Search

Large-scale computing systems provide great potential for scientific exploration. However, the complexity that accompanies these enormous machines raises challenges for both, users and operators. The effective use of such systems is often hampered by failures encountered when running applications on systems containing tens-of-thousands of nodes and hundreds-of-thousands of compute cores capable of yielding petaflops of performance. In systems of this size failure detection is complicated and root-cause diagnosis difficult. This paper describes our recent work in the identification of anomalies in monitoring data and system logs to provide further insights into machine status, runtime behavior, failure modes and failure root causes. It discusses the details of an initial prototype that gathers the data and uses statistical techniques for analysis.

We define a computational model of physical devices that have a parallel atomic operation that transforms their input, an unordered list, in such a way that their output, the sorted list, can be sequentially read off in linear time. We show that several commonly-used scientific laboratory techniques (from biology, chemistry, and physics) are instances of the model and we provide experimental implementations.

In this paper, we analysis the effect of partial occlusions in scenes captured using digital holography. We reconstruct the scene from different perspectives. These reconstructions are then combined, allowing one to overcome foreground occlusions that are obscuring one’s view of the scene. The analysis in this paper is carried out with the aid of the Wigner distribution function, allowing us to visualize the energy of the object wavefield and the occluding object wavefield in phase space. We show that by iteratively selecting different views, the original scene can be reconstructed efficiently. This technique would be useful in situations where transmission of the whole digital hologram, or exhaustive reconstruction of every perspective, was not feasible. We provide results using optically captured digital holograms of real-world objects, and simulated occlusions.