EEG MEMD performance benchmark single GPU

This data set contains execution time benchmark results for the GPU implementation of the Multivariate Empirical Mode Decomposition algorithm, which is a novel method to perform time-frequency decomposition of non-stationary multi-channel signals. The benchmarks used single GPUs (V100 and other GPUs) and varied the number of electrodes, the length of the measurement data and the number of the used direction vectors. The benchmark programs were written in CUDA 10.2. In addition to the total execution time, the speedup is also calculated, showing the performance improvement compared to a CPU-based MATLAB implementation.

Data and Resources

Additional Info

Field Value
Last Updated April 4, 2023, 10:29 (UTC)
Created April 4, 2023, 10:29 (UTC)