X-ScaleSolutions

Are you attending the 2025 Hot Interconnects conference during August 20-22, 2025? The conference is being held in a virtual manner and the registration is free. Very happy to indicate that several members of my team and I will be participating in this conference to present a technical paper on "Characterizing Communication Patterns in Distributed Large Language Model Inference" and two tutorials (1. "Principles and Practice of Scalable and Distributed Deep Neural Networks Training and Inference" and 2. "High-Performance and Smart Networking Technologies for HPC and AI"). The detailed program is available from https://lnkd.in/gZZ9UTRD. Looking forward to seeing many of you at the conference in a virtual manner and interacting with you!! #hpc #ai #bigdata #interconnect #deeplearning #inference Hot Interconnects Lang X. Kaushik Kandadi Suresh Quentin Anthony Nawras Alnaasan Benjamin Michalowicz Dhabaleswar K. (DK) Panda

Interested in learning the latest and greatest in Message Passing Interface (MPI) and Programming Environment implementations with the latest CPU, GPU, and Interconnect technologies, and their impacts on #hpc #ai #deeplearning #machinelearning #bigdata #cloud and #quantum? The 13th annual MVAPICH User group (MUG) conference with an exciting program will be held in a hybrid manner from August 18-20, 2025, in Columbus, OH, USA. All interested parties are welcome to attend the event. Online attendance is free. However, one must register for the conference in advance to get the WebEx link. The registration link is available from the following URL: https://lnkd.in/d4qhR9aP Highlights of the conference include: - Presentations on many topics related to HPC and AI: High-Performance MPI Implementations over various networks and accelerators, Communication Offload, In-Network Computing, HPC network design in finance, Performance tuning for HPC and AI applications, Applications-level optimizations for HPC and AI on many platforms, MPI-based communication for Deep Learning, Energy-efficient MPI, Optimization of collective communication over CXL, Ultra Ethernet, High-Throughput and Scale-up Ethernet, CXL-based Memory Pooling, High-Performance Inference, Quantum Accelerated HPC, and Overview of upcoming products for accelerating HPC and AI from many vendors - Two keynote talks: Dan Stanzione from Texas Advanced Computing Center (TACC), USA and Manoj Wadekar, Meta. - Eight tutorials and demos from Broadcom, Enfabrica, Keysight Technologies, NVIDIA, ParaTools, Ltd., The Ohio State University, University of Oregon, and X-ScaleSolutions - 19 invited/contributed talks from many organizations (AMD, Broadcom, Center For Development Of Advanced Computing (C Dac) (India), Enfabrica, ETRI (South Korea), Hewlett Packard Enterprise, Jump Trading Group, KAUST (King Abdullah University of Science and Technology) (Saudi Arabia), Keysight Technologies, Konkuk University (South Korea), National Science Foundation (NSF), NVIDIA, ParaTools, Ltd., San Diego Supercomputer Center, The Ohio State University, Technical University of Munich (Germany), University of Tennessee, Knoxville, University of Oregon, and X-ScaleSolutions) - 12 poster presentations (Florida State University, Northeastern University, San Diego State University, The Ohio State University, UC San Diego, University of Maryland, The University of Texas at San Antonio, and University of Oregon) - 9 short presentations from the MVAPICH and HiDL project members - A talk on the Future Roadmap of the MVAPICH project for HPC and AI The detailed conference program is available from https://lnkd.in/g939PXZN Thanks to all sponsors for the MUG '25 conference – Broadcom, Enfabrica, Keysight, NSF, NVIDIA, Ohio Supercomputer Center, The Ohio State University, Paratools, and X-ScaleSolutions!! Join us for the conference!! The MVAPICH Team