Li Zeng
Research Subjects
Li Zeng has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Xin Chen
Huawei Technologies Co Ltd China
2 shared publications - Entong Li
Huawei Technologies Co Ltd China
1 shared publication - Jincheng Lu
Huawei Technologies Co., Ltd
1 shared publication - Jinhua Zhou
Huawei Technologies Co Ltd
1 shared publication - Jinqi Huang
Huawei Technologies Co Ltd China
1 shared publication - Qiheng You
Huawei Technologies Co., Ltd
1 shared publication
Latest Publications
- Open Access
High-Throughput LLM inference on Heterogeneous Clusters
Nowadays, many companies possess various types of AI accelerators, forming heterogeneous clusters. Efficiently leveraging these clusters for high-throughput large language model (LLM) inference services can significantly reduce costs and expedite task processing. However, LLM inference on heterogeneous clusters presents two main challenges. Firstly, different deployment configurations can result in vastly different performance. The number of possible configurations is large, and evaluating the effectiveness of a specific setup is complex. Thus, finding an optimal configuration is not an easy task. Secondly, LLM inference instances within a heterogeneous cluster possess varying processing capacities, leading to different processing speeds for handling inference requests. Evaluating these capacities and designing a request scheduling algorithm that fully maximizes the potential of each instance is challenging. In this paper, we propose a high-throughput inference service system on heterogeneous clusters. First, the deployment configuration is optimized by modeling the resource amount and expected throughput and using the exhaustive search method. Second, a novel mechanism is proposed to schedule requests among instances, which fully considers the different processing capabilities of various instances. Extensive experiments show that the proposed scheduler improves throughput by 122.5% and 33.6% on two heterogeneous clusters, respectively.
Open AccessHigh-Throughput LLM inference on Heterogeneous Clusters
Nowadays, many companies possess various types of AI accelerators, forming heterogeneous clusters. Efficiently leveraging these clusters for high-throughput large language model (LLM) inference services can significantly reduce costs and expedite task processing. However, LLM inference on heterogeneous clusters presents two main challenges. Firstly, different deployment configurations can result in vastly different performance. The number of possible configurations is large, and evaluating the effectiveness of a specific setup is complex. Thus, finding an optimal configuration is not an easy task. Secondly, LLM inference instances within a heterogeneous cluster possess varying processing capacities, leading to different processing speeds for handling inference requests. Evaluating these capacities and designing a request scheduling algorithm that fully maximizes the potential of each instance is challenging. In this paper, we propose a high-throughput inference service system on heterogeneous clusters. First, the deployment configuration is optimized by modeling the resource amount and expected throughput and using the exhaustive search method. Second, a novel mechanism is proposed to schedule requests among instances, which fully considers the different processing capabilities of various instances. Extensive experiments show that the proposed scheduler improves throughput by 122.5% and 33.6% on two heterogeneous clusters, respectively.
- 18 August 2023
- Published by Springer Nature
- p. 462-477
KBQA: Accelerate Fuzzy Path Query on Knowledge Graph
- 18 August 2023
- Published by Springer Nature
- p. 462-477
KBQA: Accelerate Fuzzy Path Query on Knowledge Graph