31 August 2026 to 4 September 2026
US/Pacific timezone

Accelerating the Full Particle Transformer on the AMD Versal AI Engine

2 Sept 2026, 11:48
12m
QI Auditorium

QI Auditorium

Presentation Contributed Talks

Speaker

Yuheon Joh (University of California San Diego)

Description

While transformer-based models such as the Particle Transformer have achieved strong performance offline, deploying the full model in a low-latency, resource-constrained trigger system remains an open problem. We extend the previous implementation of a quantized, integer-only transformer for jet tagging on the AMD Versal AI Engine by adding support for LayerNorm and GELU and implementing the Particle Transformer on the AIE. The main contribution is full support for all operations used in the Particle Transformer for the AMD AIE by adding the missing operations, GELU and LayerNorm, implementing the full model on the AIE, and analyzing and optimizing performance.

Authors

Yuheon Joh (University of California San Diego) Ryan Kastner

Presentation materials