AHA! Agile Hardware Center.

Hardware/software systems must become easier and more fun to develop. We aim to enable a more agile hardware development flow, to quickly and easily modify an existing design, letting us play with the resulting system. Subprojects include:
     Frankencamera 4
     Halide for Heterogeneous Computing

CNN Blocking

Better memory behavior for Convolutional Neural Nets.

Digital Analog Design

Mixed-signal design flow

FPU Generator (FPGen)

A state-of-the-art FPU generator

Frankencamera 4

Building a heterogeneous platform for computational photography

Half Page DRAM and Charge Recycling Refresh

Reducing DRAM energy waste.

Halide for Heterogeneous Computing

Compiling Halide applications onto CPU/FPGA heterogeneous platforms.

Software Actuated Genetic Engineering (SAGE)

A CMOS lab-on-a-chip for genetic engineering, with various sensing and actuation capabilities

Stanford Extremely Efficient Computing (SEEC)

Custom ASICs can be three orders of magnitude more energy efficient than conventional CPUs, but have often-prohibitive design costs. Our goal is to create new design methods and tools to allow inexpensive development of customized chips.

Vision Processing Architecture Studies

What is the optimal architecture for vision processing?