While the first POWER8 based server of the E4's series, the OP205 was intended to be a development platform, the OP206 solution is fully compatible in Linux-based Cloud environments and is optimized for data-rich applications delivering superior data centre efficiency.
The OP206 deploys up to four NVIDIA Tesla P100 GPU and the dual socket POWER8 with NVIDIA NVLink processor - the first high-speed GPU interconnect that let data move between GPUs and CPUs 5 to 12 times faster than they can today.
GPU computing developers' and users' fundamental pain points are keeping massively parallel GPUs fed with data and this platform aims to help them do just that.
A large number of applications run on the OP206 platform. A sampling of HPC applications suited for OP206 includes Code-Saturne, LS-DYNA, OpenFOAM, GAMESS, CPMD, VASP and many others.
"Our commitment to OpenPOWER Foundation is very strong so we wanted to bring an OCP solution that could be a reference for the Foundation's eco-system", stated Cosimo Gianfreda, CTO E4 Computer Engineering. "Having initially developed the OP206 solution for a highly customized project with excellent result, we are very confident that the standard version which is now available for order, will also bring outstanding results both in terms of speed and performances. The integration of the most advanced technologies such as POWER8 CPU with NVlink and the NVIDIA Tesla P100, means that data centre environments will be able to increase throughput while being cost effective."
"The POWER8 with NVIDIA NVLink processor enables incredible speed of data transfer between CPUs and GPUs ideal for emerging workloads like AI, machine learning and advanced analytics", stated Rick Newman, Director of OpenPOWER Strategy & Market Development Europe. "The open and collaborative spirit of innovation within the OpenPOWER Foundation enables companies like E4 to leverage new technology and build cutting edge solutions to help clients grappling with the massive amounts of data in today's technology environment."
"The Power System for High Performance Computing powered by NVIDIA Tesla P100 GPUs accelerators with NVLink will facilitate high-performance analytics and enable deep learning on ever increasing mountains of data", stated Stefan Kraemer, Head of the HPC team, NVIDIA. "Performance is a key word for advanced analytics, machine learning and artificial intelligence. OpenPOWER Partners like IBM and E4 are making real advancements in the productions of accelerated systems that will thrive in such environments."
Last year E4 Computer Engineering joined the OpenPOWER Foundation and since then it has been active in developing accelerated solutions that leverage the performance of the POWER8 processor, which was designed from the ground up for Big Data. This approach is totally in line with the company strategy, which consists in becoming the reference supplier for IT solutions, focused on technological innovation, with the aim to make customers develop and implement their own research and improve the world of tomorrow.