DENVER, Nov. 14, 2017 — At the Supercomputing Conference 2017, E4 Computer Engineering announced that D.A.V.I.D.E. entered for the second time the prestigious TOP500 and Green500 lists, ranking respectively position #440 and #18. D.A.V.I.D.E., a multi-node cluster based on IBM POWER8 servers and NVIDIA Tesla P100 data center GPUs, features two relevant innovations that enhance whole system capabilities in terms of performance and energy-efficiency. D.A.V.I.D.E. entered the TOP500 and GREEN500 list in June 2017 in its air-cooled version, while the current version features liquid cooling and an innovative technology for monitoring and capping the power consumption.
D.A.V.I.D.E represents the first deployment of a novel architecture developed within the European PRACE Pre-Commercial Procurement (PCP) project. PRACE pooled together a group of European Procurers (CINECA, GENCI, JUELICH, CSC and EPCC) who tasked the PCP project to develop a Whole-System Design for Energy Efficient HPC. D.A.V.I.D.E., jointly owned by the Procurers, is currently installed at CINECA undergoing acceptance tests and is available to the entire scientific community.
D.A.V.I.D.E. is composed of 45 nodes connected via Infiniband and is based on OpenPOWER servers. Each node is a 2U Open Rack form factor chassis and hosts two IBM POWER8 Processors with NVIDIA NVLink and four NVIDIA Tesla P100 data center GPUs. Each node, named OP 206EC, where “EC” stands for “Enhanced Capabilities”, presents two major innovations: liquid Cooling, co-designed with Asetek, and an innovative technology for monitoring and capping the power consumption, co-developed with University of Bologna.
Liquid cooling is key to increase both the repeatability of the performance and the overall energy-efficiency of the system; moreover, it ensures a safe operating temperature in the silicon die, extending the lifespan of the components by lowering the thermal stresses due to higher temperature.
The second key innovation is the co-development of a new technology for measuring, monitoring and capping the power consumption of the whole system and of the single node, through the collection of data from all relevant components (processors, memory, GPU, fans) and taking actions to improve the energy efficiency without impact the overall performance and throughput.
“It has been a pleasure to work with E4 and its R&D team and we are proud to have contributed to the co-design of this innovative cluster” said John Hamill, Asetek Chief Operating Officer. “We are also pleased to be part of this European project, where it has been demonstrated that energy-efficiency and performance can perfectly coexist”.
“D.A.V.I.D.E. combines dedicated hardware, artificial intelligence, big-data technologies and smart scheduling algorithms to control and assess the system’s power consumption with no impact on performance and usage model” said Andrea Bartolini, Assistant Professor, Università di Bologna. “We are very proud to be part of the project team with E4 and its partners”.
“Co-design and actively participate to the technological innovation processes in Europe is part of our institutional mission”, said Sanzio Bassini, Director of Department of Supercomputing Applications and Innovation at CINECA. “We are ready to exploit the technological potential of this new system, which represents the latest addition to our datacenter, in order to optimize the performance of algorithms beyond HPC, in fast growing areas such as AI, machine learning and HPDA”
“We believe that D.A.V.I.D.E. demonstrated that it is possible to design cost effective technologies achieving high performance with significant improvements on energy efficiency. We thank our technology partners and CINECA for the tight collaboration that contributed to this great achievement and we look forward to further enhancing the features of the system to better support the entire scientific community”, said Cosimo Gianfreda, E4 Computer Engineering Chief Technology Officer.
About E4 Computer Engineering
Since 2002, E4 Computer Engineering has been innovating and actively encouraging the adoption of new computing and storage technologies. Because new ideas are so important, we invest heavily in research and hence in our future. Thanks to our comprehensive range of hardware, software and services, we are able to offer our customers complete solutions for their most demanding workloads on: HPC, Big-Data, AI, Deep Learning, Data Analytics, Cognitive Computing and for any challenging Storage and Computing requirements. E4. When Performance Matters.
Source: E4 Computer Engineering