DRC Computer Corporation, a leading provider of dynamically reconfigurable coprocessor modules, announced the availability of its Reconfigurable Processor Unit (RPU), the RPU110-L200. The RPU110-L200 includes an independent memory controller that provides greater memory bandwidth for maximum data flow. It also incorporates the largest FPGA available to deliver the greatest usable logic capacity for improved code efficiency and accelerated performance of compute-intensive applications.
DRC’s RPU110–L200 module plugs directly into an open processor socket in a multi-way AMD Opteron(TM) system to provide direct access to adjacent double data-rate (DDR) memory and Opteron processors at HyperTransport(TM) speed and nanosecond latency. This tight coupling between the central processing unit (CPU) and memory eliminates bandwidth and latency issues and provides a general-purpose system with supercomputer capability by running computations in hardware. Resulting speeds can be 10 to 20 times faster than if the subroutines were run in software. In addition, the RPU110-L200 is inherently efficient, requiring less power and space and reducing the number of nodes required to do the same work.
“DRC is working with Cray to deliver RPUs that offer efficient high performance computing alternatives to our customers. We believe that performance gains of 10 to 100X are possible on a number of applications when DRC’s RPUs are integrated into Cray’s systems,” says Jan Silverman, senior vice president, corporate strategy and business development for Cray. “This demonstrates yet another step forward in our Adaptive Supercomputing vision where multiple processing technologies are transparently available to the application.”
DRC offers a complete solution that addresses the critical performance and reliability needs of high-performance computing while improving overall data center efficiency. Generally, performance increases are at the expense of efficiency, but with DRC’s solutions, which combine a minimum-occupancy physical interface and controllers with maximum memory bandwidth, the exact opposite is true. DRC’s RPU110-L200 reduces power consumption, heat emissions, and space requirements in today’s data centers and delivers significant performance gains through RPU acceleration.
“The demands for high performance computing are increasing as business becomes even more global and competitive pressures increase,” explains Clay Marr, VP of sales and marketing at DRC Computer. “Markets such as financial services, oil and gas, and life sciences are demanding more powerful computing from their standards-compliant systems but also want to conserve energy — DRC meets both objectives.”
The RPU110-L200 features on-board DDR2 memory, any of three HyperTransport bus interfaces and the largest FPGAs on the market, including the Xilinx Virtex(TM)-4 LX160, and Virtex-4 LX200 FPGAs. The RPU110-L200 enables greater application acceleration in three ways: first, with three times more compute resources in the L200, more of any given application can run on the RPU; second, with 2GB of on-board RPU memory plus the 8GB of motherboard memory, the RPU110-L200 delivers more than 17GB/second of system memory bandwidth, three times the stand-alone CPU bandwidth; and third, up to 3.2GB of usable system bus bandwidth is now available.
About DRC Computer Corporation
DRC Computer Corporation is a leading provider of dynamically reconfigurable coprocessor modules that plug directly into server or workstation processor slots. With the DRC Reconfigurable Processor Unit (RPU) and DRC Development Systems, DRC addresses the HPC industry’s growing physical limitations of compute power, heat, clock speed, and density. Founded in 2004, DRC is headquartered in Sunnyvale, California. The company has received funding from Strategis Early Ventures and Capital Valley Ventures.