Super challenge, super winners
Third annual Rosie Supercomputer Super Challenge competition sees highest level of student participation
Forty-seven students competed and showcased their projects using the power of Rosie at the third annual Rosie Supercomputer Super Challenge. The record level of participants came with amazing projects all vying for the Super Challenge prizes that totaled $10,000 in cash, 10 RTX 4090s and 10 NVIDIA Jetson Orin Nanos. The competition asks students to solve an interesting problem, answer a difficult question, go beyond their course work or improve an existing process using Rosie, MSOE’s Supercomputer.
Five teams were chosen as finalists to present in front of judges and an audience. The judging panel included two MSOE Regents: Dr. Dwight Diercks ’90, NVIDIA senior vice president of software engineering and Nick Haemel ’02, NVIDIA vice president of medical imaging & system software, Dr. Derek Riley, MSOE computer science program director. Projects featured a wide range of ideas and applications using artificial intelligence technology.
· First place ($5,000 + 5 RTX 4090s): Adapting Large Language Models (LLMs) To Domains Using Retrieval Augmented Generation (RAG). Nathan Cernik, Tyler Cernik, Jennifer Madigan, Kevin Paganini and Jackson Rolando.
· Second place ($3,000 + 2 RTX 4090s): Mind over (Gray) Matter: Homologous Point Transform Applications to Brain Histology and MR Domains. Abagail Draper, Bart Gebka, Sonia Grade, Caleb Gray, Alex Neher and Mikolaj Sordyl
· Third place ($2,000 + 2 RTX 4090s): NourishNet: Proactive Severity State Forecasting of Food Commodity Prices for Global Warning Systems. Sydney Balboni, Ella Bruce, John Cisler, Caitlyn Grant, Grace Ivey, Ben Paulson, Tyge Plater and Brett Storoe.
The fourth and fifth place finishing teams were awarded with NVIDIA Jetson Orin Nanos. Additionally, a poster session was held to showcase the other projects that were submitted but not included in the top five. Two certificates for Most Novel Technical Implementation and Most Socially Beneficial were awarded after the poster session, and one lucky audience member's name was drawn to win an RTX 4090!
2024 Rosie Supercomputer Super Challenge
First place: Adapting Large Language Models (LLMs) To Domains Using Retrieval Augmented Generation (RAG). Nathan Cernik, Tyler Cernik, Jennifer Madigan, Kevin Paganini and Jackson Rolando.
Second place: Mind over (Gray) Matter: Homologous Point Transform Applications to Brain Histology and MR Domains. Abagail Draper, Bart Gebka, Sonia Grade, Caleb Gray, Alex Neher and Mikolaj Sordyl
Third place: NourishNet: Proactive Severity State Forecasting of Food Commodity Prices for Global Warning Systems. Sydney Balboni, Ella Bruce, John Cisler, Caitlyn Grant, Grace Ivey, Ben Paulson, Tyge Plater and Brett Storoe.
Dr. Dwight Diercks and Dr. John Walz
About Rosie
Rosie is the supercomputer at Milwaukee School of Engineering, and it is available to undergraduates, graduates, faculty, staff and the university’s business partners. Rosie includes three NVIDIA DGX-1 pods, each with eight NVIDIA V100 Tensor Core GPUs, and 20 servers each with four NVIDIA T4 GPUs. The nodes are joined together by Mellanox networking fabric and share 200TB of network-attached storage.
In 2024, Rosie received an upgrade that includes two DGXH100s, 16 H100 GPUs with 1.2TB of GPU memory, 224 CPU cores, and 800 GB of Infiniband networking. MSOE is the first in Wisconsin to deploy this supercomputing hardware. The DGX H100 systems deliver the scale demanded to meet the massive compute requirements of generative A.I. including large language models, recommender systems, health care research and climate science. Packing eight NVIDIA H100 GPUs per system, connected as one by NVIDIA NVLink®, each DGX H100 provides 32 petaflops of A.I. performance.