当前位置:网站首页>Intel's process roadmap for the next 10 years is exposed: 1.4nm process will be launched in 2029! How?
Intel's process roadmap for the next 10 years is exposed: 1.4nm process will be launched in 2029! How?
2022-07-27 19:44:00 【Xinzhixun】
With Intel's semiconductor manufacturing process, the speed is getting slower ,10nm The craft was also mass produced this year after many ticket jumps , Outside about “ Moore's law is dead ”、“ TSMC surpasses Intel ” The sound of is also constant . however , actually , If we look at the density of transistors , Intel 10nm The process is actually better than TSMC 7nm EUV Process of ( Intel 10nm The transistor density of the process is 100.8MTr/mm², TSMC 7nm EUV The transistor density of the process is 96.5MTr/mm²). And in the next ten years , Intel still hopes to continue to promote Moore's law .
In recent days IEEE International Conference on electronic equipment (IEDM) On ,ASML CEO Disclosed Intel 2019 ~ 2029 Chip manufacturing roadmap in , from 7nm、5nm、3nm、2nm, I have been looking forward to 1.4nm . In addition to simple digital goals , It also exposed some new technologies of Intel around the upcoming process nodes .
however , Foreign media pointed out , Actually this year 9 At the lithography conference in August , Intel has shown this slide , But this time ASML CEO Displayed PPT It is based on the previous PPT Made modifications , It is not officially released by Intel . The official original version of Intel does not disclose the number of process nodes at each time .
According to the exposure PPT Look at , It is expected that Intel's process nodes can be upgraded every two years , The first is from 2019 Year of 10nm process 、 Upgrade to 2021 Year of 7nm Extreme ultraviolet lithography (EUV) process . And then there was 2023 year 5nm,2025 Year of 3nm,2027 Year of 2nm, until 2029 Year of 1.4nm —— This is equivalent to 12 The position of the silicon atoms . obviously , Intel is planning to return to the rhythm of major upgrading every two years .
Besides , Between the two generations of process nodes , Intel will also introduce “ +” and “ ++” The iterative version of , To fully extract the performance of each generation of process . The exception is 10nm, It is already in 10nm + The stage of . So in 2020 and 2021 year , We will see 10nm ++ and 10nm +++.
ASML Think , Intel can do this at an annual pace , But there is also a special team to ensure that a complete process node can overlap with another process node . Interestingly , The slide even mentioned reverse migration (Back Porting), This is a capability considered in chip design ( In the older “ ++” Remake on the node ).
Although Intel is separating chip design from process nodes , But because the process has been locked , Its implementation is not a simple thing . From the slides , Reverse migration still has some limitations , Like the first generation 7nm The design can be reverse ported to 10nm +++,5nm Both support fallback to 7nm ++,3nm Portable to 5nm ++, as well as 2nm To 3nm ++.
It's important to point out that , This is not the first time Intel has mentioned reverse porting hardware design . Because the current generation 10nm The manufacturing process is lagging behind , Intel has considered 10nm and 10nm My retreat . Usually with the development of process nodes , Each major version iteration has a different team in charge . But the slide points out , Intel is developing 10nm Optimization and 7nm Series process technology .
Looking forward to the future , We will also see based on 10nm Chip design 7nm product 、 be based on 7nm The design of the 5nm product 、 And based on 5nm The design of the 3nm product .
It is worth mentioning that ,2023 Year of 5nm Process node , It happens to be ASML Sell it High NA EUV The time of the lithography machine , In addition, Intel has been considering new materials and new transistor designs .
Besides , In this year's IEDM The meeting , Intel also introduced that it can realize more advanced manufacturing process “2D Self assembly ” material , The size is about 0.3nm . Of course , There are still many problems to be solved if we want to succeed in mass production .
For self-assembly materials, please refer to previous articles of smart core :《 The new direction of chip manufacturing industry :“ Self assembly ” Technical analysis 》
edit : Smart core - Ranger sword source :TechSpot、cnbeta
边栏推荐
- [basic knowledge of deep learning - 45] distance calculation methods commonly used in machine learning
- 5W奖金池/面向高校,2022法律科技创新大赛报名火热进行中
- go-zero单体服务使用泛型简化注册Handler路由
- Golang sets the domestic image, vscode configures the golang development environment, and vscode debugs the golang code
- ReferenceError: __ dirname is not defined in ES module scope
- C language: 9. Return in main function
- C language preprocessing instruction
- [basic knowledge of deep learning - 49] kmeans
- 【深度学习目标检测系列 - 01】目标检测是什么
- Hdu1573 x problem [univariate linear congruence equations]
猜你喜欢
![[basic knowledge of deep learning - 46] Bayesian theorem and conditional probability formula](/img/9f/b9d7503404e068495fd8613df29366.png)
[basic knowledge of deep learning - 46] Bayesian theorem and conditional probability formula

Kettle references external scripts to complete phone number cleaning, de duplication and indentation

C language: 10. Input stream, output stream, error stream

A lock faster than read-write lock. Don't get to know it quickly

成本高、落地难、见效慢,开源安全怎么办?

低代码实现探索(四十五)业务参数

c语言:11、管道

C language: 14. Preprocessing

电商商城小程序项目完整源码(微信小程序)

c语言:5、多维数组
随机推荐
Anaconda下安装Talib库
王牌代码静态测试工具Helix QAC 2022.2中的新增功能(2)
C language: 14. Preprocessing
5W bonus pool / for colleges and universities, 2022 legal science and technology innovation competition is in progress
C language: 9. Return in main function
嵌入式C语言对次数固定的循环的优化
C language: 12. GDB tool debugging C program
【深度学习基础知识 - 45】机器学习中常用的距离计算方法
Hdu1573 x problem [univariate linear congruence equations]
Yanghui triangle
SQL time processing (SQL server\oracle)
Under the heat wave of Web3.0, the ecological shock of Mensa struck
c语言:11、管道
c语言:10、输入流,输出流,错误流
[basic knowledge of deep learning - 43] concept of odds ratio
【深度学习基础知识 - 42】逻辑回归详解
全球手机芯片设计领域首家!紫光展锐荣膺TMMi4级国际认证
Map和Set
估值超156亿元!华勤通讯完成10亿元B轮融资!高通创投、英特尔资本领投
我想咨询下,我们的maxcompute spark程序需要访问redis,开发环境和生产环境redi