Hybrid Memory Project and OpenCAPI - Michael Miller, Rambus
Super-Connecting the World’s Number One Supercomputers - Scot Schultz, Mellanox Technologies
Mellanox continues leadership as the best interconnect worldwide for both HPC and AI. With the latest intelligent interconnect devices, such as Mellanox Quantum HDR 200Gb/s and ConnectX-6 are enabling the next generation of in-network co-processing. With a more effective mapping of communication between devices in the system dramatically increases system performance.
Keynote: Michael Cunningham, Executive Vice President, Red Hat
Architecture Effects When Interacting with GPU Technologies - Christopher M. Sullivan, Oregon State University
Oregon State University’s Center for Genome Research and Biocomputing (CGRB) works closely with its users to determine the best computing platform for computational research. In the past several years we have seen many differences associated with architecture and GPU interactions. As we continued our testing we found tremendous benefits from using the OpenPOWER platform with Power9 processors and NVIDIA GPU. In an effort to truly determine the best platform for different types of GPU processing we created a set of test. We will be showing all the current data comparing some of the most bleeding-edge platforms and a demo of the machines racing each other. Our goal is to provide users with the ability to bring online the best platform for the work they are planning with GPU technology.
OpenPower Solution Builder Community - John Pace, Mark III & Christopher Sullivan, Oregon State University
The OpenPower Solution Builder Community is for those who design, implement, operate and maintain Power and OpenPower solution stacks. The Group is self-governing and provides a means for builders to share insights and innovation. There are three primary elements 1) questions & answers and pointers to relevant information, 2) best practices on building hardware-software solution stacks and reference architectures, 3) innovations of applications, libraries, methods, and approaches. The initial focus is solution stacks around the IBM POWER9 AC922 integrated GPU Server but will expand to additional systems and workloads. You do not need to be an OpenPower member to participate. This talk will discuss the community formation, topics that are active, and will solicit participation.
Developing Applications using Acceleration Hardware on Power - Gerrit Huizenga, IBM
This session will share the goals of the newly chartered Accelerated Application Ecosystem Working Group, which will help identify resources needed when porting accelerated applications to Power specific hardware accelerators. This session will provide an overview of some of the available accelerators, some of the existing porting resources, and solicit information from ISVs and software developers on any gaps in the software ecosystem today.
Keynote: Aaron Sullivan, Director of Hardware Engineering, Facebook
Protected Execution Facility On Power - Guerney Hunt, Ram Pai & Michael Anderson, IBM
IBM POWER9 architecture introduces Protected Execution Facility (PEF) which enables secure memory, which can only be accessed when the processor is in a new higher privileged mode called Ultravisor mode. Some facilities that were previously hypervisor privileged are now ultravisor privileged. This talk presents architectural changes for PEF, the Protected Execution Ultravisor, and new services called ultra calls. Combined with enhancements to the Virtual Machine format a new class of VMs called Secure Virtual Machines (SVM) is enabled. SVMs utilize hypervisor services and run in secure memory under the control of the Ultravisor. SVMs are protected from attacks from the Hypervisor, other privileged software entities, and privileged users on the system. POWER9 systems can concurrently support both normal VMs and SVMs. All firmware and tools for PEF will be Open Sourced.