Back to Volume
Paper: CUDA-Accelerated SVM for Celestial Object Classification
Volume: 442, Astronomical Data Analysis Software and Systems XX (ADASSXX)
Page: 119
Authors: Peng, N.; Zhang, Y.; Zhao, Y.
Abstract: Recently, the development in highly parallel Graphics Processing Units (GPUs) provides us a new method to solve advanced computation problems. We introduce an automated method called Support Vector Machine (SVM) based on Nvidia's Compute Unified Device Architecture (CUDA) platform for classifying celestial objects. SVM has been proved a good algorithm for separating quasars from stars, but it takes a lot of time for training and predicting with large samples. Using the data adopted from the Sloan Digital Sky Survey (SDSS) Data Release Seven (DR7), CUDA-accelerated SVM shows achieving greatly improved speedups over commonly used SVM software running on a CPU. It achieves speedups of 1.25–9.96× in training and 9.29–364.4× in predicting. This approach is effective and applicable for quasar selection in order to compile an input catalog for the Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST).
Back to Volume