MILLION: Mastering Long-Context LLM Inference Via Outlier-Immunized KV Product Quantization
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912315788492800 |
|---|---|
| author | Wang, Zongwu Xu, Peng Liu, Fangxin Hu, Yiwei Sun, Qingxiao Li, Gezi Li, Cheng Wang, Xuan Jiang, Li Guan, Haibing |
| author_facet | Wang, Zongwu Xu, Peng Liu, Fangxin Hu, Yiwei Sun, Qingxiao Li, Gezi Li, Cheng Wang, Xuan Jiang, Li Guan, Haibing |
| contents | Large language models (LLMs) are increasingly utilized for complex tasks requiring longer context lengths, with some models supporting up to 128K or 1M tokens. This trend, however, presents significant challenges in inference speed and memory management. Quantization emerges as a promising approach to address the widening gap between LLM size and memory capacity. However, traditional quantization schemes often yield suboptimal compression results for KV caches due to two key factors: i) On-the-fly quantization and de-quantization, causing significant performance overhead; ii) Prevalence of outliers in KV values, challenging low-bitwidth uniform quantization. To this end, we propose MILLION, a novel quantization framework achieving low-bitwidth KV cache through product quantization. First, we conduct a thorough analysis of KV cache distribution, revealing the limitations of existing quantization schemes. Second, we introduce a non-uniform quantization algorithm based on product quantization, which efficiently compresses data while preserving accuracy. Third, we develop a high-performance GPU inference framework with efficient attention kernel and pipeline design for MILLION that leverages sparse computation and asynchronous quantization, significantly enhancing inference speed. Comprehensive evaluation results demonstrate that MILLION can achieve 4 bits quantization with trivial perplexity and accuracy loss, and achieve 2.09x end-to-end performance gains at 32K context length. Code is released at https://github.com/ZongwuWang/MILLION. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_03661 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MILLION: Mastering Long-Context LLM Inference Via Outlier-Immunized KV Product Quantization Wang, Zongwu Xu, Peng Liu, Fangxin Hu, Yiwei Sun, Qingxiao Li, Gezi Li, Cheng Wang, Xuan Jiang, Li Guan, Haibing Distributed, Parallel, and Cluster Computing I.2.0 Large language models (LLMs) are increasingly utilized for complex tasks requiring longer context lengths, with some models supporting up to 128K or 1M tokens. This trend, however, presents significant challenges in inference speed and memory management. Quantization emerges as a promising approach to address the widening gap between LLM size and memory capacity. However, traditional quantization schemes often yield suboptimal compression results for KV caches due to two key factors: i) On-the-fly quantization and de-quantization, causing significant performance overhead; ii) Prevalence of outliers in KV values, challenging low-bitwidth uniform quantization. To this end, we propose MILLION, a novel quantization framework achieving low-bitwidth KV cache through product quantization. First, we conduct a thorough analysis of KV cache distribution, revealing the limitations of existing quantization schemes. Second, we introduce a non-uniform quantization algorithm based on product quantization, which efficiently compresses data while preserving accuracy. Third, we develop a high-performance GPU inference framework with efficient attention kernel and pipeline design for MILLION that leverages sparse computation and asynchronous quantization, significantly enhancing inference speed. Comprehensive evaluation results demonstrate that MILLION can achieve 4 bits quantization with trivial perplexity and accuracy loss, and achieve 2.09x end-to-end performance gains at 32K context length. Code is released at https://github.com/ZongwuWang/MILLION. |
| title | MILLION: Mastering Long-Context LLM Inference Via Outlier-Immunized KV Product Quantization |
| topic | Distributed, Parallel, and Cluster Computing I.2.0 |
| url | https://arxiv.org/abs/2504.03661 |