Replication Data for: Load-Balancing for Scalable Simulations with Large Particle Numbers (doi:10.18419/darus-1851)

View:

Part 1: Document Description
Part 2: Study Description
Part 3: Data Files Description
Part 4: Variable Description
Part 5: Other Study-Related Materials
Entire Codebook

Replication Data for: Load-Balancing for Scalable Simulations with Large Particle Numbers

doi:10.18419/darus-1851

DaRUS

2021-06-07

1

Hirschmann, Steffen, 2021, "Replication Data for: Load-Balancing for Scalable Simulations with Large Particle Numbers", https://doi.org/10.18419/darus-1851, DaRUS, V1, UNF:6:WXqt4IDAIDGnJB9tcwMhdg== [fileUNF]

Replication Data for: Load-Balancing for Scalable Simulations with Large Particle Numbers

doi:10.18419/darus-1851

Hirschmann, Steffen (Universität Stuttgart)

SFB 716

EXC 2075 - 390740016

DaRUS

Hirschmann, Steffen

Hirschmann, Steffen

2021-05-14

https://doi.org/10.18419/darus-1851

Computer and Information Science

Parallel algorithms in computer science (68W10), Distributed algorithms (68W15), Approximation algorithms (68W25), Parallel numerical computation (65Y05), Applications to the sciences (65Z05), Performance evaluation, queueing, and scheduling in the context of computer systems (68M20), Turbulent combustion; reactive turbulence (76F80), Combustion (80A25)

This dataset contains input data, scripts, etc. for replicating the numerical experiments in Steffen Hirschmann's dissertation.<br> The data is prefixed with folders indicating the specific experiment as is the processing metadata.<br> The experiments are: <ul> <li>00_periodicity_experiment: Experiment in Chapter 5.1. Agglomerate moves over boundary and we inspect how well different partitioning methods perform.</li> <li>01_homogeneous_scaling: Experiment in Chapter 8.1. Weak scaling of ESPResSo with and without our additions of a homogeneous fluid.</li> <li>02_droplet_formation: Experiment in Chapter 8.1. Comparison of the behavior of a non-homogeneous particle distribution with and without load-balancing.</li> <li>03_coupled_lbm_md: Experiment in Chapter 8.2. Weak scaling evaluation of our joint Lattice-Boltzmann (LBM) Molecular Dynamics (MD) partitioning.</li> <li>04_lb_adaptions: Experiments in Chapter 8.3. Simulation of a spinodal decomposition with different load-balancing methods and different variants.</li> <li>05_heterogeneity: Script for evaluating the heterogeneity measure defined in Chapter 7.1.</li> <li>06_soot_particle_agglomeration: Script to simulate the soot particle agglomeration scenario. Methodology described in Chapter 4, simulation described in Chapter 8.4.</li> </ul> The indications to chapters reference the publication described in "Related Publication".

program source code

input data

See "Related Publication".

Hirschmann, S.: Load-Balancing for Scalable Simulations with Large Particle Numbers. PhD thesis. University of Stuttgart, 2021, submitted.

Hirschmann, S.: Load-Balancing for Scalable Simulations with Large Particle Numbers. PhD thesis. University of Stuttgart, 2021, submitted.

• Number of cases: 12

• No. of variables per record: 1

• Type of File: text/tab-separated-values

UNF:6:E2YeVaO+v2VpzS+UkjzWWg==

• Number of cases: 10

• No. of variables per record: 1

• Type of File: text/tab-separated-values

UNF:6:0J2k1mxPjtjCMq6q8Vbw6g==

• x graph p4est - x graph p4est
• nproc default p4est - nproc default p4est

f63047 Location:

Variable Format: character

Notes: UNF:6:E2YeVaO+v2VpzS+UkjzWWg==

f63027 Location:

Variable Format: character

Notes: UNF:6:0J2k1mxPjtjCMq6q8Vbw6g==

Label:

agglomerate-1954.txt

Text:

Positions of the particles constituting the agglomerate. One particle position as X, Y, Z coordinates per row. Positions are measured in multiples of 20nm.

text/plain

Label:

plot-results.py

text/x-python

Label:

plot-setup.py

Text:

Plots the setup. Input is a txt file of the particle positions (agglomerate-1954.txt).

text/x-python

Label:

simulate-agglomerate.py

text/x-python

Label:

generate-grid-mpiio-data.cc

text/x-c++src

Label:

measurement.tcl

text/x-tcl

Label:

plot-scaling.py

text/x-python

Label:

lj-default-1700ppp.dat

Text:

Average, minimum and maximum runtime as well as imbalance every 4000 time steps. Data generated using the standard Cartesian domain decomposition of ESPResSo.

text/x-fixed-field

Label:

lj-reba-1700ppp.dat

Text:

Average, minimum and maximum runtime as well as imbalance every 4000 time steps. Data generated using the SFC-based spatial domain decomposition optimized for the number of particles per subdomain.

text/x-fixed-field

Label:

parse-output-files.py

text/x-python

Label:

plot.py

text/x-python

Label:

simulation_script.tcl

text/x-tcl

Label:

lbm-md-coupled-sim.py

text/x-python

Label:

run.sh

application/x-shellscript

Label:

extract-all.awk

application/x-awk

Label:

spinodal-load-balancing.py

text/x-python

Label:

spinodal_decomposition.py

text/x-python

Label:

calculate_heterogeneity.py

text/x-python

Label:

case6-100mio.py

text/x-python

Label:

generate-uniform-random-mpiio-data.cc

text/x-c++src

Label:

espresso-4.1.2+generic_dd-1.zip

Text:

ESPResSo 4.1.2 with generic_dd module to use librepa as backend to ESPResSo.

application/zip

Label:

espresso-old_custom.zip

Text:

Old version of ESPResSo with custom additions to perform coupled load-balancing.

application/zip

Label:

flowfield_patch_for_ESPResSo-4.1.2+generic_dd.diff

Text:

Patch to allow for the inclusion of an external time-dependent flow field like in our soot particle agglomeration simulation methodology.

text/x-patch

Label:

measurement_patch_for_ESPResSo-4.1.2+generic_dd.diff

Text:

Patch to add measurements to ESPResSo-4.1.2+generic_dd

text/x-patch