Portrait of Ali Bakhoda

Ali Bakhoda

Principal Performance Architect

About

Currently I am working on performance modeling of neural network accelerators at Azure Hardware Systems & Infrastructure group.

After joining Microsoft Research I was mainly focused on designing and developing a functional and cycle-level simulator for evaluating Explicit Data Graph Execution (EDGE) architectures. This simulator provided a flexible framework for research into next-generation EDGE architectures. The functional full-system model was capable of booting Windows with a simple simulated LCD with mouse and keyboard working.

I graduated from UBC in May 2014. My PhD research was focused on developing throughput-effective Network-on-Chips for compute accelerators such as modern GPUs. I am also one of the original developers of GPGPU-Sim. GPGPU-Sim provides a detailed simulation model of a contemporary GPU running CUDA workloads.