OSDI 2024 论文评述 Day 3 Session 9: Data Management IPADS-SYS 上海交通大学并行与分布式系统研究所官方知乎账号 阅读全文 【RG Q&A Summary】[OSDI'24] Taming Throughput-Latency Tradeoff in LLM Inference with Sarathi-Serve USTC-NHPCC 中国科学技术大学-国家高性能计算中心-先进数据系统实验室 ...
本文是 OSDI 2024 Day 2 第三个 session 的论文介绍,包含以下五篇论文: When will my ML Job finish? Toward providing Completion Time Estimates through Predictability-Centric Scheduling Optimizing Resource Allocation in Hyperscale Datacenters: Scalability, Usability, and Experiences μSlope: High Compression ...
OSDI 2024 论文评述www.zhihu.com/column/c_1794301172137467905 Automatically Reasoning About How Systems Code Uses the CPU cache Rishabh Iyer, Katerina Argyraki, and George Candea, EPFL 本工作提出了 CFAR 系统,CFAR 可以自动地静态分析一个程序对 Memory 的可能访问序列,并依据此序列自动分析程序对 CPU...
本文是 OSDI 2024 Day 3 第一个 session 的论文介绍,包含以下四篇论文: DSig: Breaking the Barrier of Signatures in Data Centers Ransom Access Memories: Achieving Practical Ransomware Protection in Cloud with DeftPunk Secret Key Recovery in a Global-Scale End-to-End Encryption System Flock: A Frame...
OSDI 2024 论文评述www.zhihu.com/column/c_1794301172137467905 SquirrelFS: using the Rust compiler to check file-system crash consistency Hayley LeBlanc, Nathan Taylor, James Bornholt, and Vijay Chidambaram, University of Texas at Austin 本工作提出用Rust语言的Typestate来帮助开发者简化文件系统crash ...
本文是 OSDI 2024 Day 1 第三个 session 的论文介绍,包含以下四篇论文: ACCL+: an FPGA-Based Collective Engine for Distributed Applications Beaver: Practical Partial Snapshots for Distributed Cloud Services Fast and Scalable In-network Lock Management Using Lock Fission Chop Chop: Byzantine Atomic Broadca...
本文是 OSDI 2024 Day 2 第一个 session 的论文介绍,包含以下五篇论文: Enabling Tensor Language Model to Assist in Generating High-Performance Tensor Programs for Deep Learning Ladder: Enabling Efficient Low-Precision Deep Learning Computing through Hardware-aware Tensor Transformation Caravan: Practical Onl...
第18 届 OSDI 于 2024 年 7 月 10 日 - 7 月 12 日在美国圣克拉拉召开。本次会议共收到 272 篇投稿,接收 49 篇论文,录取率为 18%。同时,本次会议还接收了 4 篇来自 OSDI 23 Revise and Resubmit 的文章,总共有 53 篇文章。 本次大会共选出 3 篇最佳论文,分别是: ...
OSDI 2024 论文评述www.zhihu.com/column/c_1794301172137467905 dLoRA: Dynamically Orchestrating Requests and Adapters for LoRA LLM Serving Bingyang Wu, Ruidong Zhu, and Zili Zhang,School of Computer Science, Peking University;Peng Sun,Shanghai AI Lab;Xuanzhe Liu and Xin Jin,School of Computer ...
IPADS-SYS 上海交通大学并行与分布式系统研究所官方知乎账号 来自专栏 · OSDI 2024 论文评述 201 人赞同了该文章 目录 收起 Taming Throughput-Latency Tradeoff in LLM Inference with Sarathi-Serve 背景 动机和挑战 设计 系统实现 测试评估 Q&A ServerlessLLM: Low-Latency Serverless Inference for Large Lang...