Wonbeom Lee, Jungi Lee, Junghwan Seo, and Jaewoong Sim, Seoul National University
Transformer-based large language models (LLMs) demonstrate impressive performance across various natural language processing tasks. Serving LLM inference for generating long contents, however, poses a challenge due to the enormous memory footprint of the transient state, known as the key-value (KV) cache, which scales with the sequence length and batch size. In this paper, we present InfiniGen, a novel KV cache management framework tailored for long-text generation, which synergistically works with modern offloading-based inference systems. InfiniGen leverages the key insight that a few important tokens that are essential for computing the subsequent attention layer in the Transformer can be speculated by performing a minimal rehearsal with the inputs of the current layer and part of the query weight and key cache of the subsequent layer. This allows us to prefetch only the essential KV cache entries (without fetching them all), thereby mitigating the fetch overhead from the host memory in offloading-based LLM serving systems. Our evaluation on several representative LLMs shows that InfiniGen improves the overall performance of a modern offloading-based system by up to 3.00× compared to prior KV cache management methods while offering substantially better model accuracy.
OSDI '24 Open Access Sponsored by
King Abdullah University of Science and Technology (KAUST)
Open Access Media
USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.
author = {Wonbeom Lee and Jungi Lee and Junghwan Seo and Jaewoong Sim},
title = {{InfiniGen}: Efficient Generative Inference of Large Language Models with Dynamic {KV} Cache Management},
booktitle = {18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24)},
year = {2024},
isbn = {978-1-939133-40-3},
address = {Santa Clara, CA},
pages = {155--172},
url = {https://www.usenix.org/conference/osdi24/presentation/lee},
publisher = {USENIX Association},
month = jul
}
![](https://www.usenix.org/sites/default/files/usenix_artifact_evaluation_available_125_update.png)
![](https://www.usenix.org/sites/default/files/usenix_artifact_evaluation_functional_125.png)
![](https://www.usenix.org/sites/default/files/usenix_artifact_evaluation_reproduced_125.png)