I just graduated from Stanford with a
Ph.D. in Computer Science. My adviser
was Mark Horowitz, whose
research group work on a wide
variety of projects. I first started working on the architecture design of
Smart Memories, a
reconfigurable chip multi-processor that supports different parallel execution
models. My focus was on the multi-threading support in hardware. Along the
way, I worked on a few simulators, a
transaction programming model and various parallel applications.
Eventually that led to my dissertation work on
characterizing a set of probabilistic inference algorithms (based on Bayesian
and Markov networks) in the contexts of their scalability on chip
multi-processors and their resilience against transient faults in the
circuit.