TPC Logo reads "Trillion Parameter Consortium. AI For Science"

Fall 2026 TPC Hackathon

Fall 2026 TPC Hackathon

Hosted by the Leibniz Supercomputing Centre (LRZ)
9-11 September 2026

Calls available to:

participate in the hackathon
lead a breakout group
Call Closed

Event Overview

The Trillion Parameter Consortium (TPC) seeks to foster open, collaborative research on the development and deployment of large and massive scale AI models for science. EuroTPC is Europe’s coordinated presence within the TPC, connecting Europe’s leading high-performance computing centres, AI Factories and research institutions to TPC’s international governance, working groups and scientific activities.

Through hackathons, the TPC community has catalyzed the acceleration of many individual and multi-institutional projects. LRZ with be hosting the Fall 2026 edition of the TPC hackathon. The hackathon will start on Wednesday, September 9, 2026 to Friday, September 11, 2026, and will take place in Garching, near Munich, Germany.

Over the course of the event, computational scientists and researchers will work hands-on to build a first prototype of their ideas, or port, accelerate, scale and optimize existing AI applications for a variety of CPU and GPU architectures. Each team will be matched with an experienced mentor — well-versed in both programming techniques and the relevant application domain — to help drive tangible progress and speedups through targeted use of programming models, libraries, and optimization tools.

Working Group Breakout Descriptions

  • WG1: Develop workflows and agentic systems, collaborate on various implementations for Pandemic Preparedness Engine (PPX), a project funded from the Coalition for Epidemic Preparedness Innovations (CEPI).
  • WG2: Benchmark llama.cpp backends and GPU vendors for performance and energy efficiency across chat, reasoning, and vision-language models to find the most power-efficient hardware/software/model combinations.
  • WG3: Cost-aware LLM evolutionary algorithm design (multi-fidelity + Bayesian optimization + tiered model routing), benchmarked on GPU kernel-generation tasks on LRZ HPC infrastructure.
  • WG4: LLM-agent system that profiles HPC hardware, runs targeted micro-benchmarks, and suggests informed code optimizations grounded in real measurements.

Registration

  • This Hackathon is open for international participants from academia and industry.
  • Registration does not guarantee participation. We will evaluate your application and inform you upon acceptance.
  • Please register with your official e-mail address to prove your affiliation.

If you’d like to participate as part of a team, you can do so through LRZ’s official platform:
Deadline: 24 August 2026

Register Here

Agenda

Wed, Sept. 9, 2026

  • 10:00-11:00 Welcome and Introduction of Teams (+ coffee)
  • 11:00-13:00 Breakout
  • 13:00-14:00 Lunch
  • 14:00-14:30 Updates on TPC
  • 14:30-16:00 Breakout
  • 16:00-16:15 Coffee Break
  • 16:15-18:00 Breakout
  • 18:00-18:30 Daily Progress Updates by each Team

Thu, Sept. 10, 2026

  • 08:30-10:30 Breakout
  • 10:30-11:00 Coffee Break
  • 11:00-12:45 Breakout
  • 12:45-13:45 Lunch
  • 14:00 onwards Social event (brewery tour + dinner)
  • 18:00-18:30 Daily Progress Updates by each Team

Fri, Sept. 11, 2026

  • 08:30-10:30 Breakout
  • 10:30-11:00 Coffee Break
  • 11:00-13:00 Breakout
  • 13:00-14:00 Lunch
  • 14:00-16:00 Final updates
  • 16:00-16:30 Closing & Adjourn

Program Committee

  • Ajay Navilarekal (Leibniz Supercomputing Centre, Germany)
  • Miguel Vázquez (Barcelona Supercomputing Center, Spain)
  • Arvind Ramanathan (Argonne National Laboratory, USA)