Efficient routing in Dragonfly networks has proven to be challenging. Studies have shown that the current state-of-the-art Universal Globally Adaptive Routing (UGAL) for the Dragonfly network makes sub-optimal routing decisions in some situations when the decisions are made using local information. The advent of Software-Defined Networking (SDN) offers new opportunities for optimizing UGAL routing by providing a global network view. In this work, we study the potential of leveraging SDN technology to improve UGAL routing on Dragonfly networks. In particular, we develop techniques that utilize the global flow information available to an SDN network to improve the latency estimation and path selection in UGAL, which leads to better routing performance. We performed extensive simulation with synthetic traffic and application workloads, and concluded that by incorporating the global flow information in SDN, the performance of UGAL can be significantly improved.

New frontiers in simulation-based teacher training have been unveiled with the advancement of artificial intelligence (AI). Integrating AI into virtual student agents increases the accessibility and affordability of teacher training simulations, but little is known about how preservice teachers interact with AI-powered student agents. This study analyzed the discourse behavior of 15 preservice teachers who undertook simulation-based training with AI-powered student agents. Using a framework of ambitious science teaching, we conducted a pattern analysis of teacher and student talk moves, looking for evidence of academically productive discourse. Comparisons are made with patterns found in real classrooms with professionally trained science teachers. Results indicated that preservice teachers generated academically productive discourse with AI-powered students by using ambitious talk moves. The pattern analysis also revealed coachable moments where preservice teachers succumbed to cycles of unproductive discourse. This study highlights the utility of analyzing classroom discourse to understand human-AI communication in simulation-based teacher training.

Compute nodes on high performance computing (HPC) platforms are increasingly equipped with multiple GPUs. This results in increased computational capacity per node, and reduction in the total number of nodes or endpoints in the system. This trend changes the computation and communication balance in comparison to pre-GPU era HPC platforms, which warrants a new study of hardware architectural parameters. In this work, we leverage the end-to-end system simulation capabilities of TraceR-CODES and study the impact of several hardware design parameters on the performance of realistic HPC workloads. We focus on three crucial hardware parameters: (1) number of GPUs per node, (2) network link bandwidth, and (3) network interface controller (NIC) scheduling policies, in the context of two popular network topologies – fat-tree and dragonfly.

The Dragonfly network organizes routers into groups, with connectivity within each group provided by local links and connectivity between groups provided by global links. The specification of Dragonfly leaves many options for arranging global links. In this work, we study global link arrangement for practical Dragonfly topologies where (1) there are multiple global links connecting each pair of groups, and (2) the global link bandwidth is similar to the local link bandwidth. We found that existing global link arrangement schemes such as the absolute, relative and circulant-based arrangements do not specify an important component in global connectivity for practical Dragonfly, which we call per-router arrangement. Per-router arrangement determines how the global links from each individual router are connected. We integrate per-router arrangement into existing schemes, develop a unified algorithm to compute a large class of global link arrangements for practical Dragonfly, and carry out an extensive simulation study to evaluate different global link arrangement schemes. Our results indicate that the existing understanding of the global link arrangement does not apply to practical Dragonfly: contradict to the existing understanding that global link arrangement does not make significant difference in performance when global links have similar bandwidth as local links, per-router arrangement significantly impacts the network performance of practical Dragonfly. We identify the schemes that yield high performance for practical Dragonfly.

The Dragonfly network has been deployed in the current generation supercomputers and will be used in the next generation supercomputers. The Universal Globally Adaptive Load-balance routing (UGAL) is the state-of-the-art routing scheme for Dragonfly. In this work, we show that the performance of the conventional UGAL can be further improved on many practical Dragonfly networks, especially the ones with a small number of groups, by customizing the paths used in UGAL for each topology. We develop a scheme to compute the custom sets of paths for each topology and compare the performance of our topology-custom UGAL routing (T-UGAL) with conventional UGAL. Our evaluation with different UGAL variations and different topologies demonstrates that by customizing the routes, T-UGAL offers significant improvements over UGAL on many practical Dragonfly networks in terms of both latency when the network is under low load and throughput when the network is under high load.