Results 21 to 30 of about 1,316 (216)

Integrating parallelism and asynchrony for high-performance software development [PDF]

open access: yesE3S Web of Conferences, 2023
This article delves into the crucial roles of parallelism and asynchrony in the development of high-performance software programs. It provides an insightful exploration into how these methodologies enhance computing systems' efficiency and performance ...
Zaripova Rimma   +2 more
doaj   +1 more source

Synchronous Distributed Training With Runtime-Adaptive Mechanisms in Hybrid Cloud Environments

open access: yesIEEE Access
The rapid growth of large datasets and AI models has necessitated scalable and powerful computing resources, driving the extension of training workloads to the cloud to balance cost and performance. Among distributed learning strategies, synchronous data-
Tuan Anh Vuong   +3 more
doaj   +1 more source

Collective Communication Performance Evaluation for Distributed Deep Learning Training

open access: yesApplied Sciences
In distributed deep learning, the improper use of the collective communication library can lead to a decline in deep learning performance due to increased communication time.
Sookwang Lee, Jaehwan Lee
doaj   +1 more source

Dynamically Changing Parallelism with the Asynchronous Sequential Data Flows

open access: yesМоделирование и анализ информационных систем, 2020
A statically typed version of the data driven functional parallel computing model is proposed. It enables a representation of dynamically changing parallelism by means of asynchronous serial data flows.
Alexander I. Legalov   +3 more
doaj   +1 more source

A snapshot of parallelism in distributed deep learning training

open access: yesRevista Colombiana de Computación
The accelerated development of applications related to artificial intelligence has generated the creation of increasingly complex neural network models with enormous amounts of parameters, currently reaching up to trillions of parameters.
Hairol Romero-Sandí   +2 more
doaj   +5 more sources

Impact of Design Decisions on Performance of Embarrassingly Parallel .NET Database Application

open access: yesVietnam Journal of Computer Science
The implementation of parallel applications is always a challenge. It embraces many distinctive design decisions that are to be taken. The paper presents issues of parallel processing with use of .NET applications and popular Database Management Systems (
Piotr Karwaczyński   +6 more
doaj   +1 more source

Parallelism Analysis of Subroutine-Level Speculative in HPEC [PDF]

open access: yesJisuanji gongcheng, 2020
Effective application of Thread-Level Speculation(TLS) technology can improve the hardware resource utilization of multicore chips,and has acquired successful results in automatic parallelization of multiple serial applications.However,it lacks efficient
WANG Xinyi, WANG Yaobin, LI Ling, YANG Yang, BU Deqing, LIU Zhiqin
doaj   +1 more source

Distinct Microstructural Characteristic Lengths Defining Notch Fatigue Crack Initiation and Propagation, and Defect Tolerance in Advanced Steels

open access: yesAdvanced Engineering Materials, EarlyView.
Schematic representation of modes of microcrack nucleation, growth and temporary arrestment at different boundaries, with the criteria for crack propagation in three scenarios: (1) large notch, (2) defect/small sharp notch, and (3) long crack. This work revisits and integrates results on the fatigue behavior of advanced bainitic steels (in particular ...
Lucia Morales‐Rivas
wiley   +1 more source

A Parallelised ROOT for Future HEP Data Processing [PDF]

open access: yesEPJ Web of Conferences, 2019
In the coming years, HEP data processing will need to exploit parallelism on present and future hardware resources to sustain the bandwidth requirements. As one of the cornerstones of the HEP software ecosystem, ROOT embraced an ambitious parallelisation
Piparo Danilo   +6 more
doaj   +1 more source

Model Parallelism With Subnetwork Data Parallelism

open access: yesCoRR
Pre-training large neural networks at scale imposes heavy memory demands on accelerators and often requires costly communication. We introduce Subnetwork Data Parallelism (SDP), a distributed training framework that partitions a model into structured subnetworks trained across workers without exchanging activations.
Singh, Vaibhav   +4 more
openaire   +3 more sources

Home - About - Disclaimer - Privacy