High Performance Fabrics for AI and HPC
AI models continue to grow to trillions of parameters and exascale HPC simulations are becoming a reality. As computing demands surge for AI and HPC applications, networking becomes essential for performance and scaling. This workshop focuses on bringing together industry experts and researchers to address networking challenges in building high-performance AI and HPC fabrics.
General Chairs
| Name | Title |
|---|---|
| Dr. Hemal Shah | Distinguished Engineer and Architect, Broadcom Inc. |
| Manoj Wadekar | AI Systems Technologist, Meta |
Organizers
| Name | Title |
|---|---|
| Dr. Hemal Shah | Distinguished Engineer and Architect, Broadcom Inc. |
| Dr. Mohan Kalkunte | VP of Architecture, Broadcom Inc. |
| Dr. Dhabaleswar K (DK) Panda | Professor and University Distinguished Scholar of Computer Science and Engineering, Ohio State University |
| Manoj Wadekar | AI Systems Technologist, Meta |
Workshop Goals
The workshop will bring together perspectives from academic researchers, hyperscale operators, and industry ecosystem partners. The following are the key goals of this workshop:
- Explore novel research ideas around AI/HPC fabrics and related areas
- Serve as a key venue for sharing research ideas, experiences, and best practices in building high performance AI/HPC fabrics
- Offer collaboration opportunities to academic researchers, industry technologists, and students to work together to address challenges in designing large scale high-performance fabric solutions for AI/HPC workloads
- Provide a venue for participants to present early ideas and obtain feedback
- Provide attendees with a holistic understanding of Ethernet’s technical maturity, open challenges, and opportunities for research, standardization, and community collaboration
Relevance and Impact to SC Attendees:
The SC Conference is a premier venue for discussing AI/HPC technologies, and with the increasing scale of networking into HPC/AI environments, this workshop is highly relevant. SC attendees, including academic researchers, students, and industry professionals, will benefit from:
- Collaboration opportunities with leaders in high-performance AI/HPC networking
- Insights into cutting-edge networking infrastructure technologies for AI/HPC workloads
- Opportunities to engage with researchers in the field and learn new ideas
- A better understanding of deployment challenges with large-scale AI/HPC networks
Topics for the Workshop:
- AI/HPC fabric architecture
- Telemetry and diagnostics for AI/HPC fabrics
- Systems software for AI/HPC fabrics
- AI/HPC application networking patterns
- Switching and routing technologies for AI/HPC fabrics
- Congestion control in AI/HPC Networks
- Scale-out, Scale-Up, and Scale-across Ethernet Solutions for AI/HPC
- Ethernet transport enhancements
- Ethernet Fabric load balancing techniques
- Protocol and interfaces for AI/ML communications
- High-performance host-network interface architectures
- Software for monitoring, controlling, and managing Ethernet networks
Paper Submission:
Please submit papers covering original work, work-in-progress, or position papers. Submissions are limited to 10 two-column pages (U.S. letter – 8.5 x 11), excluding the bibliography, using the IEEE proceedings template. The IEEE conference proceeding templates for LaTeX and MS Word provided by IEEE eXpress Conference Publishing are available for download. See the templates at https://www.ieee.org/conferences/publishing/templates. Papers should be submitted via Linklings at https://submissions.supercomputing.org/. Please choose this workshop under the SC26 Workshops header.
Timeline
| Name | Date |
| Call for Papers | May 15, 2026 |
| Abstract submission due | August 8, 2026 |
| Paper submissions due | August 15, 2026 |
| Notification to authors of acceptance | September 04, 2026 |
| Camera ready papers due | September 25, 2026 |
| Workshop takes place | November 2026 |
Program Committee
| Name | Title |
|---|---|
| Daniele De Sensi | Sapienza University of Rome, Italy |
| Albert Greenberg | Uber |
| Torsten Hoefler | ETH Zurich, Switzerland |
| Hugh Holbrook | Arista |
| Abdul Kabbani | Microsoft |
| Adnan Khaleel | Google Inc. |
| Balaji Prabhakar | Stanford University |
| Muhammad Shahbaz | University of Michigan |
| Sameer Shende | University of Oregon |
| Rip Sohan | AMD |
| Hari Subramoni | The Ohio State University |
| Rajeev Thakur | Argonne National Lab |
| Brian Towles | OpenAI |
| Ron Brightwel | Sandia National Laboratories |
For Questions, Email Us At:
sc26-wksp137-cfp.pdl@broadcom.com
To submit papers, use SC26 submission site at: https://submissions.supercomputing.org/
