Leveraging SmartNICs for HPC and Data Center Applications

Title: Leveraging SmartNICs for HPC and Data Center Applications

Speakers: Jeffrey Young (Georgia Institute of Technology), Rich Graham (NVIDIA), Oscar Hernandez (Oak Ridge National Laboratory), and Richard Vuduc (Georgia Institute of Technology)

Abstract: The past few years have witnessed a surge in the number of advanced network adapters, known as “SmartNICs”, that offer additional functionalities beyond standard packet processing capabilities. These devices often feature programmable lightweight processing cores, FPGAs, and even CPU- and GPUbased platforms capable of running separate operating systems. Though primarily aimed at data center operations, such as infrastructure management, packet filtering, and I/O acceleration, SmartNICs are increasingly being explored for high-performance computing (HPC) application acceleration. This half-day tutorial offers an in-depth exploration of the state-of-the-art for SmartNICs and the emerging software ecosystems supporting them. Attendees will engage in hands-on exercises to better understand how to use SmartNICs for HPC application acceleration, including MPI collective operation offloading, OpenMP remote offload, and algorithmic modifications to maximize on-board processing power. Participants will have the opportunity to execute these exercises using cutting-edge SmartNICs like NVIDIA’s BlueField-3 Data Processing Unit (DPU). The tutorial presenters will further discuss optimizing applications to harness SmartNICs as communication accelerators in HPC systems.

Bios:

Jeffrey Young is a senior research scientist in Georgia Tech’s School of Computer Science. With a background in computer architecture, his main research interests have focused on the intersection of high-performance computing and novel accelerators including GPUs, Xeon Phi, FPGAs, and Arm SVE processors. He is the director of a novel architecture testbed, the CRNCH Rogues Gallery, that aims to simplify and democratize access to novel post-Moore accelerators in the neuromorphic, reversible, and novel networking spaces. He received his PhD in computer engineering in 2013 from Georgia Tech’s ECE department.

Dr. Richard Graham is the Senior Director of HPC Technology at NVIDIA’s Networking Business unit. His main area of expertise revolves around HPC network software and hardware capabilities for present and upcoming HPC technologies. Before joining Mellanox/NVIDIA, Dr. Graham accumulated thirteen years of experience at Los Alamos National Laboratory and Oak Ridge National Laboratory, where he held technical and administrative positions in computer science. His technical focus encompassed communication libraries and application analysis tools. Additionally, Dr. Graham played a significant role as a co-founder of the Open MPI collaboration and served as the chairman of the MPI 3.0 standardization efforts.

Oscar Hernandez holds a PhD in Computer Science and currently works at Oak Ridge National Laboratory (ORNL). Dr. Hernandez conducts research on programming models, compilers, and tools deployed at supercomputers such as Summit and Frontier at the Leadership Computing Facility (OLCF). At ORNL, he has contributed to the standardization of parallel languages and APIs for accelerated nodes, including OpenACC/OpenMP, as well as communication libraries and frameworks like OpenSHMEM and UCX. Furthermore, he has been involved with the Exascale Computing Project, leading various initiatives to implement these technologies on Exascale systems. Oscar has also collaborated closely with application teams, including the CAAR, INCITE, and ALCC projects, as well as numerous projects funded by DOE, DoD, NSF, and Industrial Partners within the Oil & Gas industry. Additionally, he has extensive experience delivering tutorials at various events, including Supercomputing, ISC, Exascale Computing Annual Meeting, and for NSF.

Richard (Rich) Vuduc is an Associate Professor at the Georgia Institute of Technology in the School of Computational Science and Engineering, a department devoted to the study of computer-based modeling and simulation of natural and engineered systems. His research lab, The HPC Garage is interested in high-performance computing, with an emphasis on algorithms, performance analysis, and performance engineering. He is a recipient of a DARPA Computer Science Study Group grant; an NSF CAREER award; a collaborative Gordon Bell Prize in 2010; Lockheed-Martin Aeronautics Company Dean’s Award for Teaching Excellence (2013); and Best Paper Awards at the SIAM Conference on Data Mining (SDM, 2012) and the IEEE Parallel and Distributed Processing Symposium (IPDPS, 2015), among others. Most recently, Dr. Vuduc has led an effort to map high-performance applications like LAMMPS and MueLu to Data Processing Unit (DPUs) with early results resulting in an IPDPS 2022 Best Paper nominee.