Case Study: NASA Langley Research Center achieves 40× faster acoustic analysis with MathWorks MATLAB and GPU computing

NASA Langley Research Center Accelerates Acoustic Data Analysis with GPU Computing

NASA Langley Research Center needed to speed up analysis of acoustic data from wind tunnel tests of novel aircraft designs, where up to 126 microphones produced large datasets and legacy Fortran code took 20–40 minutes to process each recording. To enable in‑facility, between‑run analysis and quicker identification of hardware or setup problems, the team turned to MathWorks tools, implementing the workflow with MATLAB along with Parallel Computing Toolbox (and Signal Processing Toolbox for windowing and FFTs).

Engineers reimplemented the processing algorithm in MATLAB and used MathWorks’ Parallel Computing Toolbox to run FFTs and matrix operations on an NVIDIA K20 GPU, enabling GPU acceleration with minimal code changes. The change delivered a 40× speedup—processing time dropped to under a minute (with GPU enablement achieved in about 30 minutes)—allowing NASA to analyze data after each recording and fix issues immediately.

NASA

Christopher Bahr

Research Aerospace Engineer