism is needed that enables the power
of the grid to be accessed with ease. Using portal technology, organizations can aggregate
information, standardize tools, and provide global access to HPTC applications and information,
giving organizations the agility they need to provide distributed shared services to users through a
single point of access. The Sun Grid Engine Portal (GEP) is an enabling technology aimed at making Grid Computing easily applied to enterprise challenges. Designed to support HPTC environments, it provides a distributed architecture that can deliver compute and data resources over the Web,
making resources available where and when they are needed most.
Grid Computing: A New Technology for the Advanced Web
For HPTC users, Grid Computing promises a new era of collaborative computing for individuals, workgroups, engineering and research teams alike. Grid Computing takes advantage of the opportunities provided by the Advanced Web to transform technical computing into an information utility. Computational Grids link multiple computing nodes spread across different computing domains. Additionally, Grids use the Internet technology to match computing resource supply with demand. This paper provides an introduction to Grid Computing, and describes how Sun is providing solutions that make Grid Computing a reality today.
The Sun Storage Value Proposition for High Performance Computing
High performance computing environments typically involve highly complex problems involving huge data sets and challenging data management requirements. Storage Area Networks (SANs) can provide a more efficient and cost-effective solution by providing a single, logical view of data and a single point for managing storage. The ability to share dataset across a SAN can greatly reduce lead time for solving complex HPTC problems. Sun's new fully integrated HPC SAN solution delivers data movement, data management, and data availability advantages to HPC that more traditional direct attached storage models just can't match.
|
|
 |
|
|
