bit-wise reverse

Matters Computational-ideas, algorithms, source code

This is a book for the computationalist, whether a working programmer or anyone interested in methods of computation. The focus is on material that does not usually appear in textbooks on algorithms.

Linux Shell Scripting Cookbook

Table of Contents Preface 1 Chapter 1: Shell Something Out 7 Introduction 7 Printing in the terminal 9 Playing with variables and environment variables 12 Doing math calculations with the shell 17 Playing with file descriptors and redirection 19 Arrays and associative arrays 25 Visiting aliases 27 Grabbing information about terminal 29 Getting, setting dates, and delays 30 Debugging the script 33 Functions and arguments 35 Reading the output of a sequence of commands 38 Reading "n" characters without pressing Return 40 Field separators and iterators 41 Comparisons and tests 44 Chapter 2: Have a Good Command 49 Introduction 50 Concatenating with cat 50 Recording and playback of terminal sessions 53 Finding files and file listing 55 Playing with xargs 63 Translating with tr 69 Checksum and verification 72 Sorting, unique and duplicates 75 Temporary file naming and random numbers 80 Splitting files and data 81 ii Table of Contents Slicing file names based on extension 84 Renaming and moving files in bulk 86 Spell checking and dictionary manipulation 89 Automating interactive input 90 Chapter 3: File In, File Out 95 Introduction 96 Generating files of any size 96 Intersection and set difference (A-B) on text files 97 Finding and deleting duplicate files 100 Making directories for a long path 103 File permissions, ownership, and sticky bit 104 Making files immutable 109 Generating blank files in bulk 110 Finding a symbolic link and its target 111 Enumerating file type statistics 113 Loopback files and mounting 115 Creating ISO files, Hybrid ISO 117 Finding difference between files, patching 120 head and tail – printing the last or first 10 lines 122 Listing only directories – alternative methods 125 Fast command-line navigation using pushd and popd 126 Counting number of lines, words, and characters in a file 128 Printing directory tree 129 Chapter 4: Texting and Driving 131 Introduction 132 Basic regular expression primer 132 Searching and mining "text" inside a file with grep 136 Column-wise cutting of a file with cut 142 Frequency of words used in a given file 146 Basic sed primer 147 Basic awk primer 150 Replacing strings from a text or file 156 Compressing or decompressing JavaScript 158 Iterating through lines, words, and characters in a file 161 Merging multiple files as columns 162 Printing the nth word or column in a file or line 163 Printing text between line numbers or patterns 164 Checking palindrome strings with a script 165 Printing lines in the reverse order 169 Parsing e-mail addresses and URLs from text 171 iii Table of Contents Printing n lines before or after a pattern in a file 172 Removing a sentence in a file containing a word 174 Implementing head, tail, and tac with awk 175 Text slicing and parameter operations 177 Chapter 5: Tangled Web? Not At All! 179 Introduction 180 Downloading from a web page 180 Downloading a web page as formatted plain text 183 A primer on cURL 183 Accessing Gmail from the command line 188 Parsing data from a website 189 Image crawler and downloader 191 Web photo album generator 193 Twitter command-line client 195 define utility with Web backend 197 Finding broken links in a website 199 Tracking changes to a website 200 Posting to a web page and reading response 203 Chapter 6: The Backup Plan 205 Introduction 205 Archiving with tar 206 Archiving with cpio 211 Compressing with gunzip (gzip) 212 Compressing with bunzip (bzip) 215 Compressing with lzma 217 Archiving and compressing with zip 219 squashfs – the heavy compression filesystem 220 Cryptographic tools and hashes 222 Backup snapshots with rsync 224 Version control based backup with Git 227 Cloning hard drive and disks with dd 230 Chapter 7: The Old-boy Network 233 Introduction 233 Basic networking primer 234 Let's ping! 241 Listing all the machines alive on a network 243 Transferring files 247 Setting up an Ethernet and wireless LAN with script 250 Password-less auto-login with SSH 253 Running commands on remote host with SSH 255

Verilog位宽定义：降序与升序写法的原理、差异与工程实践

在数字电路设计与硬件描述语言中，向量（Vector）是表示多位信号的基础数据结构，其位序定义直接影响数据的存储、访问与解释方式。Verilog/SystemVerilog支持两种位宽定义语法：降序（Descending）写法 `[high:low]` 和升序（Ascending）写法 `[low:high]`。从原理上看，这两种写法定义了向量中最高有效位（MSB）与最低有效位（LSB）的不同索引映射关系，本质上是内存布局与位序约定的差异。正确的位序选择对于确保数据在模块间正确传递、与外部IP或总线协议无缝对

O3 Mini不是模型，而是轻量级思维链推理的工程范式

在大模型落地实践中，'思维链（Chain-of-Thought）'能力与'边缘设备部署'需求日益交汇，催生出一类以观察能力为内核、3秒延迟为硬约束、4GB显存为运行边界的轻量化推理范式。其技术本质并非新模型发布，而是对o1-mini等强推理模型的能力锚定与工程压缩——通过知识蒸馏或INT4量化，在消费级GPU上复现多步逻辑推演能力。这类方案广泛应用于本地AI助手、嵌入式决策终端及离线教育工具等场景，核心挑战在于平衡KV Cache开销、Tokenizer吞吐与数值精度。本文聚焦Phi-3-medium与EX

AMD ROCm本地AI部署实战：从LM Studio到生产级推理

ROCm是AMD构建的开源异构计算平台，不同于CUDA的封闭生态，它以分层解耦架构支持GPU加速AI推理。其核心原理在于HIP运行时对CUDA API的语义映射、ROCm Runtime对AMD GPU wavefront调度器的深度适配，以及Infinity Cache等硬件特性的软件协同优化。技术价值体现在打破NVIDIA生态依赖，为中小团队提供高性价比本地AI基础设施。典型应用场景包括轻量大模型（如Mistral 7B）在消费级显卡（RX 7900 XTX）上的低延迟推理、教育科研环境中的可复现实验部