Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2411.01611 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909376271351808 |
|---|---|
| author | Erdogan, Lutfi Eren Kanakagiri, Vijay Anand Raghava Keutzer, Kurt Dong, Zhen |
| author_facet | Erdogan, Lutfi Eren Kanakagiri, Vijay Anand Raghava Keutzer, Kurt Dong, Zhen |
| contents | One of the major bottlenecks for efficient deployment of neural network based recommendation systems is the memory footprint of their embedding tables. Although many neural network based recommendation systems could benefit from the faster on-chip memory access and increased computational power of hardware accelerators, the large embedding tables in these models often cannot fit on the constrained memory of accelerators. Despite the pervasiveness of these models, prior methods in memory optimization and parallelism fail to address the memory and communication costs of large embedding tables on accelerators. As a result, the majority of models are trained on CPUs, while current implementations of accelerators are hindered by issues such as bottlenecks in inter-device communication and main memory lookups. In this paper, we propose a theoretical framework that analyses the communication costs of arbitrary distributed systems that use lookup tables. We use this framework to propose algorithms that maximize throughput subject to memory, computation, and communication constraints. Furthermore, we demonstrate that our method achieves strong theoretical performance across dataset distributions and memory constraints, applicable to a wide range of use cases from mobile federated learning to warehouse-scale computation. We implement our framework and algorithms in PyTorch and achieve up to 6x increases in training throughput on GPU systems over baselines, on the Criteo Terabytes dataset. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_01611 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Stochastic Communication Avoidance for Recommendation Systems Erdogan, Lutfi Eren Kanakagiri, Vijay Anand Raghava Keutzer, Kurt Dong, Zhen Distributed, Parallel, and Cluster Computing Artificial Intelligence Computer Vision and Pattern Recognition Information Retrieval Machine Learning One of the major bottlenecks for efficient deployment of neural network based recommendation systems is the memory footprint of their embedding tables. Although many neural network based recommendation systems could benefit from the faster on-chip memory access and increased computational power of hardware accelerators, the large embedding tables in these models often cannot fit on the constrained memory of accelerators. Despite the pervasiveness of these models, prior methods in memory optimization and parallelism fail to address the memory and communication costs of large embedding tables on accelerators. As a result, the majority of models are trained on CPUs, while current implementations of accelerators are hindered by issues such as bottlenecks in inter-device communication and main memory lookups. In this paper, we propose a theoretical framework that analyses the communication costs of arbitrary distributed systems that use lookup tables. We use this framework to propose algorithms that maximize throughput subject to memory, computation, and communication constraints. Furthermore, we demonstrate that our method achieves strong theoretical performance across dataset distributions and memory constraints, applicable to a wide range of use cases from mobile federated learning to warehouse-scale computation. We implement our framework and algorithms in PyTorch and achieve up to 6x increases in training throughput on GPU systems over baselines, on the Criteo Terabytes dataset. |
| title | Stochastic Communication Avoidance for Recommendation Systems |
| topic | Distributed, Parallel, and Cluster Computing Artificial Intelligence Computer Vision and Pattern Recognition Information Retrieval Machine Learning |
| url | https://arxiv.org/abs/2411.01611 |