PuSH - Publication Server of Helmholtz Zentrum München

Götz, M.* ; Debus, C.* ; Coquelin, D.* ; Krajsek, K.* ; Comito, C.* ; Knechtges, P.* ; Hagemeier, B.* ; Tarnawa, M.* ; Hanselmann, S.* ; Siggel, M.* ; Basermann, A.* ; Streit, A.*

HeAT - A distributed and GPU-accelerated tensor framework for data analytics.

Big Data, 276-287 (2020)
DOI Order publishers version
To cope with the rapid growth in available data, the efficiency of data analysis and machine learning libraries has recently received increased attention. Although great advancements have been made in traditional array-based computations, most are limited by the resources available on a single computation node. Consequently, novel approaches must be made to exploit distributed resources, e.g. distributed memory architectures. To this end, we introduce HeAT, an array-based numerical programming framework for large-scale parallel processing with an easy-to-use NumPy-like API. HeAT utilizes PyTorch as a node-local eager execution engine and distributes the workload on arbitrarily large high-performance computing systems via MPI. It provides both low-level array computations, as well as assorted higher-level algorithms. With HeAT, it is possible for a NumPy user to take full advantage of their available resources, significantly lowering the barrier to distributed data analysis. When compared to similar frameworks, HeAT achieves speedups of up to two orders of magnitude.
Additional Metrics?
Edit extra informations Login
Publication type Article: Journal article
Document type Scientific Article
Keywords Big Data Analytics ; Dask ; Gpu ; Heat ; High-performance Computing ; Machine Learning ; Message Passing Interface ; Model Parallelism ; Numpy ; Parallel Application Frameworks ; Pytorch ; Tensor Framework
ISSN (print) / ISBN 2167-6461
e-ISSN 2167-647X
Journal Big Data
Quellenangaben Volume: , Issue: , Pages: 276-287 Article Number: , Supplement: ,
Publisher Mary Ann Liebert
Reviewing status Peer reviewed
Institute(s) Helmholtz AI - FZJ (HAI - FZJ)
Helmholtz AI - KIT (HAI - KIT)