Read about the MANGO project and the the importance of FPGAs for High Performance Computing.
Read more on HPCwire
Latest news on EU heterogeneous HPC research and development was presented at the TERATEC 2016 Forum.
Software-level runtime management is inherently intertwined with system-wide control of physical parameters, primarily power consumption and temperature. To gain a holistic understanding of such deep interplay, MANGO will explore proactive thermal and power management in order to minimize energy consumption by creating for the first time a link between hardware and software effects, involving all layers modeling in HPC server, rack, and datacenter conception
MANGO architecture will rely on an innovative concept, where general-purpose compute nodes (GN) are finely intertwined with deeply heterogeneous acceleration nodes (HN) ensuring sustainable performance as high as 100 PFLOPS. GNs provide bulk high-performance computing power, also including highend graphics computing acceleration. On the other hand, HNs will be essentially on-node clusters of next-generation manycore chips coupled with deeply customized heterogeneous computing resources.
In addition to mere performance and power efficiency, it is of paramount importance to meet new nonfunctional requirements posed by emerging classes of applications. In particular, a growing number of HPC applications demand some form of time-predictability, or more generally Quality-of-Service (QoS), particularly in those scenarios where correctness depends on both performance and timing requirements and the failure to meet either of them is critical.