Details of the publication

As the world becomes increasingly digitalised, data centres and data transmission networks are emerging as an important source of energy demand.

In this post, I want to continue building out an incremental model, but rather than focusing on emissions calculations I want to create a model to estimate energy use.

Towards environmentally sustainable computational science

Details of the publication

Deep Learning (DL) frameworks such as PyTorch and TensorFlow include runtime infrastructures responsible for executing trained models on target hardware, managing memory, data transfers, and multi-accelerator execution, if applicable. Additionally, it is a common practice to deploy pre-trained models on environments distinct from their native development settings. This led to the introduction of interchange formats such as ONNX, which includes its runtime infrastructure, and ONNX Runtime, which work as standard formats that can be used across diverse DL frameworks and languages. Even though these runtime infrastructures have a great impact on inference performance, no previous paper has investigated their energy efficiency. In this study, we monitor the energy consumption and inference time in the runtime infrastructures of three well-known DL frameworks as well as ONNX, using three various DL models. To have nuance in our investigation, we also examine the impact of using different execution providers. We find out that the performance and energy efficiency of DL are difficult to predict. One framework, MXNet, outperforms both PyTorch and TensorFlow for the computer vision models using batch size 1, due to efficient GPU usage and thus low CPU usage. However, batch size 64 makes PyTorch and MXNet practically indistinguishable, while TensorFlow is outperformed consistently. For BERT, PyTorch exhibits the best performance. Converting the models to ONNX usually yields significant performance improvements but the ONNX converted ResNet model with batch size 64 consumes approximately 10% more energy and time than the original PyTorch model.

Cet article qui résume un rapport de Greenspector explore les impacts environnementaux des différents éléments d'une page web. À travers une méthodologie rigoureuse de mesure, il révèle l'importance de la compréhension du "poids" numérique de chaque composant pour une écoconception web efficace, tout en remettant en question certains modèles mentaux sur les impacts environnementaux du numérique.

A calculator that allows users to translate abstract greenhouse gas amounts into concrete terms that are easy to understand.

The JRC supports the implementation of European sustainable product policies such as EU Ecolabel, Green Public Procurement, Eco-design of Energy...

[vc_row type=”in_container” full_screen_row_position=”middle” column_margin=”default” column_direction=”default” column_direction_tablet=”default” column_direction_phone=”default” scene_position=”center” top_padding=”3%” bottom_padding=”1%” constrain_group_2=”yes” bottom_padding_tablet=”0″ constrain_group_4=”yes” constrain_group_6=”yes” text_color=”dark” text_align=”left” row_border_radius=”none” row_border_radius_applies=”bg” overflow=”visible” overlay_strength=”0.3″ gradient_direction=”left_to_right” shape_divider_position=”bottom” bg_image_animation=”none” gradient_type=”default” shape_type=””][vc_column...

Regular readers of Uptime Institute’s annual data center survey, the longest running of its kind, already know that the industry average power usage effectiveness (PUE, a ratio of total site power and IT power) has trended sideways in recent years. Since 2020, it has been stuck in the 1.55 to 1.59 band. Even going back […]

This contributed volume provides illuminating perspectives on reducing power consumption in IP-based video distribution networks (VDNs).

Request PDF | On Jan 1, 2014, Olivier Philippot and others published Characterization of the energy consumption of websites: Impact of website implementation on resource consumption | Find, read and cite all the research you need on ResearchGate