Utilizing GPUs for benchmarking an autotuning system

Spring 2011

Student
Rachita Chandra
Advisor
Franz Franchetti
Project description

The goal of this project is to utilize NVIDIA GPUs (Graphics Processing Unit) and AMD APUs (Accelerated Processing Unit) in order to investigate the efficiency of an autotuning system for a random walk boolean satisfiability solver.

The efficiency of the code generated by the autotuning system will be gauged by measuring the generated code on both the APU and GPU.

Benchmarking in the research will focus on 2 areas: 1. Comparing GPU performance with APU performance 2. Comparing GPU performance with Intel multi-core processor performance.

Return to project list