November 13, 2011
Merrick3 High Performance Computing Platform
Using advanced FPGA technology Merrick3 delivers an unrivalled scalable processing platform. Used stand alone, in a large multi-board system, or as a personal computer accelerator, it can deliver a processing capability up to 1000 times that of the more conventional X86 processor.
The Merrick3 is based on an array of 24 Xilinx Spartan-6 FPGAs and 16 DDR3 memories. Supported by two further FPGAs offering high speed data communications and nearly instantaneous on the fly reconfiguration the platform delivers processing performance generally unobtainable in a standard personal computer.
Merrick3 is based on the standard PCI Express card format. The standard PCI Express interface offers high bandwidth data transfer between the board and a hosting motherboard. Using standard ATX motherboards, within standard PC cases, a small but high performance HPC system can be formed for difficult science and technology processing tasks. Going further, and for more advanced processing requirements, Enterpoint can supply turnkey solutions based on custom designed motherboards to construct systems of up to 10,000 boards and a resultant massive processing capability.
"The combination of low power and enormous logic capability of the Xilinx Spartan-6 FPGA family has allowed us to build a high performance processing (HPC) platform that delivers 1000X performance of a X86 processor whilst consuming a similar, or even smaller, operating power envelope. This not only brings an obvious operational cost saving to completing processing tasks but more significantly results in smaller lower cost cooling systems," says John Adair CEO of Enterpoint.
Lamachan2 Extreme I/O Solution
Using a Xilinx Spartan-6 FPGA to support a high performance PCI Express interface and large I/O count Lamachan2 offers an extreme I/O solution for High Performance Computing (HPC) systems. It is also capable of forming the basis of a low cost processing accelerator or even a combined I/O / accelerator processing solution.
Lamachan2 supports I/O solutions up to and including 266 I/O. Plugging in standard DIL modules, available directly from Enterpoint, allows the rapid construction of a wide range of I/O, communications, and processing solutions. Typically systems formed this way could include 256 200KHz ADC channels, communication interfaces up to 8 gigabit Ethernet interfaces, or even systems based on up to four XC6SLX150 X2 Coprocessors for low cost, high performance, algorithm processing solutions.
Lamachan2 can also be used as a support board to systems based on Enterpoint's high performance Merrick3 boards and other future products in this product range. A high speed expansion interface, with high bandwidth capability, allows high speed, low latency, connections to adjacent Merrick3 boards without using a host motherboard for data routing. The expansion interface can also be used to link together multiple Lamachan2 boards to form extended I/O or communication interfaces.
"Lamachan2 brings offers an I/O capability previously unobtainable with standard I/O capture products. The enormous logic capability of the Xilinx Spartan-6 FPGA, used at the heart of Lamachan2, not only allows us to offer a standard product with enormous potential but also allows low cost semi-custom solutions with extended high performance features," says John Adair CEO of Enterpoint.
Merrick3 and Lamachan2 will be showing at SC11, November 15th to 17th, in Seattle USA and is available to order immediately. For further information, please contact email@example.com.
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?