Listar ICC_Articles por autoría "53eaed22-5bd1-4d20-b771-17485699afcf"
Mostrando ítems 1-3 de 3
-
Analyzing the impact of the MPI allreduce in distributed training of convolutional neural networks
Castelló, Adrián; Catalán Carbó, Mar; Dolz, Manuel F.; Quintana-Orti, Enrique S.; Duato, José Springer (2022-01-10)For many distributed applications, data communication poses an important bottleneck from the points of view of performance and energy consumption. As more cores are integrated per node, in general the global performance ... -
PyDTNN: A user-friendly and extensible framework for distributed deep learning
Barrachina Mir, Sergio; Castelló, Adrián; Catalán Carbó, Mar; Dolz, Manuel F.; Mestre Miravet, Jose Ignacio Springer (2021-02-22)We introduce a framework for training deep neural networks on clusters of computers with the following appealing properties: (1) It is developed in Python, exposing an amiable interface that provides an accessible entry ... -
Using machine learning to model the training scalability of convolutional neural networks on clusters of GPUs
Barrachina Mir, Sergio; Castelló, Adrián; Catalán Carbó, Mar; Dolz, Manuel F.; Mestre Miravet, Jose Ignacio Springer (2021-08-30)In this work, we build a general piece-wise model to analyze data-parallel (DP) training costs of convolutional neural networks (CNNs) on clusters of GPUs. This general model is based on i) multi-layer perceptrons (MLPs) ...