Related papers: Four Algorithms on the Swapped Dragonfly
This paper describes the Swapped Dragonfly. It is a two-parameter family of diameter three interconnection networks, D3(K,M), which are linearly scalable in M. Although D3(K,M) is a Dragonfly, it differs from standard Dragonflies in many…
High-radix interconnects such as Dragonfly and its variants rely on adaptive routing to balance network traffic for optimum performance. Ideally, adaptive routing attempts to forward packets between minimal and non-minimal paths with the…
Most deep neural networks (DNNs) consist fundamentally of convolutional and/or fully connected layers, wherein the linear transform can be cast as the product between a filter matrix and a data matrix obtained by arranging feature tensors…
Much like classical supercomputers, scaling up quantum computers requires an optical interconnect. However, signal attenuation leads to irreversible qubit loss, making quantum interconnect design guidelines and metrics different from…
In this paper, we study systems of distributed entities that can actively modify their communication network. This gives rise to distributed algorithms that apart from communication can also exploit network reconfiguration in order to carry…
Dragonfly interconnect is a crucial network technology for supercomputers. To support exascale systems, network resources are shared such that links and routers are not dedicated to any node pair. While link utilization is increased,…
We give both efficient algorithms and hardness results for reconfiguring between two connected configurations of modules in the hexagonal grid. The reconfiguration moves that we consider are "pivots", where a hexagonal module rotates around…
The diversity-multiplexing tradeoff of the dynamic decode-and-forward protocol is characterized for the half-duplex three-terminal (m,k,n)-relay channel where the source, relay and the destination terminals have m, k and n antennas,…
In this work the optimal diversity-multiplexing tradeoff (DMT) is investigated for the multiple-input multiple-output fading multiple-access channels with no power constraints (infinite constellations). For K users (K>1), M transmit…
We consider single-source single-sink (ss-ss) multi-hop networks, with slow-fading links and single-antenna half-duplex relays. We identify two families of networks that are multi-hop generalizations of the well-studied two-hop network:…
The Dragonfly topology is currently one of the most popular network topologies in high-performance parallel systems. The interconnection networks of many of these systems are built from components based on the InfiniBand specification.…
Transformers achieve state-of-the-art results across many tasks, but their uniform application of quadratic self-attention to every token at every layer makes them computationally expensive. We introduce DTRNet (Dynamic Token Routing…
Motivated by the need for robust and fast distributed computation in highly dynamic Peer-to-Peer (P2P) networks, we study algorithms for the fundamental distributed agreement problem. P2P networks are highly dynamic networks that experience…
DPillar has recently been proposed as a server-centric datacenter network and is combinatorially related to (but distinct from) the well-known wrapped butterfly network. We explain the relationship between DPillar and the wrapped butterfly…
We study the k-pair communication problem for quantum information in networks of quantum channels. We consider the asymptotic rates of high fidelity quantum communication between specific sender-receiver pairs. Four scenarios of classical…
We study multi-stream transmission in the $K \times N \times K$ channel with interfering relay nodes, consisting of $K$ multi-antenna source--destination (S--D) pairs and $N$ single-antenna half-duplex relay nodes between the S--D pairs. We…
This paper studies the setup of a multiple-relay network in which $K$ half-duplex multiple-antenna relays assist in the transmission between a/several multiple-antenna transmitter(s) and a multiple-antenna receiver. Each two nodes are…
Convolutional neural networks have become the main tools for processing two-dimensional data. They work well for images, yet convolutions have a limited receptive field that prevents its applications to more complex 2D tasks. We propose a…
In this paper, we consider networks with topologies described by some connected undirected graph ${\mathcal{G}}=(V, E)$ and with some agents (fusion centers) equipped with processing power and local peer-to-peer communication, and…
The scalability of quantum computing is currently limited by physical, technological, and architectural constraints that hinder the integration of a large number of qubits within a single quantum processor. Distributed quantum computing…