Case Study: University College London achieves overnight completion of week-long wireless Monte Carlo simulations with MathWorks (MATLAB Parallel Server and Parallel Computing Toolbox)

University College London Accelerates Simulations of Next-Generation Wireless Communications Algorithms with Parallel Computing

University College London’s Institute of Communications and Connected Systems (ICE) needed to accelerate computationally intensive Monte Carlo simulations to evaluate next‑generation wireless waveforms (like SEFDM) and estimate bit error rates as low as 10^-9. Single‑core workstations made these experiments impractically slow, so the ICE group turned to MathWorks tools—MATLAB, Parallel Computing Toolbox, and MATLAB Parallel Server—to explore parallel execution on UCL’s Myriad cluster.

Using MathWorks’ toolchain, two students profiled and optimized the MATLAB code (vectorization, preallocation) and tested parallel patterns (parfor, parfeval) on a local workstation and then scaled to the multinode Myriad cluster with MATLAB Parallel Server. The result: simulations that once took six days now finish in about eight hours, the team can run billion‑bit experiments to produce more accurate BER curves, and researchers gain major productivity and deadline advantages.

University College London

Ryan Grammenos

University College London