Modeling Scalability of Distributed Machine Learning

Abstract

Abstract-Present day machine learning is computationally intensive and processes large amounts of data. It is implemented in a distributed fashion in order to address these scalability issues. The work is parallelized across a number of computing nodes. It is usually hard to estimate in advance how many nodes to use for a particular workload. We propose a simple framework for estimating the scalability of distributed machine learning algorithms. We measure the scalability by means of the speedup an algorithm achieves with more nodes. We propose time complexity models for gradient descent and graphical model inference. We validate our models with experiments on deep learning training and belief propagation. This framework was used to study the scalability of machine learning algorithms in Apache Spark.

Throughput Prediction of Asynchronous SGD in TensorFlow

“…The model is aware of different convolution computation strategies, including matrix multiplication and Fast Fourier Transform. [20] models scalability based only on hardware specifications. [14] models training on GPUs.…”