At the AMD Torrenza Initiative seminar this week, Celoxica showed the results of a seismic processing benchmark conducted for an undisclosed client that achieved a claimed 28 fold performance improvement. Celoxica’s ‘accelerated computing solution’ consists of a ‘field programmable gate array’ (FPGA), a hardware number crunching co-processor and HyperTransport interconnect. Celoxica enables FPGA co-processing an algorithm library and C compiler.
Jussel
Celoxica VP Jeff Jussel said, ‘We have demonstrated accelerated computing for customers in industries such as oil exploration, financial analysis and life sciences. We can now also quantify the financial advantages of our FPGA co-processing solution with a new calculator that figures ROI on hardware and support costs, power savings and improved revenue opportunity from increased computing performance.’
Goddard
AMD senior director Mike Goddard added, ‘HyperTransport and AMD64 provide a co-processor expansion capability that enables this FPGA-based acceleration. Celoxica’s integrated acceleration hardware and API demonstrate the successful of our Torrenza initiative.’ AMD’s Torrenza initiative improves support for specialized coprocessors in 64bit AMD Opteron-based systems.
© Oil IT Journal - all rights reserved.