{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/relayattention-for-efficient-large-language","title":"RelayAttention for Efficient Large Language Model Serving with Long System Prompts","arxiv_id":"2402.14808","date":"2024-02-22","proceeding":null,"authors":["Lei Zhu","Xinjiang Wang","Wayne Zhang","Rynson W. H. Lau"],"abstract":"A practical large language model (LLM) service may involve a long system prompt, which specifies the instructions, examples, and knowledge documents of the task and is reused across requests. However, the long system prompt causes throughput/latency bottlenecks as the cost of generating the next token grows w.r.t. the sequence length. This paper aims to improve the efficiency of LLM services that involve long system prompts. Our key observation is that handling these system prompts requires heavily redundant memory accesses in existing causal attention computation algorithms. Specifically, for batched requests, the cached hidden states (\\ie, key-value pairs) of system prompts are transferred from off-chip DRAM to on-chip SRAM multiple times, each corresponding to an individual request. To eliminate such a redundancy, we propose RelayAttention, an attention algorithm that allows reading these hidden states from DRAM exactly once for a batch of input tokens. RelayAttention is a free lunch: it maintains the generation quality while requiring no model retraining, as it is based on a mathematical reformulation of causal attention. We have observed significant performance improvements to a production-level system, vLLM, through integration with RelayAttention. The improvements are even more profound with longer system prompts.","url_abs":"https://arxiv.org/abs/2402.14808v3","url_pdf":"https://arxiv.org/pdf/2402.14808v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"relayattention-for-efficient-large-language","repo_url":"https://github.com/rayleizhu/vllm-ra","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"large-language-model","task_name":"Large Language Model"}],"methods":[{"method_slug":null,"method_name":null}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2402.14808","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2402.14808"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/rayleizhu/vllm-ra","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"ran":3},"by_repo_kind":{"official":{"samples":3,"ran":3,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"9f09d0beacdd35d1","entry":"get_lock","repo":"rayleizhu/vllm-ra","repo_kind":"official","path":"vllm/model_executor/weight_utils.py","file_url":"https://github.com/rayleizhu/vllm-ra/blob/HEAD/vllm/model_executor/weight_utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"9f09d0beacdd35d1"}},{"code_sha256_prefix":"13adeb89dc2bcd5c","entry":"init_logger","repo":"rayleizhu/vllm-ra","repo_kind":"official","path":"vllm/logger.py","file_url":"https://github.com/rayleizhu/vllm-ra/blob/HEAD/vllm/logger.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"13adeb89dc2bcd5c"}},{"code_sha256_prefix":"8be17b096145392b","entry":"prepare_hf_model_weights","repo":"rayleizhu/vllm-ra","repo_kind":"official","path":"vllm/model_executor/weight_utils.py","file_url":"https://github.com/rayleizhu/vllm-ra/blob/HEAD/vllm/model_executor/weight_utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"8be17b096145392b"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}