Parallel Implementation of Irregular Terrain Model on NVIDIA Graphic Processing Units

University of Arizona
posted on 02/06/2010

• request more info 
•   
•  LICENSE ONLINE

Background:  Irregular Terrain Models (ITM) predict long-term median transmission loss of a radio signal.  The model can be applied to various engineering problems such as preliminary esitmates for system design, cellular radio telephones, and conventional land mobile systems.  In a mobile environment, rapid changes in terrain profiles and and radio coverage demads a real time response which requires millions of calculations per second.

To improve on the performance of ITMs, researchers at The University of Arizona have researched the use of multicore processors, such as nVidia's graphics processing units (GPU's) and IBM Cell Broadband Engine (Cell BE) to achieve improved ITM performance over mainstream microprocessors.  By exploring the unique advantages these platforms present, our researchers were able to reduce the execution time of running ITM processes by a factor of 140x on the GPU platform compared to general purpose processors.  Additionally, our researchers demonstrated that by leveraging the GPU platform they were able to achive cost savings through lower power consumption.

Advantages:
*higher performance over other platforms
*lower power consumption*lower cost for performance

Applications:
*dynamic map generation for communications networks
*ability to rapidly identify unused broadcast/communications spectrum

Researchers: Dr. Ali Akoglu, Yang Song

Status:  Available for non-exclusive licensing

Refer to Case #UA10-036