paper-with-me

홈 › Papers

Towards a Flexible and High-Fidelity Approach to Distributed DNN Training Emulation

2024-05-05 · Banruo Liu, Mubarak Adetunji Ojewale, Yuhan Ding, Marco Canini

We propose NeuronaBox, a flexible, user-friendly, and high-fidelity approach to emulate DNN training workloads. We argue that to accurately observe performance, it is possible to execute the training workload on a subset of real nodes and emulate the networked execution environment along with the collective communication operations. Initial results from a proof-of-concept implementation show that NeuronaBox replicates the behavior of actual systems with high accuracy, with an error margin of less than 1% between the emulated measurements and the real system.

📄 PDF Abstract BibTeX arXiv:2405.02969

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

A Latent Representation Learning Framework for Hyperspectral Image Emulation in Remote Sensing

2026-03-23 · Chedly Ben Azizi, Claire Guilloteau, Gilles Roussel, Matthieu Puigt arxiv

Synthetic hyperspectral image (HSI) generation is essential for large-scale simulation, algorithm development, and mission design, yet traditional radiative transfer models remain computationally expensive and proposed e…

Representation Learning

A Distributed Emulation Environment for In-Memory Computing Systems

2025-10-09 · Eleni Bougioukou, Anastasios Petropoulos, Nikolaos Toulgaridis, Theodoros Chatzimichail 외 arxiv

In-memory computing technology is used extensively in artificial intelligence devices due to lower power consumption and fast calculation of matrix-based functions. The development of such a device and its integration in…

Real-time Digital RF Emulation -- II: A Near Memory Custom Accelerator

2024-06-13 · Mandovi Mukherjee, Xiangyu Mao, Nael Rahman, Coleman DeLude 외

A near memory hardware accelerator, based on a novel direct path computational model, for real-time emulation of radio frequency systems is demonstrated. Our evaluation of hardware performance uses both application-speci…

On the Value of Tokeniser Pretraining in Physics Foundation Models

2026-03-05 · Hadi Sotoudeh, Payel Mukhopadhyay, Ruben Ohana, Michael McCabe 외 arxiv

We investigate the impact of tokeniser pretraining on the accuracy and efficiency of physics emulation. Modern high-resolution simulations produce vast volumes of data spanning diverse physical regimes and scales. Traini…

Computational Efficiency

Emulation of physical processes with Emukit

2021-10-25 · Andrei Paleyes, Mark Pullin, Maren Mahsereci, Cliff McCollum 외

Decision making in uncertain scenarios is an ubiquitous challenge in real world systems. Tools to deal with this challenge include simulations to gather information and statistical emulation to quantify uncertainty. The …

Bayesian OptimizationDecision MakingDecision Making Under UncertaintyExperimental Design