The lines between traditional supercomputers, or high-performance computing (HPC) systems, running complex science and engineering applications have blurred with the advent of cloud computing, AI and data analytics. HPC has spread its wings and is finding major applications in enterprise data centers. Transaction processing, data mining and software development have become additional use cases alongside traditional HPC applications such as weather simulations and computational fluid dynamics.
At SC19 this week in Denver, we will be showcasing how FPGA-based acceleration is driving important advances in AI and HPC. Specifically, we will offer demonstrations of our real-time computing platform, development tools and ecosystem partner solutions including AWS, Microsoft Azure, IBM, Samsung, and many more. Additionally, we are participating in multiple speaker sessions, including a panel discussion titled, “Within Reach: The Last Mile to Exascale” with our CTO, Ivo Bolsens.
Here’s the summary of Xilinx speakers and booth demonstrations. Come see us in booth #301.
Title: Within Reach: The Last Mile to Exascale
Tuesday, November 19 at 3:30 pm – 5 pm | Room 201-203 Presenter: Ivo Bolsens, senior vice president and chief technology officer, Xilinx
Today, the supercomputing industry is in a global race to exascale computing. Each brigade is vying for the title of ‘first’ or ‘best’ in exascale. This inflection point for computing is creating a new wave of competition within the industry – and internationally. That competition is spurring a wave of innovation. The question is no longer ‘if’ but when and, more importantly, shifts to what will be possible in this next great computational era. Regardless of who first breaks the tape, the new technologies and exotic engineering approaches are poised to provide immense industry benefits. The panel will discuss the unique innovations in the coming exascale systems based on an open platform approach, including new technical breakthroughs, software and programming tools, and the practical applications for HPC scenarios.
“Different” doesn’t mean “Difficult”: FPGA Programming Demystified
FPGAs are a powerful resource to drive memory-intensive simulations. The programming model is different than other devices, but that doesn’t mean it has to be difficult. The heterogeneous computing compiler from Xilinx – Vitis – in combination with a series of optimized FPGA compute libraries makes FPGA programming more approachable than ever before. Learn about the tool and see a demo of a compute and memory intensive binomial options pricing model (BOPM) in action.
The Intersection of HPC, Quantitative Finance, and FPGAs on Azure Thursday, November 21 at 10:30 am | Microsoft Azure booth #633 Presenter: Rajiv Jain, Financial Segment Marketing Manager, Xilinx
Learn how the financial services industry can leverage HPC tools running on Xilinx-enabled Azure FPGA VMs to power their pricing and risk management. We'll outline how the Xilinx Quantitative Finance Library provides accelerated functions and pre-built pricing models to allow quants and developers to quickly build and deploy accelerated solutions with the efficiency of the Azure platform. These libraries and tools represent the future of customized software developer tools for hardware acceleration.
Vitis Unified Software Platform
Demonstration of the Vitis Unified Software Platform workflow. Learn how to quickly develop, analyze, and deploy accelerated applications from cloud to edge.
Samsung SmartSSD See the future of computational storage as it performs high-speed computations on the data where it is stored. Combines fast Samsung SSDs and Xilinx FPGA horsepower for highly efficient parallel computation, secure storage, retrieval, processing, and analysis.