Chameleon: a Heterogeneous and Disaggregated Accelerator System for Retrieval-Augmented Language Models
This work addresses the computational inefficiency in serving RALMs for AI applications, offering a scalable solution with significant performance improvements, though it is incremental as it builds on existing accelerator technologies.
The paper tackled the problem of efficiently serving Retrieval-Augmented Language Models (RALMs) by proposing Chameleon, a heterogeneous accelerator system that integrates LLM and vector search accelerators in a disaggregated architecture, resulting in up to 2.16x reduction in latency and 3.18x speedup in throughput compared to a hybrid CPU-GPU architecture.
A Retrieval-Augmented Language Model (RALM) combines a large language model (LLM) with a vector database to retrieve context-specific knowledge during text generation. This strategy facilitates impressive generation quality even with smaller models, thus reducing computational demands by orders of magnitude. To serve RALMs efficiently and flexibly, we propose Chameleon, a heterogeneous accelerator system integrating both LLM and vector search accelerators in a disaggregated architecture. The heterogeneity ensures efficient serving for both inference and retrieval, while the disaggregation allows independent scaling of LLM and vector search accelerators to fulfill diverse RALM requirements. Our Chameleon prototype implements vector search accelerators on FPGAs and assigns LLM inference to GPUs, with CPUs as cluster coordinators. Evaluated on various RALMs, Chameleon exhibits up to 2.16$\times$ reduction in latency and 3.18x speedup in throughput compared to the hybrid CPU-GPU architecture. The promising results pave the way for adopting heterogeneous accelerators for not only LLM inference but also vector search in future RALM systems.