Direct Access, High-Performance Memory Disaggregation with DIRECTCXL Direct Access, High-Performance Memory Disaggregation with DIRECTCXL
基于CXL2.0的内存分解实现，写的非常细致，重点和RDMA做比较
2023-10-29 Paper
A RDMA CXL 内存分解
cgroup和namespace cgroup和namespace
1.1. 介绍内核控制组cgroups是一项内核功能，允许为进程分配和限制硬件和系统资源。 创建：在使用systemd（systemd是System Management Daemon的简写（在UNIX系统中，后台进程都按照惯例以d结尾，因
2023-10-29 Basics
cgroup
x86架构内核调试环境搭建QEMU+GDB x86架构内核调试环境搭建QEMU+GDB
对内核修改后一般先在虚拟机测试在编译到物理机内
2023-10-29 Basics
qemu gdb
不同的总线和内存条组成 不同的总线和内存条组成
一些关于DDR-T更多的信息在论文Characterizing the Performance of Intel Optane Persistent Memory——A Close Look at its On-DIMM Bufferin
2023-10-26 Basics
CXL bus DDR PCIe
X86架构下64位的用户空间和内核空间 X86架构下64位的用户空间和内核空间
大多书本讲解了32为操作系统，那么64位到底有什么不同呢？
2023-10-26 Basics
kernelspace userspace namespace
Linux出现文件很少但磁盘容量已满的情况 Linux出现文件很少但磁盘容量已满的情况
虚拟机磁盘显示99%的容量，但是我的文件只有2.5G，虚拟机一共100G，其他容量来自哪里？使用ncdu磁盘分析工具sudo ncdu /，这会显示磁盘上的文件和目录，按大小排序。可以导航到占用大量空间的目录，并查看哪些文件或子目录占用了大
2023-10-26 Basics
ncdu log rotation
Transparent memory offloading in datacenters Transparent memory offloading in datacenters
在不显著影响性能的情况下自动地offload DRAM到更便宜的存储介质中为数据中心节约内存
2023-10-15 Paper
Hybrid Memory Systems A
Exploring Memory Access Similarity to Improve Irregular Application Performance for Distributed Exploring Memory Access Similarity to Improve Irregular Application Performance for Distributed
也属于异构内存，通过分析并行程序的相似性减少重复操作来提高性能。
2023-10-15 Paper
A 堆叠DRAM
Exploring the Design Space of Page Management for Multi-Tiered Memory Systems Exploring the Design Space of Page Management for Multi-Tiered Memory Systems
这一篇属于翻译。结合PM的混合分层NUMA架构，写的非常仔细的一篇。
2023-10-13 Paper
Hybrid Memory Systems A NUMA
Micron’s Perspective on Impact of CXL on DRAM Bit Growth Rate Micron’s Perspective on Impact of CXL on DRAM Bit Growth Rate
Micron白皮书
2023-10-13 Livre blanc
CXL
Characterizing the Performance of Intel Optane Persistent Memory Characterizing the Performance of Intel Optane Persistent Memory
一篇Optane的测评，和其他测评不同的在于这篇同时测评了Optane一代和二代。
2023-10-13 Paper
A Optane NVDIMMs
An empirical guide to the behavior and use of scalable persistent memory An empirical guide to the behavior and use of scalable persistent memory
探究了Optane的其他浅在特性
2023-10-12 Paper
A Optane NVDIMMs
A Summary of hybrid main memory systems A Summary of hybrid main memory systems
混合内存文献总结
2023-10-09 Paper
Review
An Operating System Level Data Migration Scheme An Operating System Level Data Migration Scheme
对性能和能耗进行建模，指导页面迁移。
2023-10-05 Paper
B Hybrid Memory Systems
Characterizing and Optimizing Hybrid DRAMPM Main Memory System with Application Awareness Characterizing and Optimizing Hybrid DRAMPM Main Memory System with Application Awareness
从并发应用的角度讨论混合内存的使用，可以感知程序的带宽敏感和延迟敏感
2023-10-04 Paper
B Hybrid Memory Systems
Going Vertical in Memory Management Going Vertical in Memory Management
片上缓存容量敏感
2023-09-15 Paper
A Chip on
Dancing in the Dark Profiling for Tiered Memory Dancing in the Dark Profiling for Tiered Memory
我们评估了不同的内存访问收集方法，并提出了一种混合分层内存方法，该方法提供对TMA的全面可见性
2023-09-15 Paper
B Hybrid Memory Systems 内存监控工具
Data Center Ethernet and RDMA Issues at Hyperscale Data Center Ethernet and RDMA Issues at Hyperscale
数据中心以太网和远程直接内存访问:亚毫秒级的问题
2023-08-22 Paper
RDMA
Towards an Adaptable Systems Architecture for Memory Tiering at WarehouseScale Towards an Adaptable Systems Architecture for Memory Tiering at WarehouseScale
面向仓库级内存分层的自适应分层系统。
2023-08-20 Paper
Hybrid Memory Systems A Warehouse-Scale Computing
Merchandiser Data Placement on Heterogeneous Memory for TaskParallel HPC Applications with LoadBalance Awareness Merchandiser Data Placement on Heterogeneous Memory for TaskParallel HPC Applications with LoadBalance Awareness
在高性能并行计算中研究异构内存数据放置问题。
2023-08-19 Paper
Hybrid Memory Systems A
Transparent Page Placement for CXL Enabled Tiered Memory Transparent Page Placement for CXL Enabled Tiered Memory
针对自己公司数据中心的工作负载的特点，修改内核页面放置策略
2023-08-19 Paper
Hybrid Memory Systems A
Aware Data Structure Refinement and Placement for Heterogeneous Memory Systems Aware Data Structure Refinement and Placement for Heterogeneous Memory Systems
提出了一种内存管理方法，该方法结合了动态数据结构细化步骤以及异构内存系统上的放置算法
2023-08-18 Paper
Hybrid Memory Systems Rxiv
P2Cache An Application-Directed Page Cache for Improving Performance of Data-Intensive Applications P2Cache An Application-Directed Page Cache for Improving Performance of Data-Intensive Applications
发现一些特殊的I/O现有的缓存策略无法适用。设计应用程序导向的页面缓存，用于提高数据密集型应用程序的性能。
2023-08-17 Paper
O Cache
Supporting Superpages and Lightweight Page Migration in Hybrid Memory Systems Supporting Superpages and Lightweight Page Migration in Hybrid Memory Systems
提出了一种称为Rainbow的新的内存管理机制来弥合超级页面和轻量级页面迁移之间的根本冲突
2023-05-02 Paper
Hybrid Memory Systems A
2 / 3