Accelerating signal processing algorithms with GPUs and MATLAB
Several MATLAB® toolboxes for signal processing and communications contain highly optimized GPU functions that run on NVIDIA GPUs to reduce computation time. Although execution speed varies by application, users have achieved speedups of 30x for wireless communication system simulations.
Signal processing and communications algorithms contain structurally parallel data flows that involve iterative, computationally intensive, and time-consuming mathematical operations. NVIDIA GPUs contain thousands of highly specialized cores that operate in parallel to reduce execution time of these algorithms and accelerate simulation.
Existing signal processing and communications algorithms can run on NVIDIA GPUs with minimal code changes.
|Product Name||GPU Support|
|Signal Processing Toolbox™||1D Cross Correlation
2D Cross Correlation
|Phased Array System Toolbox™||
Clutter simulation on GPU in an end-to-end airborne radar system
|Additional Products||All of products with GPU and parallel computing support|
Examples and How To
- Accelerate Correlation with GPUs - Example
- Using GPUs To Accelerate Turbo Coding Bit Error Rate Simulations for Communications Systems - Example
- GPU Acceleration of Clutter Simulation for Radar Systems - Example
- Acceleration of Clutter Simulation Using GPU and Code Generation – Example
- GPUs for Signal Processing and Communications Algorithms - NVIDIA Article
- Signal Processing and Communications - MathWorks Consulting
- Compare GPUs Using Standard Numerical Benchmarks in MATLAB - File Exchange Download
- GPU-Enabled MATLAB Functions - Documentation
See also: MATLAB GPU computing, low-pass filter, high-pass filter, research with MATLAB