Search papers, labs, and topics across Lattice.
RAGPerf is introduced as a benchmarking framework designed to characterize the system behaviors of Retrieval-Augmented Generation (RAG) pipelines by decoupling the workflow into modular components like embedding, indexing, retrieval, reranking, and generation. It allows users to configure parameters, model real-world scenarios with diverse datasets and query distributions, and supports various embedding models, vector databases, and LLMs. The framework automates the collection of performance and accuracy metrics, demonstrating negligible overhead through comprehensive experiments.
Pinpointing performance bottlenecks in RAG pipelines just got easier: RAGPerf offers a modular benchmarking framework to dissect and optimize each component.
We present the design and implementation of a RAG-based AI system benchmarking (RAGPerf) framework for characterizing the system behaviors of RAG pipelines. To facilitate detailed profiling and fine-grained performance analysis, RAGPerf decouples the RAG workflow into several modular components - embedding, indexing, retrieval, reranking, and generation. RAGPerf offers the flexibility for users to configure the core parameters of each component and examine their impact on the end-to-end query performance and quality. RAGPerf has a workload generator to model real-world scenarios by supporting diverse datasets (e.g., text, pdf, code, and audio), different retrieval and update ratios, and query distributions. RAGPerf also supports different embedding models, major vector databases such as LanceDB, Milvus, Qdrant, Chroma, and Elasticsearch, as well as different LLMs for content generation. It automates the collection of performance metrics (i.e., end-to-end query throughput, host/GPU memory footprint, and CPU/GPU utilization) and accuracy metrics (i.e., context recall, query accuracy, and factual consistency). We demonstrate the capabilities of RAGPerf through a comprehensive set of experiments and open source its codebase at GitHub. Our evaluation shows that RAGPerf incurs negligible performance overhead.