This is a repository for GPU acceleration of Machine Learning algorithms, including CNN and MLP. This repository is expected to continuously be maintained.
Leading increased performance in benchmark tasks and to enable discovery of complex
high-level features, scaling up deep learning algorithms have earned researchers’
interests. Recent works train deep neural networks with deep layers and very high
dimension of parameters with the assistance of vast amount of computing power. The
Graphics Processing Units (GPUs) have been applied successfully in many areas for parallel
computing in recent years. Compared with the traditional CPU cluster, GPU has an obvious
advantage of low cost of hardware and electricity consumption.
Deep learning algorithms including Deep Neural Networks (DNN), whose forward and
backward propagation contains many inner products or matrix multiplications, which can
utilize the parallelism of GPU.
In this work, we are going to perform optimizations on the implementations of DNN on
GPU and heterogeneous platforms including AMD APU, Intel CPU + Nvidia GPU, AMD CPU,
and Intel CPU. Evaluation and analysis of the implementations are compared on respective
platforms.
Copyright 2016.
For any questions, feel free to let me know.