OpenGADGET3 User Meeting

Europe/Ljubljana
Faculty of Mathematics and Physics, University of Ljubljana

Faculty of Mathematics and Physics, University of Ljubljana

Jadranska 19 1000 Ljubljana
Description

Cosmological simulation codes are nowadays the most advanced tools to capture the complexity of the evolution of the cosmic structure and of its components. This codes, that are run on massively parallel HPC infrastructures, have now to be revised and adapted to exascale facilities.

This workshop will bring together collaborators involved in the development and application of the OpenGADGET cosmological simulation code and in various aspects of extragalactic astrophysics and cosmology.

Participants
  • Alice Damiano
  • Anna Ivleva
  • Cinthia Ragone-Figueroa
  • Daniel Karner
  • Dunja Fabjan
  • Frederick Groth
  • Geray Karademir
  • Gian Luigi Granato
  • Giuseppe Murante
  • Laura Di Federico
  • Luca Tornatore
  • Massimiliano Parente
  • Michela Esposito
  • Milena Valentini
  • Paramita Barai
  • Stefano Borgani
  • Tirso Marin Gilabert
  • +11
    • 14:00 15:30
      Registration 1h 30m
    • 15:30 16:00
      On the transition from proto-clusters to nearby clusters 30m

      I will provide an overview of the current status of simulations of proto-clusters and how they evolve to the population of mature low-redshift clusters. In particular, I will critically discuss open issues related to: (a) the strong star formation in proto-clusters; (b) the low level of star formation in nearby BCGs; (c) the metal share in stars and ICM.

      Speaker: Stefano Borgani (Istituto Nazionale di Astrofisica (INAF))
    • 16:00 16:30
      SubDLe: Substructure identification with Deep Learning 30m

      The identification of substructures within halos in cosmological hydrodynamical simulations is a fundamental step to identify the simulated counterparts of real objects, namely galaxies. For this reason, substructure finders play a crucial role in extracting relevant information from the simulation outputs. They are based on physically-motivated definitions of substructures, performing multiple steps of particle-by-particle operations, thus computationally expensive. The purpose of this work is to develop a fast algorithm to identify substructures in simulations. The final aim, besides a faster production of subhalo catalogues, is to provide an algorithm fast enough to be applied with a fine time-cadence during the evolution of the simulations. We chose to apply the architecture of a well known Fully Convolutional Network, U-Net, to the identification of substructures within the mass density field of the simulation. We have developed SubDLe (Substructure identification with Deep Learning), an algorithm which combines a 3D generalization of U-Net and a Friends-of-Friends algorithm, and trained it to reproduce the identification of substructures performed by the SubFind algorithm in a set of zoom-in cosmological hydrodynamical simulations of galaxy clusters. For the feasibility study presented in this work, we have trained and tested SubDLe on galaxy clusters at z=0, using a NVIDIA P100 GPU. We focused our tests on the version of the algorithm working on the identification of purely stellar substructures, stellar SubDLe. Our stellar SubDLe is capable of identifying the majority of galaxies in the challenging high-density environment of galaxy clusters in short computing times. This result has interesting implications in view of the possibility of integrating fast subhalo finders within simulation codes, that can take advantage of accelerators available on state-of-art computing nodes.

      Speaker: Michela Esposito (Istituto Nazionale di Astrofisica (INAF))
    • 16:30 17:00
      coffee break 30m
    • 17:00 17:30
      The future (and past) of galaxy clusters and superclusters 30m

      Galaxy superclusters are exciting laboratories for a number of interesting processes in galaxy and cluster formation and evolution. Using GadgetIO, we simulate a cosmological box implementing constrained initial conditions from the local universe far into the future, obtaining final collapse regions for a number of local structures. While these regions contain interesting information about the environments they collapse in themselves, it is also possible to define a new type of zoom-in initial conditions from their associated lagrangian volumes. These zoom-in simulations of local clusters will provide insight about our cosmic neighborhood in unprecedented detail.

      Speaker: Benjamin Seidel (University Observatory Munich)
    • 17:30 18:00
      The Role of Viscosity in Galaxy Clusters 30m

      The evolution of galaxy clusters is highly influenced by the dynamics of the Intracluster Medium (ICM), which governs crucial aspects. This includes mixing, turbulence processes, and galaxy interactions within the cluster environment. Among the factors influencing the ICM dynamics, the impact of viscosity is still under debate. Understanding the effect of viscosity on the evolution of galaxy clusters is fundamental for comprehending gas properties and the underlying dynamics within the ICM.

      By conducting a thorough study, we aim to highlight the implications that viscosity introduces compared to inviscid simulations. These implications encompass morphological differences, larger density fluctuations, and the intricate interplay with dynamo amplification, among other fundamental effects. Our results challenge prior assumptions, especially concerning the constraints on viscosity within the ICM. This study is expected to enhance our understanding of ICM dynamics and contribute to our knowledge of galaxy cluster evolution.

      Speaker: Tirso Marin Gilabert (University Observatory of Munich)
    • 18:00 18:30
      Projection Bias in galaxy cluster weak-lensing for Euclid Survey 30m

      Galaxy clusters are the most massive gravitationally bound objects in the Universe. Their number density offers a sensitive probe for the growth rate of large scale structure and the underlying cosmology. A way to produce this type of analysis is relying on wide-field optical imaging and photometric surveys, since they are capable of providing both large cluster samples and weak gravitational lensing mass calibration. The combination of information from three cosmological observables such as cluster counts, clustering (i.e. 2-point correlation function) of clusters and weak lensing mass makes it possible to break some degeneracies between cosmological parameters and parameters defining scaling relations between lensing masses and richness, allowing for much more refined constraints on them via Bayesian inference techniques. In this scenario the photometric survey of galaxy clusters that is being carried out by the Euclid space telescope [Laureijs et al., 2011], will reach unprecedented sensitivity and, in turn, unprecedented precision on the expected constraints on cosmological parameters. In order to harvest the full potential of Euclid's data a good understanding of systematics is mandatory.

      In this perspective, my PhD project will concentrate on the characterisation of the systematics related to projection effects and their impact on the scaling relation between weak lensing masses and observed richness of galaxy clusters to be identified in the Euclid wide survey. As already discussed in the literature [e.g. Fumagalli et al., 2023, and references therein], projection effects occur when multiple foreground and background objects along the same line of sight are mistakenly associated with a galaxy cluster, increasing the apparent richness of the cluster. While this effect is arguably of relative importance in the SDSS cluster survey, it has been shown to produce a significant effect on the cosmological posteriors derived from the Dark Energy Survey (Costanzi et al. [2019, 2021], Salcedo et al. [2023]). We then expect that, if not accurately characterised, projection effects should have an even stronger impact in severely limiting the cosmological constraining power of the Euclid photometric cluster survey.

      Different methods to estimate and mitigate the projection effect have been recently proposed (e.g. Sunayama [2023], Wu et al. [2022]). The common approach these method shares, is to find a suitable estimator for the boost in observed lensing signal induced by the bias (e.g the ratio between the observed signal from a richness-selected sample and the signal expected from the underlying halo mass PDF). Once this estimator has been properly crafted, one can think of applying some mitigation methods that can range from simulating different galaxy models to apply multi-wavelength or multi-tracers techniques. The end result for all these methodologies should be a model capable of removing the bias effect from the cosmological parameters inference.

      Starting from the work done in the above papers, my PhD project, that will develop within the activities of the Science Working Group of Galaxy Clusters (SWG-CL) of the Euclid Consortium (EC) will consist in finding efficient way (exploring some of the different approaches mentioned above) to quantify projection effects by using suitably designed mocks Euclid photometric galaxy surveys. These synthetic Euclid mock cluster surveys will be generated using a large set of in-house cosmological N-body simulations and populating them with galaxies. In this way, I will aim at quantifying the impact of projection effects on cosmological posteriors, and to define a model to correct for them.

      Speaker: Roberto Ingrao (University of Trieste)
    • 18:30 19:30
      break & thematic discussion 1h
    • 09:00 09:30
      Neighbourhood Issues: The Impact of Environment on Galaxy Properties at Cosmic Noon 30m

      Recent observations of galaxies at high redshift have challenged our understanding of galaxy formation: some have extremely high star formation rates unmatched by anything seen at present day, others show signs of rapid inflows or outflows, and indications of very massive black holes. In particular observations of massive quenched galaxies present when the Universe was only 1Gyr old have proven difficult to reproduce. I will show that this is because simulations require both large volumes as well as very high resolutions to capture these observed properties. This is especially true as many of such galaxies live in environments that will eventually collapse into massive groups or clusters at present day, but at high redshift are typically spread over a large area of sky as the protocluster and protogroup structures are only beginning to assemble.

      Using one of the largest sets of fully hydrodynamical cosmological simulations, the Magneticum simulation suite, I show how quenched galaxies are formed at such high redshifts, how they are connected to the environment they form in, and how the feedback from these galaxies enriches the surrounding gas with metals. Furthermore, I discuss the properties of the galaxies in the first emerging protocluster cores, demonstrating that the hot gas atmospheres of the protoclusters start to build up at redshifts around z=4, and how this impacts the galaxy populations in these environments. Finally, I will show that the extent of the future collapsing regions can be traced by probing the large scale environment, thus allowing for protocluster regions to be distinguished from proto groups and massive galaxies that will starve in isolation.

      Speaker: Rhea-Silvia Remus
    • 09:30 10:00
      Universal Blacksmith: What Metals Reveal About the Histories of Galaxies 30m

      As increasingly more detailed observational data allows the probing of multiple metal lines in both stars and gas, a new window has opened into understanding the formation and accretion of galaxies' stars. In particular the ratio of alpha-elements such as oxygen and magnesium to iron contain information about the amount of enrichment from faster acting supernovae type II versus slower acting type Ia. Using the hydrodynamical cosmological simulation Magneticum Pathfinder I probe the evolution through cosmic time of different metals both locked into stars as well as within the gas of galaxies. I show how [alpha/Fe] and the stellar mass can constrain the galaxies' formation redshift, how the contribution of SN type Ia rapidly rises until z~1 before saturating, and also how individual galaxies evolve in the [alpha/Fe] vs [Fe/H] plane. Finally, I discuss where the observed high [alpha/Fe] at cosmic dawn originate from and how these earliest galaxies reach super-solar metallicities so quickly.

      Speaker: Lucas Kimmig
    • 10:00 10:30
      Tracing In- & Outflows of Galaxies in the Magneticum Pathfinder Simulations 30m

      The process of gas accreting from the cosmic web onto galaxies is a fundamental aspect of galaxy formation. Simulations provide the tools to directly trace gas flows over time and are therefore crucial to test models and assumptions about accretion modes and geometry, which can then be used to interpret observational results. Using the Magneticum Pathfinder simulation box 4 (uhr), I trace gas particles flowing in and out of galaxies to relate modes of accretion to the evolution of galaxy features with respect to scaling relations. By the virtue of gas particles in this simulation suite being able to convert only a fraction of their mass into stars, each particle can undergo the process of being accreted, form stars, absorb stellar feedback and flow out of the ISM. I retrieve accretion rates between the galaxies and their virial radii from particle positions, phase space extrapolation, shell volumes and mass evolution and test the prediction of corresponding star formation rates to compare different methods of mass flow measurements in a Lagrangian simulation. By using the built-in two-phase gas model as well as a simple temperature criterion, I present how the cold and hot accretion modes relate to the formation histories and the connection between environment, circum-galactic medium and interstellar medium as well as the prevalence of the galactic fountain concept.

      Speaker: Silvio Fortuné (USM, LMU Munich)
    • 10:30 11:00
      Coffee break 30m
    • 11:00 11:30
      Stellar Populations of Shells and Streams - Where They Came from and How They Got There 30m

      Tidal features in the outskirts of galaxies detected by low surface brightness observations provide a unique pathway to study their assembly history. The upcoming Vera C. Rubin Observatory will provide a vast number of galaxies exhibiting such features, while integral field unit observations enable the study of their stellar population properties. I employ the hydrodynamical cosmological simulation Magneticum Pathfinder to study the stellar properties of shells and streams. Tracing the stellar particles of tidal features back in time allows me to identify the progenitor satellite galaxies and connect their properties to the tidal feature they formed. I will present spatially resolved maps of the stellar velocity dispersion, mass, age, and metallicity in galaxies exhibiting shells and streams as well as the behavior of these properties within the tidal features themselves. I find that shells and streams generally appear as depressions in the velocity dispersion, while only some streams and shells appear younger and more metal rich than their surroundings. Furthermore, I will discuss the connection between these properties and the radial velocity fraction, gas and stellar mass, and the depth of the gravitational potential of the features’ progenitor satellite galaxies. Finally, I will compare the spatial extent and orientation of the feature to the half-mass radius and the orbit of the progenitor.

      Speaker: Johannes Stoiber (University Observatory, Faculty of Physics, Ludwig-Maximilians-Universität München)
    • 11:30 12:00
      Utilizing idealized setups for a targeted investigation of formation pathways of dwarf galaxies 30m

      Observations of interacting galaxies in a cluster environments indicate that the triggered star formation activity inside the tidal tails is extending far beyond what we would expect from mergers in the field. Allowing to probe the possible parameter space, isolated setups are posing as the ideal testing ground to investigate the details and consequences of such behavior, as well as for a targeted examination of observed objects. Utilizing the code OpenGADGET-3, I present three hydrodynamic simulations of a major merger inside a galaxy cluster in high resolution, where the initial conditions were tuned to correspond to the observed merger NGC 5291 in cluster Abell 3574. Compared to isolated galaxy merger simulations, I find strong changes in the merger morphology, as well as drastically increased star formation activities due to ram pressure. In particular, this environmental influence is efficiently assisting the formation of tidal dwarf galaxies, which form in the pre-enriched gas ejected by the merger. By comparing our simulations to observational data, I demonstrate that such a process is capable of reproducing characteristics of a full variety of observed dwarf galaxy types. Comparing their contribution to the observed galaxy mass function in clusters, I estimate that ~30% of dwarf galaxies in galaxy cluster environments may have been formed through stripping from mergers.

      Speaker: Anna Ivleva (University Observatory Munich)
    • 12:00 14:30
      Lunch break & free time 2h 30m
    • 14:30 15:00
      Subresolution description for supernova remnants as sources of cosmic rays 30m

      In large-scale simulations, that also include spectral cosmic-ray physics, high-energy protons and electrons accelerated at the shocks of supernova remnants have to be described by a sub-grid model. Usually, the injected cosmic rays are represented by a simple power-law spectrum in momentum space. However, in the recent past several models for more realistic cosmic-ray spectra from supernova remnants have been published. They rely heavily on results from state-of-the-art simulations of particle acceleration at strong, collisionless shocks. By combining different approaches, tabulated spectra can be generated, which are the basis for a physically motivated sub-grid description for cosmic-ray seeding by supernova remnants in OpenGadget3. With this code simulations of galaxies in isolated and dense environments can be performed in order to check if the more realistic cosmic-ray spectra lead to quantitative differences in galactic properties and non-thermal radiation.

      Speaker: Daniel Karner (USM, LMU Munich)
    • 15:00 15:30
      OpenGADGET3 in SPACE 30m

      The next generation of supercomputers will use a heterogeneous design to achieve exaflop capabilities. While this heterogeneous design offers significantly larger computational power it also increases the complexity of such systems. This has a significant impact on simulation codes to be run on these machines. Since the focus in the past was mainly on the scalability of CPUs with the introduction of GPUs new methods have to be found to adapt the code. By not being able to use the available GPUs properly, a significant fraction of computational power would be unusable, which would result in a hard barrier for any large simulations in the future.

      Scalable Parallel Astrophysical Codes for Exascale (SPACE) is an EU Centre of Excellence focused on astrophysical and cosmological applications in the advent of exascale computation. In this session, we will introduce the CoE and its mission, as well as results from activities within the CoE for OpenGADGET3.

      Speakers: Geray Karademir (USM, LMU Munich), Dr Luca Tornatore (INAF)
    • 15:30 16:00
      Coffee break 30m
    • 16:00 17:30
      Thematic discussion
    • 19:30 22:30
      Social dinner
    • 09:00 09:30
      Latest results on the massive black holes' dynamics in OpenGadget3 30m

      The dynamics of Massive Black Holes (MBHs) is primarily driven by the dynamical friction force, the drag induced by the surrounding "sea" of matter. This force anchors MBHs in the core of massive galaxies and leads to the formation of a close-pair binary BH system during mergers. Despite its fundamental role in governing the BH dynamics, the complexity of its description still represents a challenge for both analytical and numerical calculations. Recently, we have developed and extensively tested a new dynamical friction correction in OpenGADGET3 that proved to be an efficient way to control the dynamics of MBH in various environments.

      Looking forward to exploiting the impact of an improved tracing of MBH on the galaxy evolution, we started to stress the dynamical friction technique within the context of idealised simulations of dark matter halo and galaxies. These tests provide informations on the sensitivity of BH dynamics and dynamical friction correction to mass resolution and its response to the BH’s softening. Moreover, these studies help to establish the MBHs’ sinking timescale for dynamical friction, the primary uncertainty in the MBHs’ merger rate estimation and resulting gravitational waves emission.

      Speaker: Alice Damiano (INAF)
    • 09:30 10:00
      Supermassive black hole spin evolution in cosmological simulations with OpenGadget3 30m

      Central massive black holes (BHs) and their host galaxies are thought to co-evolve also due to the complex interaction that arises when the BHs accrete gas and release a large amount of energy back into the surrounding environment.
      Important actors in this scenario are powerful jets, and the spin of the central black hole is thought to be a key physical parameter that determines their power and direction.
      To carry out a self-consistent, statistically meaningful study of the role of BH spins and related jet feedback, I implemented a sub-resolution model for cosmological hydrodynamical simulations, that evolves the BH spin due to the occurrence of misaligned gas accretion.
      The model provides a more complete physical description of the accretion process from the resolved scales, adding the important spin parameter to the BHs in simulations. I will present the results of simulations that test such a model in a fully cosmological context, which allows to carry out statistical studies on the distribution of spins and radiative efficiencies.

      Speaker: Luca Sala (Faculty of Physics at the Ludwig-Maximilians-University - University Observatory Munich)
    • 10:00 10:30
      Combined Thermal and Kinetic AGN Feedback in OpenGADGET 30m

      We study different numerical implementations of distributing the feedback energy from a central SMBH to the surrounding gas. We have incorporated a new sub-resolution model for the switching between thermal and kinetic AGN feedback modes depending on the BH accretion rate. Test simulations are being performed with cases of only thermal feedback, only kinetic feedback, and the combined thermal-kinetic scheme. In the meeting, I expect to present some results from these simulations of various AGN feedback models.

      Speaker: Paramita Barai (Istituto Nazionale di Astrofisica (INAF))
    • 10:30 11:00
      Coffee break 30m
    • 11:00 12:00
      Thematic discussion: HowTo Session
    • 12:00 14:30
      Lunch break & free time 2h 30m
    • 14:30 15:00
      Clues on SMBH growth and galaxy quenching from a large scale perspective 30m

      Galaxy evolution models typically rely on SMBH feedback to quench massive galaxies. Therefore, accurately modeling the growth of these objects in our simulations is essential. In this talk, I will present an analysis of the star formation in local galaxies across different cosmic web environments, namely voids, walls, filaments, and nodes.
      Interpreted using a semi-analytic model, the analysis shows that quenching occurs similarly in massive galaxies regardless of the environment. Consequently, the growth of SMBHs, responsible for galaxy quenching, must be environment-independent. These results serve as a critical test for our galaxy evolution simulations.

      Speaker: Massimiliano Parente (Istituto Nazionale di Astrofisica (INAF))
    • 15:00 15:30
      Intertwined Formation of H2, Dust, and Stars in Cosmological Simulations 30m

      In the metal-enriched interstellar medium, the abundance of molecular gas is primarily governed by the formation of H2 on dust grains, as well as its self-shielding and shielding by dust against photo-dissociation by the interstellar radiation field. The upcoming presentation intends to describe a sub-resolution model for forming molecular hydrogen in hydrodynamic simulation with dust description and encouraging results in predicting the properties of galaxies in cosmological boxes.

      Speaker: Cinthia Ragone Figueroa (Instituto de Astronomía Teórica y Experimental (IATE) CONICET-UNC)
    • 15:30 16:00
      Coffee break 30m
    • 16:00 16:30
      Towards a New Cooling Module in OpenGADGET3: Step 1 - Cooling Tables 30m

      In this talk, I will present the current status of the new cooling implementation in OpenGADGET3. Specifically, I will refer to the first step of this project: the creation of new cooling tables, whose values were obtained using the CLOUDY code.

      Speaker: Laura Di Federico (Universitäts-Sternwarte München)
    • 17:30 18:30
      Visit to the Computer History Museum
    • 09:00 09:30
      Modelling X-ray emission from simulations with Phox 30m

      Hydrodynamic Cosmological Simulations offer a unique perspective in studying various aspects of structure formation in the universe. In particular, they hold great value in quantifying statistical properties such as the composition and enrichment of the baryonic matter distribution in the knots and filaments of the cosmic web. While it is observationally challenging to extract information from faint, low-density gas found in filaments and outskirts of galaxies, the projected capabilities of future X-ray missions provide new pathways towards a deeper understanding of these environments. Presented here is a collection of X-ray studies performed on the hydrodynamical simulation suite "Magneticum".
      We make use of the matter properties traced by the simulation to investigate various X-ray properties of the CGM and ISM in poster-child disk and elliptical galaxies from the simulation while also taking into account contamination from stellar sources such as X-ray binaries. With these studies we aim to provide constraints and limiting factors towards observational targets.

      Speaker: Stephan Vladutescu-Zopp (University Observatory Munich)
    • 09:30 10:00
      MFM in OpenGadget3 30m

      In this talk, I will present the current status of the implementation of Meshless Finite Mass (MFM) in OpenGadget3, which I implement as an alternative to the already implemented modern SPH solver.

      MFM has several advantages over other solvers, such as developing mixing instabilities and capturing the power spectrum of subsonic turbulence. Thus, it also improves the description of turbulence in the ICM of galaxy clusters. I will show some applications on idealized setups and a study of turbulence in galaxy clusters, also comparing the impact of the hydrodynamical scheme.

      Speaker: Mr Frederick Groth (University-Observatory Munich)
    • 10:00 10:30
      Tales from the Spectral Cosmic Ray Model 30m

      I'll give an update on the state of the Spectral Cosmic Ray model in OpenGadget3.

      Speaker: Ludwig Böss (University Observatory Munich)
    • 10:30 11:00
      Coffee break 30m
    • 11:00 12:00
      Thematic discussion: Final remarks & discussion
    • 12:00 14:00
      Lunch break 2h