https://quhongbin.github.io/posts/Recent content in 文章 on 我的博客Hugozh-CNhttps://quhongbin.github.io/posts/1/01/c-templates/Mon, 01 Jan 0001 00:00:00 +0000https://quhongbin.github.io/posts/1/01/c-templates/<h1 id="c-templates">C++ Templates</h1> <p>Templates are C++'s mechanism for <strong>generic programming</strong> &mdash; writing code that works with any type without repeating yourself. Unlike runtime polymorphism (virtual functions), templates are resolved entirely at compile time: the compiler generates a separate copy of the function or class for each set of type arguments actually used. This is called <strong>template instantiation</strong>.</p> <p>Templates are the foundation of the C++ Standard Library: <code>std::vector&lt;int&gt;</code>, <code>std::map&lt;std::string, double&gt;</code>, <code>std::sort(begin, end)</code> &mdash; all templates. They power everything from containers to algorithms to smart pointers. Understanding templates deeply means understanding how modern C++ achieves its combination of abstraction and zero-overhead performance.</p>https://quhongbin.github.io/posts/1/01/operating-system-operating-system/Mon, 01 Jan 0001 00:00:00 +0000https://quhongbin.github.io/posts/1/01/operating-system-operating-system/<h1 id="process">Process</h1> <ul> <li>A basic resources unit which is assigned with memory in OS - Has isolated memory address <blockquote> <p>arly System</p> </blockquote> </li> </ul> <h1 id="thread">Thread</h1> <ul> <li>A basic unit is called by CPU</li> <li>rely on and within [<em>Process</em>]</li> <li>use the virtual address from the Process, but the Threads must be in same process</li> </ul> <h1 id="coroutine">coroutine</h1> <ul> <li> <hr> </li> </ul> <table> <thead> <tr> <th><strong>dimension</strong></th> <th><strong>Process</strong></th> <th><strong>thread</strong></th> <th><strong>coroutine</strong></th> </tr> </thead> <tbody> <tr> <td><strong>caller</strong></td> <td>os's kernel</td> <td>os's kernel</td> <td>user program</td> </tr> <tr> <td><strong>memory space</strong></td> <td>Individual(no shared)</td> <td>share with Memory of Process</td> <td>share with Memory of Proces</td> </tr> <tr> <td><strong>switching performance overhead</strong></td> <td>very high (us-ms)</td> <td>middle(us)</td> <td>low(ns)</td> </tr> <tr> <td><strong>parallel</strong></td> <td>hundreds of concurrent cons</td> <td>Several thousand up to tens of thousands</td> <td>hundreds of thousands to millions</td> </tr> <tr> <td><strong>concurrent ability</strong></td> <td>multicore</td> <td>multicore</td> <td>concurrent within singal thread</td> </tr> <tr> <td><strong>usage</strong></td> <td>sandbox, dense scenario of CPU-process</td> <td>Web service, dense scenario I/O</td> <td>high-concurrent gateway,async web crawler</td> </tr> </tbody> </table> <hr> <p><a href="memory_alignment.excalidraw">file:memory_alignment.excalidraw</a></p>