MeMo encodes new knowledge into a separate memory model for frozen LLMs, achieving strong performance on BrowseComp-Plus, NarrativeQA, and MuSiQue while capturing cross-document relationships and remaining robust to retrieval noise.
hub
Efficient Memory Management for Large Language Model Serving with PagedAttention
35 Pith papers cite this work. Polarity classification is still indexing.
35
Pith papers citing it
abstract
High throughput serving of large language models (LLMs) requires batching sufficiently many requests at a time. However, existing systems struggle because the key-value cache (KV cache) memory for each request is huge and grows and shrinks dynamically. When managed inefficiently, this memory can be significantly wasted by fragmentation and redundant duplication, limiting the batch size. To address this problem, we propose PagedAttention, an attention algorithm inspired by the classical virtual memory and paging techniques in operating systems. On top of it, we build vLLM, an LLM serving system that achieves (1) near-zero waste in KV cache memory and (2) flexible sharing of KV cache within and across requests to further reduce memory usage. Our evaluations show that vLLM improves the throughput of popular LLMs by 2-4$\times$ with the same level of latency compared to the state-of-the-art systems, such as FasterTransformer and Orca. The improvement is more pronounced with longer sequences, larger models, and more complex decoding algorithms. vLLM's source code is publicly available at https://github.com/vllm-project/vllm
hub tools
citation-role summary
method 2
background 1
citation-polarity summary
claims ledger
- abstract High throughput serving of large language models (LLMs) requires batching sufficiently many requests at a time. However, existing systems struggle because the key-value cache (KV cache) memory for each request is huge and grows and shrinks dynamically. When managed inefficiently, this memory can be significantly wasted by fragmentation and redundant duplication, limiting the batch size. To address this problem, we propose PagedAttention, an attention algorithm inspired by the classical virtual memory and paging techniques in operating systems. On top of it, we build vLLM, an LLM serving system
co-cited works
representative citing papers
Fine-tuned LLMs trained with reinforcement learning using verifiable rewards produce floor plans that satisfy connectivity and numerical constraints, outperforming prior methods with at least 94% relative improvement in compatibility.
NCCLZ decouples quantization and entropy coding across NCCL stack layers to enable overlapped compression, delivering up to 9.65x speedup over plain NCCL on scientific and training workloads.
Power capping is illusory in LLM decode as memory-bound operation leaves power headroom untouched on 700 W GPUs, while SM clock locking saves up to 32% energy and three DVFS classes appear across attention types.
EEP makes wide expert-parallel MoE serving survive single-rank failures with an 11s recovery pause, 8s reintegration pause, and throughput restored to 95% of pre-fault level within 52s while staying within 4.4% of a fixed-membership baseline in steady state.
Apple MPS transformer decoding shows abrupt latency spikes up to 21x in narrow decoding-budget intervals due to KV cache and execution regime shifts, absent on CPU and CUDA.
DUET improves RLVR by allocating tokens across both prompt selection and rollout length, outperforming full-budget baselines even when using only half the tokens.
vOPD stabilizes on-policy distillation gradients by subtracting a closed-form per-token negative reverse KL baseline as a detached control variate, preserving unbiasedness while lowering variance and matching expensive full-vocabulary methods.
CacheFlow cuts TTFT by 10-62% in batched LLM serving via 3D-parallel KV cache restoration and a two-pointer scheduler that overlaps recompute and I/O.
Stealth Pretraining Seeding plants persistent unsafe behaviors in LLMs via diffuse poisoned web content that activates on precise triggers and evades standard evaluation.
Language models learn to evict KV cache entries end-to-end via reinforcement learning from outcome reward alone, achieving 2-3x cache compression while maintaining accuracy on Countdown, AMC, and AIME tasks.
Sparse prefix caching via dynamic programming for optimal checkpoint placement under overlap distributions improves the Pareto frontier for recurrent and hybrid LLM serving on shared-prefix data.
LoSA caches prefix attention for stable tokens in block-wise DLMs and applies sparse attention only to active tokens, preserving near-dense accuracy while achieving 1.54x lower attention density and up to 4.14x speedup.
METIS internalizes curriculum judgment in LLM reinforcement fine-tuning by predicting within-prompt reward variance via in-context learning and jointly optimizing with a self-judgment reward, yielding superior performance and up to 67% faster convergence across math, code, and agent benchmarks.
Slipstream uses asynchronous compaction with trajectory-grounded judge validation to improve long-horizon agent accuracy by up to 8.8 percentage points and reduce latency by up to 39.7%.
HyLo upcycles Transformer LLMs into hybrids with MLA and Mamba2/Gated DeltaNet blocks via staged training and distillation, extending context to 2M tokens and outperforming prior upcycled hybrids on long-context benchmarks.
Fused compressed-domain int4 attention on Apple Silicon delivers 48x speedup and 3.2x KV cache compression for 128K-context 70B models while matching FP16 token predictions.
R^3 optimizes full scientific applications on GPUs better than tuning kernel parameters or compiler flags alone while running nearly an order of magnitude faster than modern evolutionary search methods.
Integrated solar-compute-radiator panels enable orbital satellites to achieve over 100 kW of AI inference compute per metric ton launched, supporting thousands of simultaneous large language model sessions.
MemFactory is a new unified modular framework for memory-augmented LLM agent inference and training that integrates GRPO and reports up to 14.8% relative gains on MemAgent evaluations.
Self-RAG trains LLMs to adaptively retrieve passages on demand and self-critique using reflection tokens, outperforming ChatGPT and retrieval-augmented Llama2 on QA, reasoning, and fact verification.
LLM inference should be reframed and evaluated as energy-to-token production with a Token Production Function that accounts for power, cooling, and efficiency ceilings.
The paper delivers a unified executable benchmarking suite for tool-using agents that enforces a shared evidence-admission contract across web, code, and micro-task environments.
Function-based chunking underperforms other strategies in RAG code completion by 3.57-5.64 points, with context length as the dominant factor.
citing papers explorer
-
MeMo: Memory as a Model
MeMo encodes new knowledge into a separate memory model for frozen LLMs, achieving strong performance on BrowseComp-Plus, NarrativeQA, and MuSiQue while capturing cross-document relationships and remaining robust to retrieval noise.
-
Generative Floor Plan Design with LLMs via Reinforcement Learning with Verifiable Rewards
Fine-tuned LLMs trained with reinforcement learning using verifiable rewards produce floor plans that satisfy connectivity and numerical constraints, outperforming prior methods with at least 94% relative improvement in compatibility.
-
NCCLZ: Compression-Enabled GPU Collectives with Decoupled Quantization and Entropy Coding
NCCLZ decouples quantization and entropy coding across NCCL stack layers to enable overlapped compression, delivering up to 9.65x speedup over plain NCCL on scientific and training workloads.
-
The Illusion of Power Capping in LLM Decode: A Phase-Aware Energy Characterisation Across Attention Architectures
Power capping is illusory in LLM decode as memory-bound operation leaves power headroom untouched on 700 W GPUs, while SM clock locking saves up to 32% energy and three DVFS classes appear across attention types.
-
Surviving Partial Rank Failures in Wide Expert-Parallel MoE Inference
EEP makes wide expert-parallel MoE serving survive single-rank failures with an 11s recovery pause, 8s reintegration pause, and throughput restored to 95% of pre-fault level within 52s while staying within 4.4% of a fixed-membership baseline in steady state.
-
Non-Monotonic Latency in Apple MPS Decoding: KV Cache Interactions and Execution Regimes
Apple MPS transformer decoding shows abrupt latency spikes up to 21x in narrow decoding-budget intervals due to KV cache and execution regime shifts, absent on CPU and CUDA.
-
DUET: Optimize Token-Budget Allocation for Reinforcement Learning with Verifiable Rewards
DUET improves RLVR by allocating tokens across both prompt selection and rollout length, outperforming full-budget baselines even when using only half the tokens.
-
KL for a KL: On-Policy Distillation with Control Variate Baseline
vOPD stabilizes on-policy distillation gradients by subtracting a closed-form per-token negative reverse KL baseline as a detached control variate, preserving unbiasedness while lowering variance and matching expensive full-vocabulary methods.
-
CacheFlow: Efficient LLM Serving with 3D-Parallel KV Cache Restoration
CacheFlow cuts TTFT by 10-62% in batched LLM serving via 3D-parallel KV cache restoration and a two-pointer scheduler that overlaps recompute and I/O.
-
PermaFrost-Attack: Stealth Pretraining Seeding(SPS) for planting Logic Landmines During LLM Training
Stealth Pretraining Seeding plants persistent unsafe behaviors in LLMs via diffuse poisoned web content that activates on precise triggers and evades standard evaluation.
-
Neural Garbage Collection: Learning to Forget while Learning to Reason
Language models learn to evict KV cache entries end-to-end via reinforcement learning from outcome reward alone, achieving 2-3x cache compression while maintaining accuracy on Countdown, AMC, and AIME tasks.
-
Sparse Prefix Caching for Hybrid and Recurrent LLM Serving
Sparse prefix caching via dynamic programming for optimal checkpoint placement under overlap distributions improves the Pareto frontier for recurrent and hybrid LLM serving on shared-prefix data.
-
LoSA: Locality Aware Sparse Attention for Block-Wise Diffusion Language Models
LoSA caches prefix attention for stable tokens in block-wise DLMs and applies sparse attention only to active tokens, preserving near-dense accuracy while achieving 1.54x lower attention density and up to 4.14x speedup.
-
Internalizing Curriculum Judgment for LLM Reinforcement Fine-Tuning
METIS internalizes curriculum judgment in LLM reinforcement fine-tuning by predicting within-prompt reward variance via in-context learning and jointly optimizing with a self-judgment reward, yielding superior performance and up to 67% faster convergence across math, code, and agent benchmarks.
-
Slipstream: Trajectory-Grounded Compaction Validation for Long-Horizon Agents
Slipstream uses asynchronous compaction with trajectory-grounded judge validation to improve long-horizon agent accuracy by up to 8.8 percentage points and reduce latency by up to 39.7%.
-
Long-Context Aware Upcycling: A New Frontier for Hybrid LLM Scaling
HyLo upcycles Transformer LLMs into hybrids with MLA and Mamba2/Gated DeltaNet blocks via staged training and distillation, extending context to 2M tokens and outperforming prior upcycled hybrids on long-context benchmarks.
-
Open-TQ-Metal: Fused Compressed-Domain Attention for Long-Context LLM Inference on Apple Silicon
Fused compressed-domain int4 attention on Apple Silicon delivers 48x speedup and 3.2x KV cache compression for 128K-context 70B models while matching FP16 token predictions.
-
Record-Remix-Replay: Hierarchical GPU Kernel Optimization using Evolutionary Search
R^3 optimizes full scientific applications on GPUs better than tuning kernel parameters or compiler flags alone while running nearly an order of magnitude faster than modern evolutionary search methods.
-
Reduced-Mass Orbital AI Inference via Integrated Solar, Compute, and Radiator Panels
Integrated solar-compute-radiator panels enable orbital satellites to achieve over 100 kW of AI inference compute per metric ton launched, supporting thousands of simultaneous large language model sessions.
-
MemFactory: Unified Inference & Training Framework for Agent Memory
MemFactory is a new unified modular framework for memory-augmented LLM agent inference and training that integrates GRPO and reports up to 14.8% relative gains on MemAgent evaluations.
-
Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection
Self-RAG trains LLMs to adaptively retrieve passages on demand and self-critique using reflection tokens, outperforming ChatGPT and retrieval-augmented Llama2 on QA, reasoning, and fact verification.
-
Position: LLM Inference Should Be Evaluated as Energy-to-Token Production
LLM inference should be reframed and evaluated as energy-to-token production with a Token Production Function that accounts for power, cooling, and efficiency ceilings.
-
An Executable Benchmarking Suite for Tool-Using Agents
The paper delivers a unified executable benchmarking suite for tool-using agents that enforces a shared evidence-admission contract across web, code, and micro-task environments.
-
How Does Chunking Affect Retrieval-Augmented Code Completion? A Controlled Empirical Study
Function-based chunking underperforms other strategies in RAG code completion by 3.57-5.64 points, with context length as the dominant factor.
-
VLMaxxing through FrameMogging Training-Free Anti-Recomputation for Video Vision-Language Models
Training-free adaptive reuse of stable visual state in video VLMs reduces follow-up latency by 15-36x on Qwen2.5-VL while preserving correctness on VideoMME, with smaller first-query speedups via pruning.
-
StreamIndex: Memory-Bounded Compressed Sparse Attention via Streaming Top-k
Chunked streaming top-k enables CSA indexer execution at 1M sequence length with 6.21 GB peak memory and >=0.998 recall on synthetic V4-shaped inputs.
-
EdgeFM: Efficient Edge Inference for Vision-Language Models
EdgeFM is an agent-driven framework that strips non-essential features from VLMs and packages reusable optimized kernels, achieving up to 1.49x speedup over TensorRT-Edge-LLM on NVIDIA Orin while enabling first end-to-end deployment on Horizon Journey hardware.
-
Measurement of Generative AI Workload Power Profiles for Whole-Facility Data Center Infrastructure Planning
High-resolution power profiles for AI workloads on H100 GPUs are measured and scaled to whole-facility energy demand using a bottom-up model, with the dataset made public.
-
Silicon Showdown: Performance, Efficiency, and Ecosystem Barriers in Consumer-Grade LLM Inference
Nvidia achieves 1.6x throughput with NVFP4 but hits a VRAM wall for 70B+ models, while Apple UMA enables linear scaling to 80B at 4-bit with up to 23x better energy efficiency.
-
EasyVideoR1: Easier RL for Video Understanding
EasyVideoR1 delivers an optimized RL pipeline for video understanding in large vision-language models, achieving 1.47x throughput gains and aligned results on 22 benchmarks.
-
Hierarchical vs. Flat Iteration in Shared-Weight Transformers
Hierarchical two-speed shared-weight recurrence in Transformers shows a sharp performance gap compared to independent layer stacking in empirical language modeling tests.
-
Secure On-Premise Deployment of Open-Weights Large Language Models in Radiology: An Isolation-First Architecture with Prospective Pilot Evaluation
An isolation-first on-premise architecture for open-weights LLMs in radiology achieved regulatory approval for processing PHI and showed good utility for text-anchored tasks in a one-week pilot with 22 users.
-
Seed1.5-VL Technical Report
Seed1.5-VL is a compact multimodal model that sets new records on dozens of vision-language benchmarks and outperforms prior systems on agent-style tasks.
-
Yi: Open Foundation Models by 01.AI
Yi models are 6B and 34B open foundation models pretrained on 3.1T curated tokens that achieve strong benchmark results through data quality and targeted extensions like long context and vision alignment.
-
SLM Finetuning for Natural Language to Domain Specific Code Generation in Production
Fine-tuned small language models outperform larger models in natural language to domain-specific code generation with improved performance, latency, and the ability to adapt to customer-specific scenarios without losing general capabilities.