SiPearl se asocia con Intel para ofrecer una supercomputadora a exaescala en Europa
To enable this powerful combination, SiPearl plans to use and optimize for its Rhea microprocessor the open and unified programming interface, oneAPI, created by Intel. Using this single solution across the entire heterogeneous compute node, consisting of Rhea and Ponte Vecchio, will increase developer productivity and application performance.
It will also capitalize on the relevance of CXL, the low latency interconnection and optimized sharing of memory resources between microprocessors and their peripherals – such as accelerators.
“We are delighted that Ponte Vecchio, based on the Intel Xe HPC architecture, was selected by SiPearl as a high performance computing accelerator to equip the first exascale supercomputers in Europe. Intel is proud to support SiPearl in adopting the oneAPI Unified Programming Interface to increase developer productivity. As we implement our IDM 2.0 strategy to bring cutting-edge technologies to our European partners, we see great potential in our work with SiPearl, ” dijo jeff mcveigh, Intel vice president and general manager of his division. Super Compute.
“We are delighted with this collaboration with Intel, whose Ponte Vecchio accelerator, optimized for high performance computing, is extremely innovative. The combination of Rhea, the European high-performance microprocessor we are designing, with Intel’s Ponte Vecchio accelerator using the oneAPI programming interface, will allow us to rapidly develop heterogeneous compute nodes to meet supercomputing needs. exascale in Europe “, concludes Philippe Notton, President and Founder of SiPearl.