arXiv subjects

Marion Weinzierl

Publications and source records attributed to Marion Weinzierl.

3 recordsLinked to original sources

arXiv2022

The Future of Hackathon Research and Practice

Hackathons are time-bounded collaborative events which have become a global phenomenon adopted by both researchers and practitioners in a plethora of contexts. Hackathon events are generally used to accelerate the development of, for example, scientific results and collaborations, communities, and innovative prototypes addressing urgent challenges. As hackathons have been adopted into many different contexts, the events have also been adapted in numerous ways corresponding to the unique needs and situations of organizers, participants and other stakeholders. While these interdisciplinary adaptions, in general affords many advantages - such as tailoring the format to specific needs - they also entail certain challenges, specifically: 1) limited exchange of best practices, 2) limited exchange of research findings, and 3) larger overarching questions that require interdisciplinary collaboration are not discovered and remain unaddressed. We call for interdisciplinary collaborations to address these challenges. As a first initiative towards this, we performed an interdisciplinary collaborative analysis in the context of a workshop at the Lorentz Center, Leiden in December 2021. In this paper, we present the results of this analysis in terms of six important areas which we envision to contribute to maturing hackathon research and practice: 1) hackathons for different purposes, 2) socio-technical event design, 3) scaling up, 4) making hackathons equitable, 5) studying hackathons, and 6) hackathon goals and how to reach them. We present these areas in terms of the state of the art and research proposals and conclude the paper by suggesting next steps needed for advancing hackathon research and practice.

cs.HC↗

arXiv2017

Impact of Non-potential Coronal Boundary Conditions on Solar Wind Prediction

Predictions of the solar wind at Earth are a central aspect of space weather prediction. The outcome of such a prediction, however, is highly sensitive to the method used for computing the magnetic field in the corona. We analyze the impact of replacing the potential field coronal boundary conditions, as used in operational space weather prediction tools, by non-potential conditions. For this, we compare the predicted solar wind plasma parameters with observations at 1 AU for two six-months intervals, one at solar maximum and one in the descending phase of the current cycle. As a baseline, we compare with the operational Wang-Sheeley-Arge model. We find that for solar maximum, the non-potential coronal model and an adapted solar wind speed formula lead to the best solar wind predictions in a statistical sense. For the descending phase, the potential coronal model performs best. The Wang-Sheeley-Arge model outperforms the others in predicting high speed enhancements and streamer interactions. A better parameter fitting for the adapted wind speed formula is expected to improve the performance of the non-potential model here.

astro-ph.SR↗

arXiv2017

Quasi-matrix-free hybrid multigrid on dynamically adaptive Cartesian grids

We present a family of spacetree-based multigrid realizations using the tree's multiscale nature to derive coarse grids. They align with matrix-free geometric multigrid solvers as they never assemble the system matrices which is cumbersome for dynamically adaptive grids and full multigrid. The most sophisticated realizations use BoxMG to construct operator-dependent prolongation and restriction in combination with Galerkin/Petrov-Galerkin coarse-grid operators. This yields robust solvers for nontrivial elliptic problems. We embed the algebraic, problem- and grid-dependent multigrid operators as stencils into the grid and evaluate all matrix-vector products in-situ throughout the grid traversals. While such an approach is not literally matrix-free---the grid carries the matrix---we propose to switch to a hierarchical representation of all operators. Only differences of algebraic operators to their geometric counterparts are held. These hierarchical differences can be stored and exchanged with small memory footprint. Our realizations support arbitrary dynamically adaptive grids while they vertically integrate the multilevel operations through spacetree linearization. This yields good memory access characteristics, while standard colouring of mesh entities with domain decomposition allows us to use parallel manycore clusters. All realization ingredients are detailed such that they can be used by other codes.

math.NA↗