A Novel Processor Architecture for FPGA Supercomputing A Novel Processor Architecture for FPGA Supercomputing
Ever-increasing computing performance is the main development driver in the world of supercomputing. Researchers have long looked at using FPGAs for co-processors as a promising way to accelerate critical applications. FPGA supercomputing is now generating more interest than ever, with major system vendors offering general-purpose computers equipped with the largest Xilinx Virtex-4 FPGAs, thus making the technology accessible to a wider audience.
There are compelling reasons to give serious consideration to FPGA supercomputing, especially given the significant performance benefits you can achieve from using FPGAs to accelerate applications. Typical applications can be accelerated 10-100x compared to CPUs, making FPGAs very attractive from a price/performance perspective.
What is even more important is that FPGAs use only a fraction of the power per computation compared to CPUs. Increased computing power with traditional clusters requires more and more electrical power. Today, the cooling problems in large computing centers have become major obstacles to increased computing performance.
Please disable any pop-up blockers for proper viewing of this Whitepaper.