• Published a research paper titled “Power and Energy Footprint of OpenMP programs using OpenMP Runtime API” at the E2SC Workshop, held in conjunction with Supercomputing Conference, 2014 at New Orleans.
• Designed a mechanism to collect the Power and Energy values of the NASA parallel programming benchmarks by varying a comprehensive list of parameters and analyze them
• Wrote a shell script to automate the process of collection of metrics
• Worked with high performance systems running Linux to connect remotely via SSH to perform the work
• Interacted with a research fellow and a research assistant to understand the requirements of the project and to solve the problems encountered during the progress
• Visualized the collected data in the form of excel charts and later included them in the research paper
• Concluded that best execution performance of a program does not guarantee best energy performance
• Parallelized C and C++ programs using parallel programming language OpenMP to speed up execution
• Enhanced the performance evaluation of the project by adding new features to the Collector API program which is written in C.
• Wrote passes for the LLVM compiler in C++ for integrating OpenSHMEM framework with LLVM