ISC’11 show preview
Steinbuch Centre for Computing (SCC) is the information technology centre of the Karlsruhe Institute of Technology (KIT). It will be
presenting its Large Scale Data Facility (LSDF), a concept for the management, archiving and analysis of scientific data. LSDF supports the complete lifecycle of data and is available to all scientific disciplines. The production started by storing images from zebrafish embryos taken by a High Content Screening microscope. KIT develops new methods which allow scientists to work on hundreds of thousands of images in parallel. Further topics at the booth will be the
The two key aspects for DRAM in HPC environments are performance/Watt and GB/core, and during the ISC 2011 show, Samsung will demonstrate its Green DDR3 technology. 30nm class Samsung Green DDR3 allows for significant savings in energy consumption while offering higher performance per Watt. Samsung 8GB and 16GB modules offer low power consumption and greatly reduced heat dissipation. Particularly in HPC environments, a ‘one DIMM per channel’ configuration provides the best system performance results. The company’s high-density modules prevent performance gains won by an expansion of memory capacity from being sacrificed by lower memory clock speeds due to ‘more than one DIMM per channel’ deployments. Live demonstrations will be taking place at booth 502.
www.samsung.com/greenmemory
Grid Computing Centre Karlsruhe (GridKa), the German Tier-1 centre for the Large Hadron Collider, as well as high-performance computing as a service, and the latest cloud computing research results at SCC. The results of advanced 100 GE hardware and protocol tests will also be presented.
www.scc.kit.edu
ScaleMP’s software-based vSMP Foundation solutions enable users to aggregate up to 128 commodity servers as a single, virtual SMP system of up to 8,192 cores and 64TB of shared RAM. These solutions deliver a high core count – in the Tflop range – and large amounts of shared memory within a single execution environment, providing scaling and performance. ScaleMP will be at booth 302.
www.scalemp.com
Schroff solutions feature standard, off-the-shelf and customer-specific integrated enclosures and front panels for open architecture
bus-based applications, including VME, VME64x, VITA31.1 (VXS), VITA 41, CompactPCI, VXI, MicroTCA, AdvancedMC and AdvancedTCA. The multifunctional cabinet and cooling platforms, combined with xTCA systems, deliver a solution that can be tailored to individual requirements.
www.schroff.co.uk
30 SCIENTIFIC COMPUTING WORLD
Solarflare will showcase its OpenOnload application acceleration middleware that increases message rates for HPC applications two to threefold, while reducing application latency by 50 per cent. New features include affinity driver to map applications to specific server cores, PCIe bus optimisations to reduce latency, acceleration of the Linux Pipe command, and a socket and stack control API that enables customers to control and fine tune acceleration options on a per-socket basis. Solarflare will also showcase its scalable
vNIC architecture (SVA) support of single root I/O virtualisation (SR-IOV) for Redhat Linux kernel-based virtual machines (KVMs) and Citrix XenServer 6.0. The company’s SR-IOV solution offers IT and datacentre managers the ability to improve their virtualised network I/O performance, while maintaining support of critical functions provided by the hypervisor. By using Solarflare OpenOnload middleware in combination with Solarflare SFN5122F 10GbE SFP+ server adapters, response times can be improved by 51 per cent for memcached queries.
www.solarflare.com
Symmetric Computing will be showcasing its line of supercomputers. The company’s Trio Departmental SuperComputing is an easy-to-program, large shared memory, single software image, Linux server with a one Tflop rating (SPECfp-2006 of 1602). Programmers can use standard threading packages to access all 96 Trio cores and up to 1.5 Tbytes of memory. Programmers don’t need to be concerned with Message Passing Interface programming, and there is no need to build complex file-access program components; programmers can just read a big dataset into memory and access it as an array. The company’s patent-pending DSMP extension to the Linux kernel has been combined with off-the-shelf server blades.
www.symmetriccomputing.com
Whamcloud implements and supports Lustre solutions in HPC centres around the world. Lustre is utilised in more than half of the TOP500 supercomputing sites today and is thought to be the best evolutionary technology for addressing many of the exascale issues of tomorrow. Eric Barton, Whamcloud CTO, will be co-hosting a Birds of a Feather (BoF) session entitled ‘Lustre for HPC
from Petascale to Exascale’ on Wednesday 22 June at 2.30pm to 3.15pm. Barton will be joined by co-host Jacques-Charles Lafoucriere, EOFS, in a discussion of the features under development and the future roadmap of this critical technology.
www.whamcloud.com
www.scientific-computing.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52