你的手机是你所拥有的最复杂的机器。让我们把它拆开来看看。
Let’s take apart your phone

原始链接: https://everythingmachine.io/phone/

现代智能手机依靠一组专门的组件来运行: * **处理器:** **CPU** 是通用大脑,利用高性能和高能效核心每秒处理数十亿条指令。**GPU** 并行工作以渲染复杂的视觉效果,而 **NPU** 则专为面部识别和模式匹配等人工智能任务而设计。 * **内存:** 数据存储分为三个层级。**高速缓存 (Cache)** 速度极快但容量较小,位于 CPU 内部。**RAM** 作为运行中应用程序的较大临时工作区,而 **闪存 (Flash storage)** 则为文件和软件提供长期、非易失性的存储空间。 * **安全性:** **安全处理单元** 充当数字保险库,隔离生物识别数据和支付凭证等敏感信息。它不会直接泄露原始机密,而是提供加密的一次性验证,以确保设备安全。 这些组件共同平衡了高速计算、视觉渲染、智能模式识别和强大的数据保护,从而提供无缝的用户体验。

Hacker News 最新 | 过往 | 评论 | 提问 | 展示 | 招聘 | 投稿 登录 你手中的手机是你所拥有过最复杂的机器。让我们把它拆开看看。(everythingmachine.io) 13 分,由 bookofjoe 在 1 小时前发布 | 隐藏 | 过往 | 收藏 | 1 条评论 帮助 jimnotgym 2 分钟前 [–] 我来这里是希望能听到人们拥有过比这更复杂的机器的故事。我抱过一个婴儿。这比手机更胜一筹吗? 回复 指南 | 常见问题 | 列表 | API | 安全 | 法律 | 申请 YC | 联系 搜索:
相关文章

原文

Central Processing Unit (CPU)

This is a general-purpose computer. It fetches an instruction, say “add 2”, from memory, applies that instruction to data, say “2”, then commits the result, “4”. Modern phone CPUs overlap these steps, allowing a single core to complete tens of billions of instructions per second.

They contain six to eight cores, some high-performance and other smaller, power-efficient ones; surging when needed and sipping energy the rest of the time.

Graphics processing unit (GPU)

The GPU renders the display you see. The CPU’s step-by-step approach doesn’t work well for controlling millions of pixels, so the GPU is built differently. Thousands of basic processing units work in parallel, each specialised for the repetitive maths that turns code into moving visuals.

The CPU decides what should appear, the GPU calculates how it should look, and passes that to the screen’s TFT layer to illuminate each pixel.

Neural Processing Unit (NPU)

The newest of the trio, appearing in flagship phones around 2017. Binary transistors are very good at maths. Neurones in your brain are better at the fuzzy work of recognising patterns. The NPU is built for the multiply-and-accumulate operations needed for artificial neural networks, using them to recognise the patterns of speech, faces, objects, and somehow power the black magic of generative AI.

Memory

Memory is groups of transistors which can be set as on or off, representing 1 or 0. Billions of these switches work together to store your data.

There are three types. Ultra-fast cache sits inside the CPU. RAM sits just outside, it’s slower but far larger. Both need to be powered to remember. This is working memory, where active software lives.

Flash storage can remember without power. It is slower than RAM, though still dozens of times faster than the spinning hard drives from a decade ago. This is where the software and your files are stored.

Secure Processing Unit

A large city has dark alleyways and crime. Instead of trying to defend every street, your phone has a separate vault processor, holding secrets like your fingerprint and payment authentication data. It keeps them locked up, but gives out single use codes, answering yes or no to questions like ‘is this their fingerprint?’

联系我们 contact @ memedata.com