English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Artificial Intelligence (AI) applications based on Deep Neural Networks (DNN)or Deep Learning (DL) have become popular due to their success in solvingproblems likeimage analysis and speech recognition. Training a DNN iscomputationally intensive and High Performance Computing(HPC) has been a keydriver in AI growth. Virtualisation and container technology have led to theconvergence of cloud and HPC infrastructure. These infrastructures with diversehardware increase the complexity of deploying and optimising AI trainingworkloads. AI training deployments in HPC or cloud can be optimised withtarget-specific libraries, graph compilers, andby improving data movement orIO. Graph compilers aim to optimise the execution of a DNN graph by generatingan optimised code for a target hardware/backend. As part of SODALITE (a Horizon2020 project), MODAK tool is developed to optimise application deployment insoftware defined infrastructures. Using input from the data scientist andperformance modelling, MODAK maps optimal application parameters to a targetinfrastructure and builds an optimised container. In this paper, we introduceMODAK and review container technologies and graph compilers for AI. Weillustrate optimisation of AI training deployments using graph compilers andSingularity containers. Evaluation using MNIST-CNN and ResNet50 trainingworkloads shows that custom built optimised containers outperform the officialimages from DockerHub. We also found that the performance of graph compilersdepends on the target hardware and the complexity of the neural network.

Optimising AI Training Deployments using Graph Compilers and Containers