Hetu v2: A General and Scalable Deep Learning System with Hierarchical and Heterogeneous Single Program Multiple Data Annotations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Haoyang, Fu, Fangcheng, Ge, Hao, Lin, Sheng, Wang, Xuanyu, Niu, Jiawen, Miao, Xupeng, Cui, Bin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918004098334720
author Li, Haoyang
Fu, Fangcheng
Ge, Hao
Lin, Sheng
Wang, Xuanyu
Niu, Jiawen
Miao, Xupeng
Cui, Bin
author_facet Li, Haoyang
Fu, Fangcheng
Ge, Hao
Lin, Sheng
Wang, Xuanyu
Niu, Jiawen
Miao, Xupeng
Cui, Bin
contents The Single Program Multiple Data (SPMD) paradigm provides a unified abstraction to annotate various parallel dimensions in distributed deep learning (DL) training. With SPMD, users can write training programs from the viewpoint of a single device, and the system will automatically deduce the tensor sharding and communication patterns. However, with the recent development in large-scale DL models, distributed training exhibits spatial and temporal workload heterogeneity, arising from both device disparities (e.g., mixed hardware, failures) and data variations (e.g., uneven sequence lengths). Such heterogeneity violates SPMD's assumption of uniform workload partitioning, which restricts its ability to express and optimize heterogeneous parallel strategies effectively. To address this, we propose HSPMD within the Hetu v2 system to achieve general and scalable DL training. HSPMD extends SPMD's annotations to support asymmetric sharding and composes standard communication primitives for hierarchical communication, all while retaining the simplicity of a single-device declarative programming model. Leveraging HSPMD, Hetu handles spatial heterogeneity through progressive graph specialization, enabling device-specific execution logic, and addresses temporal heterogeneity via dynamic graph switching. Evaluations on heterogeneous clusters, elastic training, and mixed-length data scenarios show that HSPMD matches or outperforms specialized systems, providing a flexible and efficient solution for modern large-scale model training.
format Preprint
id arxiv_https___arxiv_org_abs_2504_20490
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Hetu v2: A General and Scalable Deep Learning System with Hierarchical and Heterogeneous Single Program Multiple Data Annotations
Li, Haoyang
Fu, Fangcheng
Ge, Hao
Lin, Sheng
Wang, Xuanyu
Niu, Jiawen
Miao, Xupeng
Cui, Bin
Distributed, Parallel, and Cluster Computing
The Single Program Multiple Data (SPMD) paradigm provides a unified abstraction to annotate various parallel dimensions in distributed deep learning (DL) training. With SPMD, users can write training programs from the viewpoint of a single device, and the system will automatically deduce the tensor sharding and communication patterns. However, with the recent development in large-scale DL models, distributed training exhibits spatial and temporal workload heterogeneity, arising from both device disparities (e.g., mixed hardware, failures) and data variations (e.g., uneven sequence lengths). Such heterogeneity violates SPMD's assumption of uniform workload partitioning, which restricts its ability to express and optimize heterogeneous parallel strategies effectively. To address this, we propose HSPMD within the Hetu v2 system to achieve general and scalable DL training. HSPMD extends SPMD's annotations to support asymmetric sharding and composes standard communication primitives for hierarchical communication, all while retaining the simplicity of a single-device declarative programming model. Leveraging HSPMD, Hetu handles spatial heterogeneity through progressive graph specialization, enabling device-specific execution logic, and addresses temporal heterogeneity via dynamic graph switching. Evaluations on heterogeneous clusters, elastic training, and mixed-length data scenarios show that HSPMD matches or outperforms specialized systems, providing a flexible and efficient solution for modern large-scale model training.
title Hetu v2: A General and Scalable Deep Learning System with Hierarchical and Heterogeneous Single Program Multiple Data Annotations
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2504.20490