No abstract

Driven by growing application requirements and accelerated by current trends in microprocessor design, the number of processor cores on modern supercomputers is increasing from generation to generation. However, load or communication imbalance prevents many codes from taking advantage of the available parallelism, as delays of single processes may spread wait states across the entire machine. Moreover, when employing complex point-to-point communication patterns, wait states may propagate along far-reaching cause-effect chains that are hard to track manually and that complicate an assessment of the actual costs of an imbalance. Building on earlier work by Meira Jr. et al., we present a scalable approach that identifies program wait states and attributes their costs in terms of resource waste to their original cause. By replaying event traces in parallel both in forward and backward direction, we can identify the processes and call paths responsible for the most severe imbalances even for runs with tens of thousands of processes.

The formation of a thin current sheet in a magnetic quasiseparatrix layer ͑QSL͒ is investigated by means of numerical simulation using a simplified ideal, low-␤, MHD model. The initial configuration and driving boundary conditions are relevant to phenomena observed in the solar corona and were studied earlier by Aulanier et al. ͓Astron. Astrophys. 444, 961 ͑2005͔͒. In extension to that work, we use the technique of adaptive mesh refinement ͑AMR͒ to significantly enhance the local spatial resolution of the current sheet during its formation, which enables us to follow the evolution into a later stage. Our simulations are in good agreement with the results of Aulanier et al. up to the calculated time in that work. In a later phase, we observe a basically unarrested collapse of the sheet to length scales that are more than one order of magnitude smaller than those reported earlier. The current density attains correspondingly larger maximum values within the sheet. During this thinning process, which is finally limited by lack of resolution even in the AMR studies, the current sheet moves upward, following a global expansion of the magnetic structure during the quasistatic evolution. The sheet is locally one-dimensional and the plasma flow in its vicinity, when transformed into a comoving frame, qualitatively resembles a stagnation point flow. In conclusion, our simulations support the idea that extremely high current densities are generated in the vicinities of QSLs as a response to external perturbations, with no sign of saturation.

The recent developments in mesh-free plasma modeling using parallel tree codes are described, covering the algorithmic and performance issues and how to apply this technique to multidimensional electrostatic plasma problems. Examples of the simulations of the ion acceleration by high-intensity lasers, heating, and the dynamics of the nanostructured targets, as well as more recent applications of this technique to the simulations of edge plasmas in tokamaks and mesh-free magnetoinductive models, are given.

Three-dimensional, time-dependent numerical simulations of the dynamics of magnetic flux ropes are presented. The simulations are targeted towards an experiment previously conducted at CalTech (Bellan, P. M. and J. F. Hansen, Phys. Plasmas, 5, 1991Plasmas, 5, (1998) which aimed at simulating Solar prominence eruptions in the laboratory. The plasma dynamics is described by ideal MHD using different models for the evolution of the mass density. Key features of the reported experimental observations like pinching of the current loop, its expansion and distortion into helical shape are reproduced in the numerical simulations. Details of the final structure depend on the choice of a specific model for the mass density.