Over at the Nvidia Blog, George Millington writes that, the fourth consecutive year, the Nvidia Tesla Accelerated Computing Platform helped set new milestones in the Asia Student Supercomputer Challenge, the world’s largest supercomputer competition.
The NVIDIA DGX-1 features up to 170 teraflops of half precision (FP16) peak performance, 8 Tesla P100 GPU accelerators with 16GB of memory per GPU, 7TB SSD DL Cache, and a NVLink Hybrid Cube Mesh. Packaged with fully integrated hardware and easily deployed software, it is the world’s first system built specifically for deep learning and with NVIDIA’s revolutionary, Pascal-powered Tesla P100 accelerators, interconnected with NVIDIA’s NVLink. NVIDIA designed the DGX-1 to meet the never-ending computing demands of artificial intelligence and claims it can provide the throughput of 250 CPU-based servers delivered via a single box.
In this podcast, the Radio Free HPC team recaps the GPU Technology Conference, which wrapped up last week in San Jose.
Since Rich is traveling around in some desert somewhere, Dan and Henry go it alone and discuss the new Pascal (P1000) GPU, NVIDIA’s new server, and what happened at the concurrent OpenPOWER conference.”
“Cavium ThunderX has significant differentiation in the 64-bit ARM market as Cavium is the first ARMv8 vendor to deliver dual socket support with full ARMv8.1 implementation and significant advantage in CPU cores with 48 cores per socket. In addition, ThunderX supports large memory capacity (512GB per socket, 1TB in a 2S system) with excellent memory bandwidth and low memory latency. In addition, ThunderX includes multiple 10 GbE / 40GbE network interfaces delivering excellent IO throughput. These features enable ThunderX to deliver the core performance & scale out capability that the HPC market requires.”
In this video from the 2016 GPU Technology Conference, Jason Pai from Supermicro describes the new 1028GQ-TRT SuperServer. With support for up to four Nvidia Tesla K80 GPUs, the 1U superserver offers extreme compute density in 1U of rack space. “From HPC to Deep Learning and Big Data Analytics, denser, more powerful GPU solutions have become a necessity in order to service the next generation of GPU-accelerated applications. At GTC, Supermicro demonstrated how these applications have progressed, and how its GPU solutions are influencing this evolution.”
According to the latest Intersect360 Research site census data, of the 50 most popular application packages mentioned by HPC users, 34 have offer GPU support, including 9 of the top 10. As is evident from the number of GPU-accelerated applications available in areas such as chemical research, physics, structural analysis, and visualization, the use of this accelerator technology has become well established in the HPC user community. “According to the latest Intersect360 Research site census data, of the 50 most popular application packages mentioned by HPC users, 34 have offer GPU support, including 9 of the top 10. As is evident from the number of GPU-accelerated applications available in areas such as chemical research, physics, structural analysis, and visualization, the use of this accelerator technology has become well established in the HPC user community.”
In this video from the 2016 GPU Technology Conference, Jaan Mannik from One Stop Systems describes the GPUltima system. Delivering up to 1 Petaflop in a rack, the GPUltima is a single 19″ rack comprised of 8 OSS High Density Compute Accelerators (HDCA) each with 16 NIVIDA Dual GPUs (128 total), 16 dual-socket servers, an Infiniband Switch and an Ethernet Switch.”
Europe’s most powerful supercomputer Piz Daint is being upgraded, a move that is expected to at least double its computing power. ETH Zurich is investing around CHF 40 million to allow researchers to perform simulations, data analyses and visualizations even more efficiently in the future. Although slightly reduced in physical size, Piz Daint will become considerably more powerful as a result of the upgrade, particularly because we will be able to significantly increase bandwidth in the most important areas,” says CSCS Director Thomas Schulthess.
“Supermicro’s new SuperServer and SuperBlade platforms optimized for next generation GPU technology offer the most flexible architecture for the future of extreme parallel computing,” said Charles Liang, President and CEO of Supermicro. “Our proprietary cooling architecture supports up to four GPUs in 1U, maximizing compute density, performance and reliability under peak thermal operating conditions. Supermicro’s innovation and revolutionary solutions deliver the greatest range of options for next generation supercomputers, enabling new breakthroughs in the latest deep learning and artificial intelligence applications.”
As high performance and webscale applications become mainstream, HPE’s continued focus on this market is yielding positive results for our customers,” said Bill Mannel, vice president and general manager, HPC, Big Data and IoT Servers, HPE. “Already, more than a third of the HPC market is using HPE compute platforms to enhance scientific and business innovation and gain a competitive edge. Today’s announcement reinforces our commitment to delivering new infrastructure solutions that satisfy our customers’ insatiable need for massive compute power to fuel new applications and unlock the value of their data.”