Events 

Portal Annual Retreat 2025

Oceano Hotel, Half Moon Bay
Wednesday, August 27- Thursday, August 28, 2025

Please note: some talk recordings are not included due to pre-publication constraints

Wednesday, August 27

3:00pm Hotel check-in
4:00-4:15pm Welcome - Introduction
Mark Horowitz Faculty
4:15-4:30pm Research Overview
Fred Kjolstad Faculty
4:30-5:15pm Lightning Talks Portal Researchers and Students
5:15-6:15pm Poster Session & Reception (outside) Portal Researchers and Students
6:15-7:45pm Dinner All

Thursday, August 28

7:45am Breakfast
8:30-10:00am Session: High-Level Unified Collection IR
8:30-9:00am Introduction to the Unified Collection IR
Fred Kjolstad Faculty
9:00-9:30am Portable Compilation of the Collection IR
Olivia Hsu Postdoc, Incoming Professor CMU
9:30-10:00am Compiling Set Expressions onto Acceleration Trees
AJ Root PhD Student
10:00-10:30am Break and guest room check-out
10:30-11:30am Keynote: Compiler 2.0: Building Compilers in the Era of Machine Learning
Saman Amarasinghe Professor, Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
11:30am-12:00pm Session: Validation
11:30-12:00pm Verification and Synthesis of Spatial Implications for Bounding Volume Hierarchy Traversals
Clark Barrett Faculty
12:00-1:15pm Lunch All
1:15-2:30pm Session: Abstract Machines and Hardware Generators
1:15-1:40pm Voyager: An HLS-Based Framework for Design-Space Exploration and Generation of DNN Accelerators Kartik Prabhu PhD Student
1:40-2:05pm Zircon: Heterogeneous CGRA for Accelerating Machine Learning, Image Processing, and Sparse Tensor Algebra Applications Michael Oduoza PhD Student
2:05-2:30pm LEGO: A compiler for tiling and orchestrating sparse tensor algebra expressions Gautham Ravipati PhD Student
2:30-3:00pm Afternoon Break
3:00-4:15pm Session: Emerging Data Representations
3:00-3:25pm A 7nm SoC for LLM-based AI Reasoning featuring Variable Block Size Formatbook Quantization Christian Kubicka MS Student
3:25-3:50pm Vision-Language Alignment from Compressed Image Representation using 2D Gaussian Splatting Yasmine Omri PhD Student
3:50-4:15pm Exchangeability in Neural Network Architectures and its Application to Dynamic Pruning Luke Yi PhD Student
4:15-4:45pm Explicit Feedback Session
4:45-5:00pm Wrap up and closing thoughts Mark Horowitz Faculty