# NanoFlow: Towards Optimal Large Language Model Serving Throughput

arXiv: https://arxiv.org/abs/2408.12757 · HTML: https://mlsanity.com/paper/2408.12757/

## Key facts

- Cited by 90 papers in the map (data through 14 Oct 2026).
- Placed in Serving systems and request scheduling, one of 12 directions in Efficient LLM inference and serving.
- Builds on 22 of its 63 references that are in the map.
- First posted on arXiv 22 Aug 2024.

## Placed in

- Efficient LLM inference and serving / [Serving systems and request scheduling](https://mlsanity.com/inference/serving-systems-and-request-scheduling/)

## Builds on (most cited first, 22 of 22)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://mlsanity.com/paper/1909.08053/) (2019)
- [Fast Inference from Transformers via Speculative Decoding](https://mlsanity.com/paper/2211.17192/) (2022)
- [GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints](https://mlsanity.com/paper/2305.13245/) (2023)
- [Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning](https://mlsanity.com/paper/1702.03118/) (2017)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPU](https://mlsanity.com/paper/2303.06865/) (2023)
- [Splitwise: Efficient generative LLM inference using phase splitting](https://mlsanity.com/paper/2311.18677/) (2023)
- [Efficiently Scaling Transformer Inference](https://mlsanity.com/paper/2211.05102/) (2022)
- [Taming Throughput-Latency Tradeoff in LLM Inference with Sarathi-Serve](https://mlsanity.com/paper/2403.02310/) (2024)
- [Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference](https://mlsanity.com/paper/2406.10774/) (2024)
- [LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset](https://mlsanity.com/paper/2309.11998/) (2023)
- [Atom: Low-bit Quantization for Efficient and Accurate LLM Serving](https://mlsanity.com/paper/2310.19102/) (2023)
- [QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving](https://mlsanity.com/paper/2405.04532/) (2024)
- [SARATHI: Efficient LLM Inference by Piggybacking Decodes with Chunked Prefills](https://mlsanity.com/paper/2308.16369/) (2023)
- [DeepSpeed-FastGen: High-throughput Text Generation for LLMs via MII and DeepSpeed-Inference](https://mlsanity.com/paper/2401.08671/) (2024)
- [Automatic Horizontal Fusion for GPU Kernels](https://mlsanity.com/paper/2007.01277/) (2020)

## Built on by (most cited first, 30 of 90)

- [FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Serving](https://mlsanity.com/paper/2501.01005/) (2025)
- [LongCat-Flash Technical Report](https://mlsanity.com/paper/2509.01322/) (2025)
- [POD-Attention: Unlocking Full Prefill-Decode Overlap for Faster LLM Inference](https://mlsanity.com/paper/2410.18038/) (2024)
- [Continuum: Efficient and Robust Multi-Turn LLM Agent Scheduling with KV Cache Time-to-Live](https://mlsanity.com/paper/2511.02230/) (2025)
- [LServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention](https://mlsanity.com/paper/2502.14866/) (2025)
- [NEO: Saving GPU Memory Crisis with CPU Offloading for Online LLM Inference](https://mlsanity.com/paper/2411.01142/) (2024)
- [Laminar: A Scalable Asynchronous RL Post-Training Framework](https://mlsanity.com/paper/2510.12633/) (2025)
- [Strata: Hierarchical Context Caching for Long Context Language Model Serving](https://mlsanity.com/paper/2508.18572/) (2025)
- [Tactic: Adaptive Sparse Attention with Clustering and Distribution Fitting for Long-Context LLMs](https://mlsanity.com/paper/2502.12216/) (2025)
- [Prefill-Decode Aggregation or Disaggregation? Unifying Both for Goodput-Optimized LLM Serving](https://mlsanity.com/paper/2508.01989/) (2025)
- [PrefillOnly: An Inference Engine for Prefill-only Workloads in Large Language Model Applications](https://mlsanity.com/paper/2505.07203/) (2025)
- [BlendServe: Optimizing Offline Inference for Auto-regressive Large Models with Resource-aware Batching](https://mlsanity.com/paper/2411.16102/) (2024)
- [MoESD: Unveil Speculative Decoding's Potential for Accelerating Sparse MoE](https://mlsanity.com/paper/2505.19645/) (2025)
- [METIS: Fast Quality-Aware RAG Systems with Configuration Adaptation](https://mlsanity.com/paper/2412.10543/) (2024)
- [semi-PD: Towards Efficient LLM Serving via Phase-Wise Disaggregated Computation and Unified Storage](https://mlsanity.com/paper/2504.19867/) (2025)
- [LLM-42: Enabling Determinism in LLM Inference with Verified Speculation](https://mlsanity.com/paper/2601.17768/) (2026)
- [Seesaw: High-throughput LLM Inference via Model Re-sharding](https://mlsanity.com/paper/2503.06433/) (2025)
- [The Energy Cost of Reasoning: Analyzing Energy Usage in LLMs with Test-time Compute](https://mlsanity.com/paper/2505.14733/) (2025)
- [Nexus:Proactive Intra-GPU Disaggregation of Prefill and Decode in LLM Serving](https://mlsanity.com/paper/2507.06608/) (2025)
- [Kinetics: Rethinking Test-Time Scaling Laws](https://mlsanity.com/paper/2506.05333/) (2025)
- [HyGen: Efficient LLM Serving via Elastic Online-Offline Request Co-location](https://mlsanity.com/paper/2501.14808/) (2025)
- [Efficient and Adaptable Overlapping for Computation and Communication via Signaling and Reordering](https://mlsanity.com/paper/2504.19519/) (2025)
- [On Evaluating Performance of LLM Inference Serving Systems](https://mlsanity.com/paper/2507.09019/) (2025)
- [Accelerating Large-Scale Reasoning Model Inference with Sparse Self-Speculative Decoding](https://mlsanity.com/paper/2512.01278/) (2025)
- [TeleRAG: Efficient Retrieval-Augmented Generation Inference with Lookahead Retrieval](https://mlsanity.com/paper/2502.20969/) (2025)
- [COMET: Towards Partical W4A4KV4 LLMs Serving](https://mlsanity.com/paper/2410.12168/) (2024)
- [EPS-MoE: Expert Pipeline Scheduler for Cost-Efficient MoE Inference](https://mlsanity.com/paper/2410.12247/) (2024)
- [DynaServe: Unified and Elastic Execution for Dynamic Disaggregated LLM Serving](https://mlsanity.com/paper/2504.09285/) (2025)
- [SparseServe: Unlocking Parallelism for Dynamic Sparse Attention in Long-Context LLM Serving](https://mlsanity.com/paper/2509.24626/) (2025)
- [SPAD: Specialized Prefill and Decode Hardware for Disaggregated LLM Inference](https://mlsanity.com/paper/2510.08544/) (2025)

## Abstract (arXiv)

Large Language Models (LLMs) have resulted in a surging demand for planet-scale serving systems, where tens of thousands of GPUs continuously serve hundreds of millions of users. Consequently, throughput has emerged as a key metric that determines serving systems' performance. Due to large model sizes and memory-intensive self-attention, LLM serving has been commonly assumed to be memory-bound. Through a detailed analysis, we show that despite having memory-intensive components, end-to-end LLM serving is compute bound for most common workloads and LLMs. Alas, most existing serving engines fall short from optimal compute utilization, because the heterogeneous operations that comprise LLM serving--compute, memory, networking--are executed sequentially within a device. We propose NanoFlow, a novel serving framework that exploits intra-device parallelism, which overlaps the usage of heterogeneous resources within a single device. NanoFlow splits inputs into smaller nano-batches and duplicates operations to operate on each portion independently, enabling overlapping. NanoFlow automatically identifies the number, size, ordering, and GPU resource allocation of nano-batches to minimize the execution time, while considering the interference of concurrent operations. We evaluate NanoFlow's end-to-end serving throughput on several popular models such as LLaMA-2-70B, Mixtral 8x7B, LLaMA-3-8B, etc. With practical workloads, NanoFlow provides 1.91x throughput boost compared to state-of-the-art serving systems achieving 50% to 72% of optimal throughput across popular models.