% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@INPROCEEDINGS{Villamar:1014307,
author = {Villamar, Jose and Kelbling, Matthias and Terhorst, Dennis
and More, Heather and Tetzlaff, Tom and Senk, Johanna and
Thober, Stephan},
title = {{T}racking large-scale simulations through unified metadata
handling},
reportid = {FZJ-2023-03229},
year = {2022},
abstract = {Simulation is an essential pillar of knowledge generation
in science. The numerical models used to describe, predict,
and understand real-world systems are typically complex.
Consequently, applying these models by means of simulation
often poses high demands on computational resources, and
requires high-performance computing (HPC) or other dedicated
hardware architectures. Metadata describing the details of a
numerical experiment arise at all stages of the simulation
process: the conceptual description of the model, the model
implementation, and the tools and machines used to run the
simulation. Capturing these metadata and provenance
information along the processing chain is a vital
requirement for several purposes, e.g. reproducibility,
benchmarking and validation, assessment of the reliability
of the simulations, and data exploration [1,2]. The ability
to search, share, and evaluate metadata and provenance
traces from heterogeneous simulations and environments is a
major challenge in provenance-driven analysis. The
availability of a common metadata framework, which can be
adopted by scientists from different scientific domains,
would foster the meta-analysis of HPC simulation workflows
[3]. Here, we develop a metadata management framework for
generic HPC-based simulation research comprising concepts
and tools for efficiently generating, organizing, and
exploring metadata along a given simulation workflow. The
derived solutions cope with the modularity and flexibility
demands of rapidly progressing science and are applicable to
diverse research fields. As a proof of concept, we will
apply these solutions to use cases from environmental
research and computational neuroscience.[1] Guilyardi, E.,
et. al. (2013) doi: 10.1175/BAMS-D-11-00035.1[2] Manninen,
T., et. al. (2018) doi: 10.3389/fninf.2018.00020[3] Ivie,
P., $\&$ Thain, D. (2018). doi: 10.1145/3186266},
month = {Oct},
date = {2022-10-05},
organization = {Helmholtz Metadata Collaboration
conference, Virtual (Germany), 5 Oct
2022 - 6 Oct 2022},
subtyp = {After Call},
cin = {INM-6 / IAS-6 / INM-10 / IAS-9},
cid = {I:(DE-Juel1)INM-6-20090406 / I:(DE-Juel1)IAS-6-20130828 /
I:(DE-Juel1)INM-10-20170113 / I:(DE-Juel1)IAS-9-20201008},
pnm = {5232 - Computational Principles (POF4-523) / MetaMoSim -
Generic metadata management for reproducible
high-performance-computing simulation workflows - MetaMoSim
(ZT-I-PF-3-026) / HBP SGA3 - Human Brain Project Specific
Grant Agreement 3 (945539) / ACA - Advanced Computing
Architectures (SO-092) / Brain-Scale Simulations
$(jinb33_20220812)$},
pid = {G:(DE-HGF)POF4-5232 / G:(DE-Juel-1)ZT-I-PF-3-026 /
G:(EU-Grant)945539 / G:(DE-HGF)SO-092 /
$G:(DE-Juel1)jinb33_20220812$},
typ = {PUB:(DE-HGF)24},
url = {https://juser.fz-juelich.de/record/1014307},
}