BOOST: BOttleneck-Optimized Scalable Training Framework for Low-Rank Large Language Models

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Wang, Zhengyang, Liu, Ziyue, Zhang, Ruijie, Maurya, Avinash, Hovland, Paul, Nicolae, Bogdan, Cappello, Franck, Zhang, Zheng
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917481411510272
author Wang, Zhengyang
Liu, Ziyue
Zhang, Ruijie
Maurya, Avinash
Hovland, Paul
Nicolae, Bogdan
Cappello, Franck
Zhang, Zheng
author_facet Wang, Zhengyang
Liu, Ziyue
Zhang, Ruijie
Maurya, Avinash
Hovland, Paul
Nicolae, Bogdan
Cappello, Franck
Zhang, Zheng
contents The scale of transformer model pre-training is constrained by the increasing computation and communication cost. Low-rank bottleneck architectures offer a promising solution to significantly reduce the training time and memory footprint with minimum impact on accuracy. Despite algorithmic efficiency, bottleneck architectures scale poorly under standard tensor parallelism. Simply applying 3D parallelism designed for full-rank methods leads to excessive communication and poor GPU utilization. To address this limitation, we propose BOOST, an efficient training framework tailored for large-scale low-rank bottleneck architectures. BOOST introduces a novel Bottleneck-aware Tensor Parallelism, and combines optimizations such as online-RMSNorm, linear layer grouping, and low-rank activation checkpointing to achieve end-to-end training speedup. Evaluations on different low-rank bottleneck architectures demonstrate that BOOST achieves 1.46-1.91$\times$ speedup over full-rank model baselines and 1.87-2.27$\times$ speedup over low-rank model with naively integrated 3D parallelism, with improved GPU utilization and reduced communication overhead.
format Preprint
id arxiv_https___arxiv_org_abs_2512_12131
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle BOOST: BOttleneck-Optimized Scalable Training Framework for Low-Rank Large Language Models
Wang, Zhengyang
Liu, Ziyue
Zhang, Ruijie
Maurya, Avinash
Hovland, Paul
Nicolae, Bogdan
Cappello, Franck
Zhang, Zheng
Machine Learning
Distributed, Parallel, and Cluster Computing
The scale of transformer model pre-training is constrained by the increasing computation and communication cost. Low-rank bottleneck architectures offer a promising solution to significantly reduce the training time and memory footprint with minimum impact on accuracy. Despite algorithmic efficiency, bottleneck architectures scale poorly under standard tensor parallelism. Simply applying 3D parallelism designed for full-rank methods leads to excessive communication and poor GPU utilization. To address this limitation, we propose BOOST, an efficient training framework tailored for large-scale low-rank bottleneck architectures. BOOST introduces a novel Bottleneck-aware Tensor Parallelism, and combines optimizations such as online-RMSNorm, linear layer grouping, and low-rank activation checkpointing to achieve end-to-end training speedup. Evaluations on different low-rank bottleneck architectures demonstrate that BOOST achieves 1.46-1.91$\times$ speedup over full-rank model baselines and 1.87-2.27$\times$ speedup over low-rank model with naively integrated 3D parallelism, with improved GPU utilization and reduced communication overhead.
title BOOST: BOttleneck-Optimized Scalable Training Framework for Low-Rank Large Language Models
topic Machine Learning
Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2512.12131