惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

有赞技术团队
有赞技术团队
cs.CV updates on arXiv.org
cs.CV updates on arXiv.org
P
Palo Alto Networks Blog
C
Cisco Blogs
The Hacker News
The Hacker News
T
Threatpost
S
Schneier on Security
K
Kaspersky official blog
Spread Privacy
Spread Privacy
博
博客园_首页
钛媒体：引领未来商业与生活新知
钛媒体：引领未来商业与生活新知
NISL@THU
NISL@THU
量
量子位
cs.CL updates on arXiv.org
cs.CL updates on arXiv.org
Google DeepMind News
Google DeepMind News
Security Latest
Security Latest
博
博客园 - 司徒正美
云风的 BLOG
云风的 BLOG
博
博客园 - 叶小钗
H
Hackread – Cybersecurity News, Data Breaches, AI and More
N
News and Events Feed by Topic
爱范儿
爱范儿
P
Proofpoint News Feed
C
CERT Recently Published Vulnerability Notes
Project Zero
Project Zero
Application and Cybersecurity Blog
Application and Cybersecurity Blog
Cisco Talos Blog
Cisco Talos Blog
GbyAI
GbyAI
cs.AI updates on arXiv.org
cs.AI updates on arXiv.org
Apple Machine Learning Research
Apple Machine Learning Research
T
Tenable Blog
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
V
Vulnerabilities – Threatpost
Forbes - Security
Forbes - Security
博
博客园 - 三生石上(FineUI控件)
C
Cyber Attacks, Cyber Crime and Cyber Security
N
News and Events Feed by Topic
V
V2EX
Webroot Blog
Webroot Blog
The Register - Security
The Register - Security
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
阮一峰的网络日志
阮一峰的网络日志
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
Blog — PlanetScale
Blog — PlanetScale
M
MIT News - Artificial intelligence
Scott Helme
Scott Helme
Simon Willison's Weblog
Simon Willison's Weblog
L
LangChain Blog
W
WeLiveSecurity
Cloudbric
Cloudbric

cs.DC updates on arXiv.org

DUAL-BLADE: Dual-Path NVMe-Direct KV-Cache Offloading for Edge LLM Inference Progressive Semantic Communication for Efficient Edge-Cloud Vision-Language Models Efficient, VRAM-Constrained xLM Inference on Clients Folding Tensor and Sequence Parallelism for Memory-Efficient Transformer Training & Inference DORA: A Scalable Asynchronous Reinforcement Learning System for Language Model Training AMMA: A Multi-Chiplet Memory-Centric Architecture for Low-Latency 1M Context Attention Serving RaMP: Runtime-Aware Megakernel Polymorphism for Mixture-of-Experts Spark Policy Toolkit: Semantic Contracts and Scalable Execution for Policy Learning in Spark Internet of Everything in the 6G Era: Paradigms, Enablers, Potentials and Future Directions PolyKV: A Shared Asymmetrically-Compressed KV Cache Pool for Multi-Agent LLM Inference A Survey on Split Learning for LLM Fine-Tuning: Models, Systems, and Privacy Optimizations ITAS: A Multi-Agent Architecture for LLM-Based Intelligent Tutoring Latency and Cost of Multi-Agent Intelligent Tutoring at Scale TACO: Efficient Communication Compression of Intermediate Tensors for Scalable Tensor-Parallel LLM Training FreeScale: Distributed Training for Sequence Recommendation Models with Minimal Scaling Cost CommFuse: Hiding Tail Latency via Communication Decomposition and Fusion for Distributed LLM Training A Taxonomy and Resolution Strategy for Client-Level Disagreements in Federated Learning Usable Agent Discovery for Decentralized AI Systems Cloud to Edge: Benchmarking LLM Inference On Hardware-Accelerated Single-Board Computers Data-Free Contribution Estimation in Federated Learning using Gradient von Neumann Entropy Shard the Gradient, Scale the Model: Serverless Federated Aggregation via Gradient Partitioning Promoting Simple Agents: Ensemble Methods for Event-Log Prediction GraphLeap: Decoupling Graph Construction and Convolution for Vision GNN Acceleration on FPGA AGNT2: Autonomous Agent Economies on Interaction-Optimized Layer 2 Infrastructure FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels Stream-CQSA: Avoiding Out-of-Memory in Attention Computation via Flexible Workload Scheduling A Delta-Aware Orchestration Framework for Scalable Multi-Agent Edge Computing Federated Learning over Blockchain-Enabled Cloud Infrastructure Optimal Routing for Federated Learning over Dynamic Satellite Networks: Tractable or Not? Sherpa.ai Privacy-Preserving Multi-Party Entity Alignment without Intersection Disclosure for Noisy Identifiers Preserving Clusters in Error-Bounded Lossy Compression of Particle Data Unlocking the Edge deployment and ondevice acceleration of multi-LoRA enabled one-for-all foundational LLM UCCL-Zip: Lossless Compression Supercharged GPU Communication Training Time Prediction for Mixed Precision-based Distributed Training Robust Synchronisation for Federated Learning in The Face of Correlated Device Failure Breaking the Training Barrier of Billion-Parameter Universal Machine Learning Interatomic Potentials A Fully GPU-Accelerated Framework for High-Performance Configuration Interaction Selection with Neural Network Quantum States DataCenterGym: A Physics-Grounded Simulator for Multi-Objective Data Center Scheduling Optimizing Stochastic Gradient Push under Broadcast Communications Scepsy: Serving Agentic Workflows Using Aggregate LLM Pipelines ARGUS: Agentic GPU Optimization Guided by Data-Flow Invariants Prefill-as-a-Service: KVCache of Next-Generation Models Could Go Cross-Datacenter Cooperate to Compete: Strategic Data Generation and Incentivization Framework for Coopetitive Cross-Silo Federated Learning Exploiting Correlations in Federated Learning: Opportunities and Practical Limitations ELMoE-3D: Leveraging Intrinsic Elasticity of MoE for Hybrid-Bonding-Enabled Self-Speculative Decoding in On-Premises Serving AgileLog: A Forkable Shared Log for Agents on Data Streams Secure and Privacy-Preserving Vertical Federated Learning Event Tensor: A Unified Abstraction for Compiling Dynamic Megakernel CUTEv2: Unified and Configurable Matrix Extension for Diverse CPU Architectures with Minimal Design Overhead Record-Remix-Replay: Hierarchical GPU Kernel Optimization using Evolutionary Search NimbusGuard: A Novel Framework for Proactive Kubernetes Autoscaling Using Deep Q-Networks Taming Asynchronous CPU-GPU Coupling for Frequency-aware Latency Estimation on Mobile Edge Rebooting Microreboot: Architectural Support for Safe, Parallel Recovery in Microservice Systems A-IO: Adaptive Inference Orchestration for Memory-Bound NPUs SMART: When is it Actually Worth Expanding a Speculative Tree? ConfigSpec: Profiling-Based Configuration Selection for Distributed Edge--Cloud Speculative LLM Serving OpenCLAW-P2P v7.0-P2PCLAW: Resilient Multi-Layer Persistence, Live Reference Verification, and Production-Scale Evaluation of Decentralized AI Peer Review v7.0 -- Mathematical Corrections & Ecosystem Developments Edition DarwinNet: An Evolutionary Network Architecture for Agent-Driven Protocol Synthesis RoboECC: Multi-Factor-Aware Edge-Cloud Collaborative Deployment for VLA Models Hardware Utilization and Inference Performance of Edge Object Detection Under Fault Injection HearthNet: Edge Multi-Agent Orchestration for Smart Homes Token-Budget-Aware Pool Routing for Cost-Efficient LLM Inference Cornserve: A Distributed Serving System for Any-to-Any Multimodal Models Characterizing Performance-Energy Trade-offs of Large Language Models in Multi-Request Workflows ECHO: Elastic Speculative Decoding with Sparse Gating for High-Concurrency Scenarios Duration-Informed Workload Scheduler Domain-Adaptive Model Merging Across Disconnected Modes Why Smaller Is Slower? Dimensional Misalignment in Compressed LLMs veScale-FSDP: Flexible and High-Performance FSDP at Scale AEG: A Baremetal Framework for AI Acceleration via Direct Hardware Access in Heterogeneous Accelerators ACE-Bench: A Lightweight Benchmark for Evaluating Azure SDK Usage Correctness StreamServe: Adaptive Speculative Flows for Low-Latency Disaggregated LLM Serving Emergent Social Structures in Autonomous AI Agent Networks: A Metadata Analysis of 626 Agents on the Pilot Protocol SPEED-Bench: A Unified and Diverse Benchmark for Speculative Decoding Para-B&B: Load-Balanced Deterministic Parallelization of Solving MIP Rashomon Sets and Model Multiplicity in Federated Learning Characterizing WebGPU Dispatch Overhead for LLM Inference Across Four GPU Vendors, Three Backends, and Three Browsers Scalable Explainability-as-a-Service (XaaS) for Edge AI Systems NPU Design for Diffusion Language Model Inference PRAXIS: Integrating Program Analysis with Observability for Root-Cause Analysis BitFlipScope: Scalable Fault Localization and Recovery for Bit-Flip Corruptions in LLMs Cornfigurator: Automated Planning for Any-to-Any Multimodal Model Serving SHARe-KAN: Post-Training Vector Quantization for Cache-Resident KAN Inference Spira: Exploiting Voxel Data Structural Properties for Efficient Sparse Convolution in Point Cloud Networks Power to the Clients: Federated Learning in a Dictatorship Setting From Tokens to Layers: Redefining Stall-Free Scheduling for MoE Serving with Layered Prefill Speculative Actions: A Lossless Framework for Faster Agentic Systems InfiniPipe: Elastic Pipeline Parallelism for Efficient Variable-Length Long-Context LLM Training DPQuant: Efficient and Differentially-Private Model Training via Dynamic Quantization Scheduling HFX: Joint Design of Algorithms and Systems for Multi-SLO Serving and Fast Scaling Reliable Microservice Tail Latency Prediction via Decoupled Dual-Stream Learning and Gradient Modulation On the Surprising Effectiveness of a Single Global Merging in Decentralized Learning FedRef: Bayesian Fine-Tuning using a Reference Model to Mitigate Catastrophic Forgetting for Heterogeneous Federated Learning Sandwich: Joint Configuration Search and Hot-Switching for Efficient CPU LLM Serving MegaScale-Data: Scaling Dataloader for Multisource Large Foundation Model Training RESFL: An Uncertainty-Aware Framework for Responsible Federated Learning by Balancing Privacy, Fairness and Utility BatchLLM: Optimizing Large Batched LLM Inference with Global Prefix Sharing and Throughput-oriented Token Batching Deep Optimizer States: Towards Scalable Training of Transformer Models Using Interleaved Offloading CoreGuard: Safeguarding Foundational Capabilities of LLMs Against Model Stealing in Edge Deployment Cloudless-Training: A Framework to Improve Efficiency of Geo-Distributed ML Training
SLOs-Serve: Optimized Serving of Multi-SLO LLMs
Siyuan Chen, Zhipeng Jia, Samira Khan, Arvind Krishnamurthy, Phi · 2025-04-06 · via cs.DC updates on arXiv.org

This paper introduces SLOs-Serve, a system designed for serving multi-stage large language model (LLM) requests with application- and stage-specific service level objectives (SLOs). The key idea behind SLOs-Serve is to customize the allocation of tokens to meet these SLO requirements. SLOs-Serve uses a multi-SLO dynamic programming-based algorithm to continuously optimize token allocations under SLO constraints by exploring the full design space of chunked prefill and (optional) speculative decoding. Leveraging this resource planning algorithm, SLOs-Serve effectively supports multi-SLOs and multi-replica serving with dynamic request routing while being resilient to bursty arrivals. Our evaluation across 6 LLM application scenarios (including summarization, coding, chatbot, tool calling, and reasoning) demonstrates that SLOs-Serve improves per-GPU serving capacity by 2.2x on average compared to prior state-of-the-art systems.

此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。