Rüdiger Nather

Alice Moallemy-Oureh, Silvia Beddar-Wiesing, Yannick Nagel et al.

Temporal Point Processes (TPPs) have recently become increasingly interesting for learning dynamics in graph data. A reason for this is that learning on dynamic graph data is becoming more relevant, since data from many scientific fields, ranging from mathematics, biology, social sciences, and physics to computer science, is naturally related and inherently dynamic. In addition, TPPs provide a meaningful characterization of event streams and a prediction mechanism for future events. Therefore, (semi-)parameterized Neural TPPs have been introduced whose characterization can be (partially) learned and, thus, enable the representation of more complex phenomena. However, the research on modeling dynamic graphs with TPPs is relatively young, and only a few models for node attribute changes or evolving edges have been proposed yet. To allow for learning on fully dynamic graph streams, i.e., graphs that can change in their structure (addition/deletion of nodes/edge) and in their node/edge attributes, we propose a Marked Neural Spatio-Temporal Point Process (MNSTPP). It leverages a Dynamic Graph Neural Network to learn a Marked TPP that handles attributes and spatial data to model and predict any event in a graph stream.

Rüdiger Nather, Claudia Fohry, Mia Reitz

As the size of computational problems grows, so does the likelihood of Silent Data Corruptions (SDCs). A common defense is replication, where the computation is repeated and correct results are determined by majority voting. Asynchronous Many-Task (AMT) runtimes are generally well suited for this approach, since the inputs and outputs of task replicas can be compared, and the tasks can be recomputed if necessary. Most existing SDC protection schemes assume static tasks and dependencies. Dynamic settings are more challenging, especially in clusters, since the tasks/data must be tracked for the comparisons. This paper considers a particularly dynamic setting with task spawning at runtime, task communication through C++11-like promises/futures, conditional touches, and cluster-wide load balancing via work-first work stealing. We propose an approach that closely couples original and replica computations by cross-validating all outgoing effects when interacting with the runtime system. The approach selectively recomputes affected tasks only. We implemented the approach in the ItoyoriFBC runtime system and conducted preliminary experiments with Fibonacci and emulated $\mathcal{H}$-matrix LU decomposition benchmarks. Results show a factor of less than two increase of failure-free running times, despite full replication, which is mainly due to improved opportunities for load balancing resulting from the higher number of tasks. The overhead for failure correction was about 0.5% of the overall running time per SDC.

Rüdiger Nather

As parallel workloads grow in complexity, managing fine-grained data dependencies becomes a critical challenge. Futures offer a promising model for handling these dependencies, particularly in irregular algorithms, but they also come with the restriction of value-immutability. This immutability limits the ability to perform in-place memory updates, a necessity for high-performance linear algebra where memory recycling is paramount. In this paper, we address these limitations by introducing a new construct, await_delete, which extends traditional future semantics to allow safe value reuse once consumers are finished. Building on this extension, we present a novel future-based algorithm for the block-wise inversion of dense, symmetric matrices, motivated by a recent algorithm for finding conservation laws of dynamical systems. We implement our approach in an extended version of Taskflow and evaluate it through strong-scaling experiments. Our results demonstrate that while futures incur significant overhead on smaller problem sizes, they achieve nearly linear scaling on large matrices. We analyze the amortization threshold and show that futures are a viable high-performance tool for large-scale linear algebra.