In this special guest feature, Rob Farber writes that a study done by Kyoto University Graduate School of Medicine shows that code modernization can help Intel Xeon processors outperform GPUs on machine learning code. “The Kyoto results demonstrate that modern multicore processing technology now matches or exceeds GPU machine-learning performance, but equivalently optimized software is required to perform a fair benchmark comparison. For historical reasons, many software packages like Theano lacked optimized multicore code as all the open source effort had been put into optimizing the GPU code paths.”
In this video from ISC 2016, Greg Schmidt from Hewlett Packard Enterprise describes the new Apollo 6500 server. With up to eight high performance NVIDIA GPUs designed for maximum transfer bandwidth, the HPE Apollo 6500 is purpose-built for HPC and deep learning applications. Its high ratio of GPUs to CPUs, dense 4U form factor and efficient design enable organizations to run deep learning recommendation algorithms faster and more efficiently, significantly reducing model training time and accelerating the delivery of real-time results, all while controlling costs.
Today One Stop Systems (OSS) announced that it has completed a merger with Magma, with OSS as the surviving entity. Both companies are market leaders in PCIe expansion technology used to create high-end compute accelerators and flash storage arrays. Together they become a dominant technology leader of PCIe expansion appliances.
In this video from PASC16, Peter Messmer from Nvidia gives his perspectives on the conference and his work on co-design for high performance computing. “Using a combination of specialized rather than one type fits all processing elements offers the advantage of providing the most economical hardware for each task in a complex application. In order to produce optimal codes for such heterogeneous systems, application developers will need to design algorithms with the architectural options in mind.”
Organizations that implement high-performance computing (HPC) technologies have a wide range of requirements. From small manufacturing suppliers to national research institutions, using significant computing technologies is critical to creating innovative products and leading-edge research. No two HPC installations are the same. For maximum return, budget, software requirements, performance and customization all must be considered before installing and operating a successful environment.
A new World Record was set by the Huazhong University team at the Student Cluster Competition at ISC 2016. Using Nvidia Tesla K80 GPUs, the team recorded 12.56 teraflops on the LINPACK benchmark, while staying within a 3-Kw power consumption limit.
In this video from ISC 2016, Marc Hamilton from Nvidia describes the new DGX-1 Deep Learning Supercomputer. “The NVIDIA DGX-1 is the world’s first purpose-built system for deep learning with fully integrated hardware and software that can be deployed quickly and easily. Its revolutionary performance significantly accelerates training time, making the NVIDIA DGX-1 the world’s first deep learning supercomputer in a box.”
At ISC 2016, Supermicro debuted the latest innovations in HPC architectures and technologies including a 2U 4-Node server supporting new Intel Xeon Phi processors (formerly code named Knights Landing) with integrated or external Intel Omni-Path fabric option, together with associated 4U/Tower development workstation; 1U SuperServer supporting up to 4 GPU including the next generation P100 GPU; Lustre High Performance File system; and 1U 48-port top-of-rack network switch with 100Gbps Intel Omni-Path Architecture (OPA) providing a unique HPC cluster solution offering excellent bandwidth, latency and message rate that is highly scalable and easily serviceable.
In this video from ISC 2016, Steve Branton from Asetek describes the company’s innovative liquid cooling solutions for HPC. “Because liquid is 4,000 times better at storing and transferring heat than air, Asetek’s solutions provide immediate and measurable benefits to large and small data centers alike. RackCDU D2C is a “free cooling” solution that captures between 60% and 80% of server heat, reducing data center cooling cost by over 50% and allowing 2.5x-5x increases in data center server density. D2C removes heat from CPUs, GPUs, memory modules within servers using water as hot as 40°C (105°F), eliminating the need for chilling to cool these components.”