About

VLSI Group Overview

We study processor design, reconfigurable-computing based accelerators for image processing and machine learning, low power circuits, high-speed links, computational photography, and applying engineering to biology.

Stanford has had a research effort focusing on VLSI design since the late 70’s, though the nature of the work has changed dramatically during the past two decades. The early work in this group focused on designing high performance functional units, and the tools and techniques to support this effort.  This work led to fast adder/multiplier/divider designs, interesting memory structures and tools. Scaling and the resulting increase in design complexity has changed the focus of the group from functional units to complex chip design. 

More recent work has focused on low-power design issues. Current issues include power, since all chips are power limited, and design cost. So we work to change design methods such that one can more easily create power efficient chips with low design NRE costs. We apply this research to multiple areas including high speed links, reconfigurable computing, and efficient hardware and software methods for image processing and machine learning.

In addition, we have explored the use of tools from digital systems design, computational photography and VLSI circuits to analyze biological systems. Past work has included modeling of the Caulobacter cell cycle (w/ McAdams and Shapiro), new micro-fluidic structures (w/ Quake), and light-field microscopy (w/ Levoy). More recent work has supported efforts to improve brain mapping techniques, and understanding motion and visual processing in fruit flies.