Porting Signal processing algorithms to CuPy for precision measurement
summary
Describes porting interferometry-based signal processing used at CERN (Butterworth filter, Hilbert transform, Savitzky–Golay smoothing) from CPU to NVIDIA GPUs using CuPy, then later a C implementation. Explains the steps taken to adapt these algorithms to run in parallel on GPUs, measures performance as data size increases, and discusses optimizations to meet sub-second processing for thousands of target points.
AI-written summary · about
related talks
30mCuPy: A NumPy-compatible Library for GPU
30mCUDA in Your Python: Effective Parallel Programming on the GPU
35mPython at the Speed of Light: A Complete Tour of the CUDA Ecosystem
01h 02mExploit your GPU power with PyCUDA (and friends)
29mA Beginners Guide to GPUs for Pythonistas
41mLuxuries, necessities, and the challenges that remain
25mScale Smarter, Not Harder, with cuPyNumeric.
02h 28mGPU Accelerated Python