HP optimises multi-core for HPC

HP has introduced a programme to optimise multi-core technology in high-performance computing (HPC) applications. 

Unveiled at the International Supercomputing Conference in Dresden, HP's Multi-Core Optimisation Programme focuses on research for next-generation multi-core optimisation technologies. 

"We are entering a phase where exploiting the power of multi-core processors is critical for customers to accelerate the innovation and discovery that is directly tied to their business outcomes," said Winston Prather, vice president and general manager of HPC at HP.

HP's Multi-Core Optimisation Programme brings together developments from HP and its partners to enhance multi-core performance across a variety of HPC architectures, platforms and operating environments.

Initial partners include Accelrys, AMD, Intel, Karlsruhe Institute of Technology and TotalView Technology. 

"The emergence of multi-core technology allows HPC technology to better help solve grand challenges, and we have already seen an impact on large-scale numerical simulation in many fields of applications," said Professor Vincent Heuveline, a board member of the Steinbuch Centre for Computing at the Karlsruhe Institute.

HP has been working on the problem of multi-core optimisation since the first dual-core processors became available, using techniques such as load balancing between cores, memory and I/O.

The company recently enhanced its Simple Linux Utility for Resource Management tool to include multi-core application scheduling within HP's XC Linux cluster software.

HP has also expanded its development tools portfolio to cater for multi-core environments, including compilers, debuggers and maths libraries that are now available for a variety of processor and system architectures, operating systems and interconnects.

HP is also adding technical support to help optimise independent software vendors' applications for multi-core performance.