C. Gheller

Swiss National Supercomputing Centre, Lugano, Ticino, Switzerland

Are you C. Gheller?

Claim your profile

Publications (64)65.31 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: The detection of the diffuse gas component of the cosmic web remains a formidable challenge. In this work we study synchrotron emission from the cosmic web with simulated SKA1 observations, which can represent an fundamental probe of the warm-hot intergalactic medium. We investigate radio emission originated by relativistic electrons accelerated by shocks surrounding cosmic filaments, assuming diffusive shock acceleration and as a function of the (unknown) large-scale magnetic fields. The detection of the brightest parts of large ($>10 \rm Mpc$) filaments of the cosmic web should be within reach of the SKA1-LOW, if the magnetic field is at the level of a $\sim 10$ percent equipartition with the thermal gas, corresponding to $\sim 0.1 \mu G$ for the most massive filaments in simulations. In the course of a 2-years survey with SKA1-LOW, this will enable a first detection of the "tip of the iceberg" of the radio cosmic web, and allow for the use of the SKA as a powerful tool to study the origin of cosmic magnetism in large-scale structures. On the other hand, the SKA1-MID and SKA1-SUR seem less suited for this science case at low redshift ($z \leq 0.4$), owing to the missing short baselines and the consequent lack of signal from the large-scale brightness fluctuations associated with the filaments. In this case only very long exposures ($\sim 1000$ hr) may enable the detection of $\sim 1-2$ filament for field of view in the SKA1-SUR PAF Band1.
    01/2015;
  • [Show abstract] [Hide abstract]
    ABSTRACT: A number of scientific numerical codes can currently exploit GPUs with remarkable performance. In astrophysics, Enzo and Ramses are prime examples of such applications. The two codes have been ported to GPUs adopting different strategies and programming models, Enzo adopting CUDA and Ramses using OpenACC. We describe here the different solutions used for the GPU implementation of both cases. Performance benchmarks will be presented for Ramses. The results of the usage of the more mature GPU version of Enzo, adopted for a scientific project within the CHRONOS programme, will be summarised.
    12/2014;
  • Source
    F. Vazza, M. Brüggen, C. Gheller, P. Wang
    [Show abstract] [Hide abstract]
    ABSTRACT: The amplification of primordial magnetic fields via a small-scale turbulent dynamo during structure formation might be able to explain the observed magnetic fields in galaxy clusters. The magnetisation of more tenuous large-scale structures such as cosmic filaments is more uncertain, as it is challenging for numerical simulations to achieve the required dynamical range. In this work, we present magneto-hydrodynamical cosmological simulations on large uniform grids to study the amplification of primordial seed fields in the intracluster medium (ICM) and in the warm-hot-intergalactic medium (WHIM). In the ICM, we confirm that turbulence caused by structure formation can produce a significant dynamo amplification, even if the amplification is smaller than what is reported in other papers. In the WHIM inside filaments, we do not observe significant dynamo amplification, even though we achieve Reynolds numbers of $R_{\rm e} \sim 200-300$. The maximal amplification for large filaments is of the order of $\sim 100$ for the magnetic energy, corresponding to a typical field of a few $\sim \rm nG$ starting from a primordial weak field of $10^{-10}$ G (comoving). In order to start a small-scale dynamo, we found that a minimum of $\sim 10^2$ resolution elements across the virial radius of galaxy clusters was necessary. In filaments we could not find a minimum resolution to set off a dynamo. This stems from the inefficiency of supersonic motions in the WHIM in triggering solenoidal modes and small-scale twisting of magnetic field structures. Magnetic fields this small will make it hard to detect filaments in radio observations.
    Monthly Notices of the Royal Astronomical Society 09/2014; 445(4). · 5.23 Impact Factor
  • Source
    Conference Paper: Splotch on the Xeon Phi
    [Show abstract] [Hide abstract]
    ABSTRACT: The management and analysis of modern large-scale datasets generated by scientific experiments (observations or numerical simulations) can be very challenging due to continuously increasing sizes and complexity. Traditional data mining and analysis methods can be expensive if applied to large-scale datasets. Visual exploration and discovery can represent invaluable tools providing scientists with prompt and intuitive insights, enabling them to identify characteristics and manageable regions of interest for applying time consuming methods. They can also be effective means for communicating scientific results not only to researchers but also to members of the general public. Splotch [1] is a rendering algorithm for large-scale datasets exploiting modern High Performance Computing (HPC) architectures. Parallelism on HPC systems is achieved by deploying the MPI [2], OpenMP and CUDA paradigms [3]. In order to address emerging technologies and exploit all potential hardware within a HPC environment the algorithm is extended, building upon the MPI and OpenMP implementations, to create an offloading model exploiting Intel’s Xeon Phi using the new Many Integrated Core (MIC) architecture and the provided C++ offload extensions.
    International Supercomputing Conference, Leipzig, Germany; 06/2014
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Splotch is a rendering algorithm for exploration and visual discovery in particle-based datasets coming from astronomical observations or numerical simulations. The strengths of the approach are production of high quality imagery and support for very large-scale datasets through an effective mix of the OpenMP and MPI parallel programming paradigms. This article reports our experiences in re-designing Splotch for exploiting emerging HPC architectures nowadays increasingly populated with GPUs. A performance model is introduced to guide our re-factoring of Splotch. A number of parallelization issues are discussed, in particular relating to race conditions and workload balancing, towards achieving optimal performances. Our implementation was accomplished by using the CUDA programming paradigm. Our strategy is founded on novel schemes achieving optimised data organisation and classification of particles. We deploy a reference cosmological simulation to present performance results on acceleration gains and scalability. We finally outline our vision for future work developments including possibilities for further optimisations and exploitation of hybrid systems and emerging accelerators.
    Astronomy and Computing. 04/2014; 5:9-18.
  • Source
    Franco Vazza, Claudio Gheller, Marcus Brüggen
    [Show abstract] [Hide abstract]
    ABSTRACT: Non-thermal (relativistic) particles are injected into the cosmos by structure formation shock waves, active galactic nuclei and stellar explosions. We present a suite of unigrid cosmological simulations (up to $2048^3$) using a two-fluid model in the grid code ENZO. The simulations include the dynamical effects of cosmic-ray (CR) protons and cover a range of theoretically motivated acceleration efficiencies. For the bulk of the cosmic volume the modelling of CR processes is rather stable with respect to resolution, provided that a minimum (cell) resolution of $\approx 100 ~\rm kpc/h$ is employed. However, the results for the innermost cluster regions depend on the assumptions for the baryonic physics. Inside clusters, non-radiative runs at high resolution tend to produce an energy density of CRs that are below available upper limits from the FERMI satellite, while the radiative runs are found to produce a higher budget of CRs. We show that weak ($M \leq 3-5$) shocks and shock-reacceleration are crucial to set the level of CRs in the innermost region of clusters, while in the outer regions the level of CR energy is mainly set via direct injection by stronger shocks, and is less sensitive to cooling and feedback from active galactic nuclei and supernovae.
    Monthly Notices of the Royal Astronomical Society 01/2014; 439(3). · 5.23 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The management and analysis of modern large-scale datasets generated by scientific experiments (observations or numerical simulations) can be very challenging due to continuously increasing sizes and complexity. Traditional data mining and analysis methods can be expensive if applied to large-scale datasets. Visual exploration and discovery can represent invaluable tools providing scientists with prompt and intuitive insights, enabling them to identify characteristics and manageable regions of interest for applying time consuming methods. They can also be effective means for communicating scientific results not only to researchers but also to members of the general public. Splotch [1] is a rendering algorithm for large-scale datasets exploiting modern High Performance Computing (HPC) architectures. Parallelism on HPC systems is achieved by deploying the MPI and OpenMP paradigms [2]. As GPUs are rapidly gaining popularity within HPC communities, we developed a GPU accelerated implementation of Splotch. Our CUDA approach [3] is presented together with performance results for Fermi and Kepler. We discuss future plans to further improve performances by exploiting new CUDA features for Kepler.
    GPU Technology Conference, San Francisco, USA; 03/2013
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We present a GPU implementation of Splotch based on the CUDA programming paradigm. Splotch is an algorithm for visualization of large-scale datasets coming from astronomical observations or numerical simula tions designed to exploit effectively High Performance Computing (HPC) systems. We describe the main steps required to adapt the original Splotch to GPU architectures and discuss results on a number of tests and benchmarks.
    Astronomical Society of the Pacific Conference Series. 01/2013; 475(Astronomical Data Analysis Software and Systems XXII):103.
  • Source
    F. Vazza, M. Brueggen, C. Gheller
    [Show abstract] [Hide abstract]
    ABSTRACT: We investigate the observable effects of feedback from Active Galactic Nuclei (AGN) on non-thermal components of the intracluster medium (ICM). We have modelled feedback from AGN in cosmological simulations with the adaptive mesh refinement code ENZO, investigating three types of feedback that are sometimes called quasar, jet and radio mode. Using a small set of galaxy clusters simulated at high resolution, we model the injection and evolution of Cosmic Rays, as well as their effects on the thermal plasma. By comparing, both, the profiles of thermal gas to observed profiles from the ACCEPT sample, and the secondary gamma-ray emission to the available upper limits from FERMI, we discuss how the combined analysis of these two observables can constrain the energetics and mechanisms of feedback models in clusters. Those modes of AGN feedback that provide a good match to X-ray observations, yield a gamma-ray luminosity resulting from secondary cosmic rays that is about below the available upper limits from FERMI. Moreover, we investigate the injection of turbulent motions into the ICM from AGN, and the detectability of these motions via the analysis of line broadening of the Fe XXIII line. In the near future, deeper observations/upper-limits of non-thermal emissions from galaxy clusters will yield stringent constraints on the energetics and modes of AGN feedback, even at early cosmic epochs.
    Monthly Notices of the Royal Astronomical Society 10/2012; 428(3). · 5.23 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We present a numerical scheme, implemented in the cosmological adaptive mesh refinement code enzo, to model the injection of cosmic ray (CR) particles at shocks, their advection and their dynamical feedback on thermal baryonic gas. We give a description of the algorithms and show their tests against analytical and idealized one-dimensional problems. Our implementation is able to track the injection of CR energy, the spatial advection of CR energy and its feedback on the thermal gas in run-time. This method is applied to study CR acceleration and evolution in cosmological volumes, with both fixed and variable mesh resolution. We compare the properties of galaxy clusters with and without CRs for a sample of high-resolution clusters with different dynamical states. At variance with similar simulations based on smoothed particles hydrodynamics, we report that the inclusion of CR feedback in our method decreases the central gas density in clusters, thus reducing the X-ray and Sunyaev–Zeldovich effect from the clusters centre, while enhancing the gas density and its related observables near the virial radius.
    Monthly Notices of the Royal Astronomical Society 04/2012; 421(4):3375 - 3398. · 5.52 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We present the numerical implementation of run-time injection of Cosmic Rays energy, their spatial advection and their dynamical feedack on baryonic gas in the cosmological grid code ENZO. We discuss the results of its application to large-scale simulations showing that the CR energy inside clusters of galaxies is small compared to the gas energy (less than a few percent), while the ratio is larger near the accretion regions of clusters and filaments (about 0.1-0.3). CR feedback has a small, but significant impact on the X-ray emission and Sunyaev-Zeldovich effect from clusters.
    09/2011;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This article presents a newly developed Web portal called VisIVOWeb that aims to provide the astrophysical community with powerful visualization tools for large-scale data sets in the context of Web 2.0. VisIVOWeb can effectively handle modern numerical simulations and real-world observations. Our open-source software is based on established visualization toolkits offering high-quality rendering algorithms. The underlying data management is discussed with the supported visualization interfaces and movie-making functionality. We introduce VisIVOWeb Network, a robust network of customized Web portals for visual discovery, and VisIVOWeb Connect, a lightweight and efficient solution for seamlessly connecting to existing astrophysical archives. A significant effort has been devoted for ensuring interoperability with existing tools by adhering to IVOA standards. We conclude with a summary of our work and a discussion on future developments.
    Publications of the Astronomical Society of the Pacific 07/2011; 123. · 3.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The aim of scientific visualization (images and animations) is to create suitable visuals (images and animations) to aid scientists in easily understanding highly-complex datasets. Modern visualization can act as a catalyst in rapidly and intuitively discovering correlations and patterns in large-scale astrophysical datasets without involving numerical algorithms or CPU intensive analysis codes. We introduce VisIVODesktop 3.0, a true multi-platform environment for interactive astrophysical visualization. We outline recently added visualization functionality, namely hybrid and stereoscopic rendering. Hybrid rendering combines volume and point rendering and employs an elliptical, Gaussian distribution function. Opacity transfer functions can be finely tuned to extract interesting details within cosmological structures. We conclude with a summary of our work and future developments.
    07/2011;
  • [Show abstract] [Hide abstract]
    ABSTRACT: Nowadays digital sky surveys and long-duration, high-resolution numerical simulations using high performance computing and grid systems produce multidimensional astrophysical datasets in the order of several Petabytes. Sharing visualizations of such datasets within communities and collaborating research groups is of paramount importance for disseminating results and advancing astrophysical research. Moreover educational and public outreach programs can benefit greatly from novel ways of presenting these datasets by promoting understanding of complex astrophysical processes, e.g., formation of stars and galaxies. We have previously developed VisIVO Server, a grid-enabled platform for high-performance large-scale astrophysical visualization. This article reviews the latest developments on VisIVO Web, a custom designed web portal wrapped around VisIVO Server, then introduces VisIVO Smartphone, a gateway connecting VisIVO Web and data repositories for mobile astrophysical visualization. We discuss current work and summarize future developments.
    07/2011;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: [...] We present results for the statistics of thermal gas and the shock wave properties for a large volume simulated with three different cosmological numerical codes: the Eulerian total variations diminishing code TVD, the Eulerian piecewise parabolic method-based code ENZO, and the Lagrangian smoothed-particle hydrodynamics code GADGET. Starting from a shared set of initial conditions, we present convergence tests for a cosmological volume of side-length 100 Mpc/h, studying in detail the morphological and statistical properties of the thermal gas as a function of mass and spatial resolution in all codes. By applying shock finding methods to each code, we measure the statistics of shock waves and the related cosmic ray acceleration efficiencies, within the sample of simulations and for the results of the different approaches. We discuss the regimes of uncertainties and disagreement among codes, with a particular focus on the results at the scale of galaxy clusters. Even if the bulk of thermal and shock properties are reasonably in agreement among the three codes, yet some significant differences exist (especially between Eulerian methods and smoothed particle hydrodynamics). In particular, we report: a) differences of huge factors (10-100) in the values of average gas density, temperature, entropy, Mach number and shock thermal energy flux in the most rarefied regions of the simulations between grid and SPH methods; b) the hint of an entropy core inside clusters simulated in grid codes; c) significantly different phase diagrams of shocked cells in grid codes compared to SPH; d) sizable differences in the morphologies of accretion shocks between grid and SPH methods.
    Monthly Notices of the Royal Astronomical Society 06/2011; 418(2). · 5.23 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Splotch is a light and fast, publicly available, ray-tracer software tool which supports the effective visualization of cosmological simulations data. The algorithm it relies on is designed to deal with point-like data, optimizing the ray-tracing calculation by ordering the particles as a function of their 'depth', defined as a function of one of the coordinates or other associated parameters. Realistic three-dimensional impressions are reached through a composition of the final colour in each pixel properly calculating emission and absorption of individual volume elements.
    Astrophysics Source Code Library. 03/2011;
  • Notiziario del Cineca. 01/2011;
  • [Show abstract] [Hide abstract]
    ABSTRACT: In the painstaking effort aimed at defining standards for the inclusion of theoretical data in the VO by testing the compliance of as many simulation services as possible, some updates delivered by the Italian Theoretical Virtual Observatory (ITVO) project are presented. They include cosmological simulated data archive s and services (at Trieste and Catania Astronomical Observatories and at the CINECA supercomputing centre in Bologna) and stellar simulations data archive s and services (BaSTI, a Bag of Stellar Tracks and Isochrones, maintained at the Teramo Astronomical Observatory). Following an upgrade in BaSTI database and a new Web service endpoint for VisIVOweb in Trieste some improvements in data visualization are presented: from new plots for BaSTI tracks emphasizing star evolution's key points to their comparison with observational data, from direct D visualization of galaxy clusters to their direct online data manipulation server side using the VO-compliant VisIVOWeb service. Finally we present a new unique Web portal for the ITVO project under VObs.it, the Italian effort in the VO world, that collects all information regarding the various tokens of this theoretical virtual observatory project that are spread all over Italy.
    12/2010;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We study the properties of chaotic motions in the intra cluster medium using a set of 20 galaxy clusters simulated with large dynamical range, using the Adaptive Mesh Refinement code ENZO (e.g. Norman et al.2007). The adopted setup allows us to study the spectral and spatial properties of turbulent motions in galaxy clusters with unprecedented detail, achieving an maximum available Reynolds number of the order of R=500-1000 for the largest eddies. The correlations between the energy of these motions in the Intra Cluster Medium and the dynamical state of the host systems are studied, and the statistical properties of turbulent motions and their evolution with time support that major merger events are responsible for the injection of the bulk of turbulent kinetic energy inside cluster. Turbulence is found to account for a 20-30 per cent of the thermal energy in merging clusters, while it accounts for a 5 per cent in relaxed clusters. A comparison of the energies of turbulence and motions in our simulated clusters with present upper-limits in real nearby clusters, recently derived with XMM-Newton (Sanders et al.2010), is provided. When the same spatial scales of turbulent motions are compared, the data from simulations result well within the range presently allowed by observations. Finally, we comment on the possibility that turbulence may accelerate relativistic particles leading to the formation of giant radio halos in turbulent (merging) clusters. Based on our simulations we confirm previous semi-analytical studies that suggest that the fraction of turbulent clusters is consistent with that of clusters hosting radio halos.
    Astronomy and Astrophysics 10/2010; 529. · 4.48 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Over the past few years, the role of visualization for scientific purpose has grown up enormously. Astronomy makes an extended use of visualization techniques to analyze data, and scientific visualization has became a fundamental part of modern researches in Astronomy. With the evolution of high performance computers, numerical simulations have assumed a great role in the scientific investigation, allowing the user to run simulation with higher and higher resolution. Data produced in these simulations are often multi-dimensional arrays with several physical quantities. These data are very hard to manage and to analyze efficiently. Consequently the data analysis and visualization tools must follow the new requirements of the research. AstroMD is a tool for data analysis and visualization of astrophysical data and can manage different physical quantities and multi-dimensional data sets. The tool uses virtual reality techniques by which the user has the impression of travelling through a computer-based multi-dimensional model.
    Astrophysics Source Code Library. 10/2010;