Research

Coordinates hardware and software views of entry/exit cost when selecting processor low-power states, avoiding energy-negative sleep transitions. Granted November 12, 2024.

Augments OS thread scheduling decisions with temporal characteristics of workloads, improving placement of threads on heterogeneous cores.

GNNs bottleneck on non-contiguous memory access patterns that GPUs handle poorly; this work evaluates how dataflow architectures from emerging AI accelerators alleviate those memory and compute bottlenecks for graph neural network workloads.

Evaluates dataflow-based AI accelerators from SambaNova, Cerebras, Graphcore, and Groq against diverse workloads - deep learning primitives, benchmark models, and scientific machine learning applications - characterizing where novel architectures beat GPUs and why.

How to schedule inference workloads across the heterogeneous accelerators (CPU, GPU, VPU) on a single edge device: a platform-aware allocation approach that improves throughput and energy efficiency over naive placement.

A semi-autonomous robot that identifies and sorts recyclable waste - early work combining computer vision with embedded robotics.