Related papers: Lossy Bulk Synchronous Parallel Processing Model f…
We consider unreliable multi-hop networks serving multiple flows in which packets not delivered to their destination nodes by their deadlines are dropped. We address the design of policies for routing and scheduling packets that optimize…
The last decades have seen a surge of interests in distributed computing thanks to advances in clustered computing and big data technology. Existing distributed algorithms typically assume {\it all the data are already in one place}, and…
Cloud computing has grown rapidly in recent years, mainly due to the sharp increase in data transferred over the internet. This growth makes load balancing a key part of cloud systems, as it helps distribute user requests across servers to…
Large-scale diffusion models such as FLUX (12B parameters) and Stable Diffusion 3 (8B parameters) require multi-GPU parallelism for efficient inference. Unified Sequence Parallelism (USP), which combines Ulysses and Ring attention…
The multistage robust unit commitment (UC) is of paramount importance for achieving reliable operations considering the uncertainty of renewable realizations. The typical affine decision rule method and the robust feasible region method may…
Large-scale models rely heavily on 3D parallelism for distributed training, which utilizes tensor parallelism (TP) as the intra-operator parallelism to partition model states across GPUs. However, TP introduces significant communication…
We consider the problem of sampling $n$ numbers from the range $\{1,\ldots,N\}$ without replacement on modern architectures. The main result is a simple divide-and-conquer scheme that makes sequential algorithms more cache efficient and…
Multi-server jobs are imperative in modern cloud computing systems. A noteworthy feature of multi-server jobs is that, they usually request multiple computing devices simultaneously for their execution. How to schedule multi-server jobs…
In this paper we study the design issues in improving TCP performance over the ATM UBR service. ATM-UBR switches respond to congestion by dropping cells when their buffers become full. TCP connections running over UBR can experience low…
BitTorrent, one of the most widespread used P2P application for file-sharing, recently got rid of TCP by introducing an application-level congestion control protocol named uTP. The aim of this new protocol is to efficiently use the…
Internet performance is tightly related to the properties of TCP and UDP protocols, jointly responsible for the delivery of the great majority of Internet traffic. It is well understood how these protocols behave under FIFO queuing and what…
Consider a system in which tasks of different execution times arrive continuously and have to be executed by a set of processors that are prone to crashes and restarts. In this paper we model and study the impact of parallelism and failures…
To train modern large DNN models, pipeline parallelism has recently emerged, which distributes the model across GPUs and enables different devices to process different microbatches in pipeline. Earlier pipeline designs allow multiple…
It is a challenging task to train large DNN models on sophisticated GPU platforms with diversified interconnect capabilities. Recently, pipelined training has been proposed as an effective approach for improving device utilization. However,…
Fault tolerance overhead of high performance computing (HPC) applications is becoming critical to the efficient utilization of HPC systems at large scale. HPC applications typically tolerate fail-stop failures by checkpointing. Another…
An online backup system should be quick and reliable in both saving and restoring users' data. To do so in a peer-to-peer implementation, data transfer scheduling and the amount of redundancy must be chosen wisely. We formalize the problem…
Modern mobile and stationary devices are equipped with multiple network interfaces aiming to provide wireless and wireline connectivity either in a local LAN or the Internet. Multipath TCP (MPTCP) protocol has been developed on top of…
Belief propagation and its variants are popular methods for approximate inference, but their running time and even their convergence depend greatly on the schedule used to send the messages. Recently, dynamic update schedules have been…
Most machine learning and deep neural network algorithms rely on certain iterative algorithms to optimise their utility/cost functions, e.g. Stochastic Gradient Descent. In distributed learning, the networked nodes have to work…
HPC systems keep growing in size to meet the ever-increasing demand for performance and computational resources. Apart from increased performance, large scale systems face two challenges that hinder further growth: energy efficiency and…