Does The Convey Supercomputer Represent A New Generation Of Supercomputer?

Read Complete Research Material



Does the Convey supercomputer represent a new generation of supercomputer?

Does the Convey supercomputer represent a new generation of supercomputer?

Introduction

A supercomputer is a computer which is at the frontline of current processing capacity specifically the speed of computer. These are new generation supercomputers with extra-ordinarily fast data processing speed whose number crunching power is measured in hundreds of billions of floating point operations (gigaflops). Parallel-processing technology is actually the basis of supercomputers. These computers utilize only few but very complicated programs in modeling neurological, meteorological, nuclear reactions and economy behavior phenomenon (Markoff, 2008).

Discussion

With a hybrid supercomputer, convey computer exploded into the HPC scene in which field programmable gate array (FPGA) accelerators and x64 processors are employed. With a new generation of FPGAs, convey computer is involved in upgrading its HC-1. Convey Computer's Hybrid Core-1 super is not similar to hybrid CPU-GPU machines but for some specific applications, more sense as an accelerator is made by a more expensive FPGA because in order to accelerate very particular algorithms and to have different personalities, it can be programmed. A Linux software stack is run by the original HC-1and a Xilinx Virtex-5 FPGA is placed inside of an old Xeon socket utilizing the front-side bus architecture from the Xeon 5300 generation (Morgan, 2010).

Both the newer Quick Path Interconnect and front side bus are licensed by convey. From the chip giant, they are used modern Xeon 5500, 5600, 6500, and 7500 processors. But QPI version of the machine is not yet implemented in it but it is expected to be done soon (Morgan, 2010).

Supercomputer with the HC-1ex announced at SC10 that convey is upgraded to the latest Virtex-6 FPGA offering the elements of the prior generation of FPGAs nearly four times. It is producing somewhere between two and three times more performance executing specific algorithms ...