Scientists Build Inexpensive Super Computer Based on PLDA Solution
Project short-listed for 2006 Gordon Bell Price / Performance prize
San Jose, Calif., January 17, 2007 — PLDA today announced that Atsushi Kawai, a professor with the Saitama Institute of Technology, and Toshiyuki Fukushige, a scientist with the K&F Computing Research Co., are finalists for the Gordon Bell price / performance prize, which judges genuine scientific applications based on a performance ratio measured in sustained flops per dollar of acquisition cost.
Professors Kawai and Fukushige built two low-cost systems based on PLDA products designed to execute astrophysical N-body simulation, one of the most widely used techniques for exploring the formation and evolution of astronomical objects. Their article is published with the Super Computing International Conference for high-performance computing, networking, storage, and analysis (http://sc06.supercomputing.org/).
"We chose PLDA because their products deliver the best performance / price value," explains professor Kawai. "Our system performs a cosmological N-body simulation with 2.1 million particles with a sustained performance of 22.59 Gflops averaged over 2.95 hours. The total price of our system is $2,363, which yields a price per performance ratio of $105 / Gflops."
The first of the two systems consists of PLDA's PCIe XpressLite GX design kit, PLDA's PCIe XpressLite IP controller, and a host PC. The second system incorporates PLDA's PCI-X IP controller. Both systems yield similar results.
The XpressLite GX Design Kit, based on the Altera Stratix™ GX FPGA, is a low-cost PCI-SIG® compliant solution designed for x4 Endpoint applications. The board is delivered with the PCI-SIG compliant XpressLite IP Controller free of charge. Other deliverables include a PCI Express Testbench, reference design, Software Development Kit (SDK), and technical support provided directly by the IP design team.
The professors' article, titled $105/GFLOP Astrophysical N-Body Simulation with a Reconfigurable Add-in Card and a Hierarchical Tree Algorithm, is available here http://www.kfcr.jp/gb06paper.pdf.
About PLDA
PLDA designs and sells a wide range of ASIC and FPGA interfacing solutions for the PCI and derivative markets (PCI Express, PCI-X, and PCI). The company offers complete solutions, including IP cores, hardware, software, consulting services, and comprehensive technical support provided directly by the IP designers. For additional information about PLD Applications, please visit www.plda.com.
About the Gordon Bell Prize
The Gordon Bell Prizes recognize groundbreaking achievements for performance and scalability in several categories on genuine and specific scientific applications. In recent years, awards have been granted in four categories: Peak performance based on sustained floating point operations per second, Price per performance ratio measured in sustained flop/s per dollar of acquisition cost, Special accomplishment for innovation in scalable implementation, and Scalability achieved through language constructs.
|
Related News
- PsiQuantum and GLOBALFOUNDRIES to Build the World's First Full-scale Quantum Computer
- NEC Selects Synopsys ZeBu Server 4 Emulation Solution for Super Computer Verification
- Micro/sys, Inc. Uses Xilinx Spartan-6 FPGA to Build Integrated FPGA & ARM Cortex-A8 Processor Single-Board Computer
- GDA Technologies Licenses HyperTransport IP cores to Super computer Pioneer Cray, Inc.
- Logic Fruit Launches the Advanced Kritin iXD 6U VPX Single Board Computer (SBC)
Breaking News
- Ubitium Debuts First Universal RISC-V Processor to Enable AI at No Additional Cost, as It Raises $3.7M
- TSMC drives A16, 3D process technology
- Frontgrade Gaisler Unveils GR716B, a New Standard in Space-Grade Microcontrollers
- Blueshift Memory launches BlueFive processor, accelerating computation by up to 50 times and saving up to 65% energy
- Eliyan Ports Industry's Highest Performing PHY to Samsung Foundry SF4X Process Node, Achieving up to 40 Gbps Bandwidth at Unprecedented Power Levels with UCIe-Compliant Chiplet Interconnect Technology
Most Popular
- Cadence Unveils Arm-Based System Chiplet
- CXL Fabless Startup Panmnesia Secures Over $60M in Series A Funding, Aiming to Lead the CXL Switch Silicon Chip and CXL IP
- Esperanto Technologies and NEC Cooperate on Initiative to Advance Next Generation RISC-V Chips and Software Solutions for HPC
- Eliyan Ports Industry's Highest Performing PHY to Samsung Foundry SF4X Process Node, Achieving up to 40 Gbps Bandwidth at Unprecedented Power Levels with UCIe-Compliant Chiplet Interconnect Technology
- Arteris Selected by GigaDevice for Development in Next-Generation Automotive SoC With Enhanced FuSa Standards
E-mail This Article | Printer-Friendly Page |