Research

I am currently working on three projects related to compute and memory acceleration: 1) Updates to our OpenStack cluster that will support temperature- and power-based measurements for nodes that also include GPU accelerators. 2) An initial investigation of next-generation programming models and runtimes, such as OpenACC and OpenMP 4.0. with researchers from ORNL 3) Algorithm support for programmable compute devices like FPGAs and high-bandwidth memories. This last project is joint with Dr. Vuduc, and is related to underlying compute architecture and software being done by students from Dr. Yalamanchili and Dr. Hyesoon Kim's groups.

I recently finished my PhD proposal and am now working on the last part of my thesis, a project called Oncilla. This work is focused on using Global Address Spaces (GAS) to enable high-performance data movement between host memory (DRAM) and accelerator memory (GPUs). The target application is a data warehousing workload, TPC-H, that is the focus of GPU-related research with the Red Fox compiler framework. For more information please click the link above or check the sidebar.

I finished my Master's thesis in Fall 2008 and have a related paper and presentation posted here.

Please also feel free to check out my publications under the Publications page.