HexGen: Generative Inference of Large Language Model over Heterogeneous Environment
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2023
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910459492302848 |
|---|---|
| author | Jiang, Youhe Yan, Ran Yao, Xiaozhe Zhou, Yang Chen, Beidi Yuan, Binhang |
| author_facet | Jiang, Youhe Yan, Ran Yao, Xiaozhe Zhou, Yang Chen, Beidi Yuan, Binhang |
| contents | Serving generative inference of the large language model is a crucial component of contemporary AI applications. This paper focuses on deploying such services in a heterogeneous and cross-datacenter setting to mitigate the substantial inference costs typically associated with a single centralized datacenter. Towards this end, we propose HexGen, a flexible distributed inference engine that uniquely supports the asymmetric partition of generative inference computations over both tensor model parallelism and pipeline parallelism and allows for effective deployment across diverse GPUs interconnected by a fully heterogeneous network. We further propose a sophisticated scheduling algorithm grounded in constrained optimization that can adaptively assign asymmetric inference computation across the GPUs to fulfill inference requests while maintaining acceptable latency levels. We conduct an extensive evaluation to verify the efficiency of HexGen by serving the state-of-the-art Llama-2 (70B) model. The results suggest that HexGen can choose to achieve up to 2.3 times lower latency deadlines or tolerate up to 4 times more request rates compared with the homogeneous baseline given the same budget. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2311_11514 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | HexGen: Generative Inference of Large Language Model over Heterogeneous Environment Jiang, Youhe Yan, Ran Yao, Xiaozhe Zhou, Yang Chen, Beidi Yuan, Binhang Distributed, Parallel, and Cluster Computing Serving generative inference of the large language model is a crucial component of contemporary AI applications. This paper focuses on deploying such services in a heterogeneous and cross-datacenter setting to mitigate the substantial inference costs typically associated with a single centralized datacenter. Towards this end, we propose HexGen, a flexible distributed inference engine that uniquely supports the asymmetric partition of generative inference computations over both tensor model parallelism and pipeline parallelism and allows for effective deployment across diverse GPUs interconnected by a fully heterogeneous network. We further propose a sophisticated scheduling algorithm grounded in constrained optimization that can adaptively assign asymmetric inference computation across the GPUs to fulfill inference requests while maintaining acceptable latency levels. We conduct an extensive evaluation to verify the efficiency of HexGen by serving the state-of-the-art Llama-2 (70B) model. The results suggest that HexGen can choose to achieve up to 2.3 times lower latency deadlines or tolerate up to 4 times more request rates compared with the homogeneous baseline given the same budget. |
| title | HexGen: Generative Inference of Large Language Model over Heterogeneous Environment |
| topic | Distributed, Parallel, and Cluster Computing |
| url | https://arxiv.org/abs/2311.11514 |