GPU Acceleration of Data Processing and Analysis
By its nature, radar data tends to be highly parallel (i.e., every pulse is processed identically). This holds true for traditional detection applications as well as more specialized imaging applications. Because of this data parallelism, radar data processing lends itself well to massively-parallel processors, such as Graphics Processing Units (GPU's).
To take advantage of GPU's for parallel processing, SPC’s Radar Physics group became an early adopter of Accelereyes’s Jacket platform. Jacket serves as middleware, which allows Matlab-based calculations to be performed on NVIDIA GPU's. SPC’s success with Jacket has led to recognition by Accelereyes as a “success story.”
While much of SPC’s experience with GPU acceleration has focused on radar processing (see below), the Matlab and Jacket expertise we have developed is applicable to a wide variety of fields.
To learn how SPC’s scientists can help accelerate your processing, analysis, or modeling Matlab code, please contact us at firstname.lastname@example.org.
Click to view larger image
Radar Clutter Reduction
SPC has used GPU's to accelerate processing to reduce radar clutter. The algorithm processes raw data from marine navigation radars using a variety of thresholding techniques to extract real targets from clutter. Using an NVIDIA GPU, SPC has achieved 10x speed improvements relative to a Core i7-920 CPU and 5x improvements relative to a real-time DSP implementation.
Radar Image Formation
Radar imaging is computationally intensive. As a result, many imaging algorithms apply FFT-based approximations. While efficient, these algorithms sacrifice data fidelity for speed. Other algorithms better preserve information, but are often too slow for many applications. SPC scientists have implemented a Matlab SAR/ISAR imaging routine based on the backprojection algorithm. Using an NVIDIA GPU, we have demonstrated speedups of roughly 45x for large datasets.
Download the GPU Acceleration of SAR/ISAR Imaging Algorithms
, presented at the 2010 Antenna Measurement Techniques Association (AMTA) Symposium.
To receive more details (brochures, technical specifications, etc.) fill out the form below. This form is for internal SPC business purposes only. No information will be provided to third parties.. After submitting the form, you will be contacted by SPC personnel.