Related papers: Throughput-Optimized Networks at Scale

Modern reasoning models, such as OpenAI's o1 and DeepSeek-R1, exhibit impressive problem-solving capabilities but suffer from critical inefficiencies: high inference latency, excessive computational resource consumption, and a tendency…

We study the problem of counting the number of nodes in a slotted-time communication network, under the challenging assumption that nodes do not have identifiers and the network topology changes frequently. That is, for each time slot links…

Task sequencing (TS) is one of the core open problems in Deep Learning, arising in a plethora of real-world domains, from robotic assembly lines to autonomous driving. Unfortunately, prior work has not convincingly demonstrated the…

Tensor processing units (TPUs) are one of the most well-known machine learning (ML) accelerators utilized at large scale in data centers as well as in tiny ML applications. TPUs offer several improvements and advantages over conventional ML…

In the realm of self-supervised learning (SSL), conventional wisdom has gravitated towards the utility of massive, general domain datasets for pretraining robust backbones. In this paper, we challenge this idea by exploring if it is…

Artificial Neural Networks (ANNs) require significant amounts of data and computational resources to achieve high effectiveness in performing the tasks for which they are trained. To reduce resource demands, various techniques, such as…

Contour trees offer an abstract representation of the level set topology in scalar fields and are widely used in topological data analysis and visualization. However, applying contour trees to large-scale scientific datasets remains…

Faster pathfinding in time-dependent transport networks is an important and challenging problem in navigation systems. There are two main types of transport networks: road networks for car driving and public transport route network. The…

Pre-training of Large Language Models is often prohibitively expensive and inefficient at scale, requiring complex and invasive modifications in order to achieve high data throughput. In this work, we present Token-Superposition Training…

A low-power precision-scalable processor for ConvNets or convolutional neural networks (CNN) is implemented in a 40nm technology. Its 256 parallel processing units achieve a peak 102GOPS running at 204MHz. To minimize energy consumption…

We propose the CLEX supercomputer topology and routing scheme. We prove that CLEX can utilize a constant fraction of the total bandwidth for point-to-point communication, at delays proportional to the sum of the number of intermediate hops…

Recurrent Neural Networks (RNNs) have become the standard modeling technique for sequence data, and are used in a number of novel text-to-speech models. However, training a TTS model including RNN components has certain requirements for GPU…

With technology scaling down, hundreds and thousands processing elements (PEs) can be integrated on a single chip. Network-on-chip (NoC) has been proposed as an efficient solution to handle this distinctive challenge. In this thesis, we…

In recent years, Graph Neural Networks (GNNs) have shown superior performance on diverse real-world applications. To improve the model capacity, besides designing aggregation operations, GNN topology design is also very important. In…

In this paper, we present a new MTL framework that searches for structures optimized for multiple tasks with diverse graph topologies and shares features among tasks. We design a restricted DAG-based central network with read-in/read-out…

Temporal Neural Networks (TNNs) use time as a resource to represent and process information, mimicking the behavior of the mammalian neocortex. This work focuses on implementing TNNs using off-the-shelf digital CMOS technology. A…

Time-Sensitive Networking (TSN) is an enhancement of Ethernet which provides various mechanisms for real-time communication. Time-triggered (TT) traffic represents periodic data streams with strict real-time requirements. Amongst others,…

Direct numerical simulations (DNS) are an indispensable tool for understanding the fundamental physics of turbulent flows. Because of their steep increase in computational cost with Reynolds number ($R_{\lambda}$), well-resolved DNS are…

Most real-world networks are embedded in latent geometries. If a node in a network is found in the vicinity of another node in the latent geometry, the two nodes have a disproportionately high probability of being connected by a link. The…

All-to-All communication is a key performance bottleneck for distributed machine learning (ML) and high-performance computing (HPC) workloads, where dense traffic increasingly stresses scale-up interconnects. While these ML and HPC…