paper-with-me

Papers

Scaling Distributed Deep Learning Workloads beyond the Memory Capacity with KARMA

2020-08-26 · Mohamed Wahib, Haoyu Zhang, Truong Thao Nguyen, Aleksandr Drozd, Jens Domke, Lingqi Zhang, Ryousei Takano, Satoshi Matsuoka

The dedicated memory of hardware accelerators can be insufficient to store all weights and/or intermediate states of large deep learning models. Although model parallelism is a viable approach to reduce the memory pressure issue, significant modification of the source code and considerations for algorithms are required. An alternative solution is to use out-of-core methods instead of, or in addition to, data parallelism. We propose a performance model based on the concurrency analysis of out-of-core training behavior, and derive a strategy that combines layer swapping and redundant recomputing. We achieve an average of 1.52x speedup in six different models over the state-of-the-art out-of-core methods. We also introduce the first method to solve the challenging problem of out-of-core multi-node training by carefully pipelining gradient exchanges and performing the parameter updates on the host. Our data parallel out-of-core solution can outperform complex hybrid model parallelism in training large models, e.g. Megatron-LM and Turning-NLG.

📄 PDF Abstract BibTeX arXiv:2008.11421

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Characterizing the Efficiency of Distributed Training: A Power, Performance, and Thermal Perspective

2025-09-12 · Seokjin Go, Joongun Park, Spandan More, Hanjiang Wu 외 arxiv

The rapid scaling of Large Language Models (LLMs) has pushed training workloads far beyond the limits of single-node analysis, demanding a deeper understanding of how these models behave across large-scale, multi-GPU sys…

Analysis and Optimization of GNN-Based Recommender Systems on Persistent Memory

2022-07-25 · Yuwei Hu, Jiajie Li, Zhongming Yu, Zhiru Zhang

Graph neural networks (GNNs), which have emerged as an effective method for handling machine learning tasks on graphs, bring a new approach to building recommender systems, where the task of recommendation can be formula…

Link PredictionRecommendation Systems

FloatSOM: GPU-Accelerated, Distributed, Topology-Flexible Self-Organizing Maps

2026-04-29 · Tony Xu, Sarah Klamt, Katherine Turner, Anne Brustle 외 arxiv

GPU-accelerated Self-Organizing Map (SOM) implementations are among the most competitive options for large-scale SOM analysis, but growing dataset sizes increasingly challenge their practical use because workloads no lon…

G10: Enabling An Efficient Unified GPU Memory and Storage Architecture with Smart Tensor Migrations

2023-10-13 · Haoyang Zhang, Yirui Eric Zhou, Yuqi Xue, Yiqi Liu 외

To break the GPU memory wall for scaling deep learning workloads, a variety of architecture and system techniques have been proposed recently. Their typical approaches include memory extension with flash memory and direc…

Deep LearningGPU

OasisKV: Scaling In-Decode KV Cache Beyond HBM with Lookahead Sparse Prefetching

2026-08-08 · Can Xiao, Sukmin Cho, Junbong We, Zhixiong Niu 외 hf

Large language model (LLM) inference serving is increasingly constrained by memory rather than compute. As long-context and long-form reasoning workloads become more prevalent, the key-value (KV) cache dominates both mem…