To The Moon
Home
News
TigerAI
Log In
Sign Up
年轻的太太
+Follow
Posts · 28
Posts · 28
Following · 0
Following · 0
Followers · 0
Followers · 0
年轻的太太
年轻的太太
·
2020-11-21
??
Sorry, this post has been deleted
看
2.04K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-21
??
He Xiaopeng, CEO of XPeng Motors, suspected of responding to rumors of stealing technology from a distance
今日小鹏汽车CEO何小鹏发布朋友圈,疑似隔空回应马斯克指小鹏偷窃技术。何小鹏称,“看来昨天我们发布的包含激光雷达的小鹏下一代自动驾驶架构,让西边的某人很不爽,连续用pigu发声。我想说的是,造谣早就证明是无法打败任何竞争对手的,明年开始,在中国的自动驾驶你、要有思想准备被我们打的找不着东,至于国际,我们会相遇的。” 海量资讯、精准解读,尽在新浪财经APP 责任编辑:尹悦
He Xiaopeng, CEO of XPeng Motors, suspected of responding to rumors of stealing technology from a distance
看
1.95K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-21
??
The book return of True Perfect Diary is nearly 1.4 billion US dollars. Can it redeem the angel investment winter?
投资完美日记对真格基金来说意义重大,是可以记在这家机构投资历史上,乃至中国天使投资历史上的案例。 完美日记创办四年即上市,又是个标准的风险投资的好故事。但从不同的视角出发,看到意义也不尽相同。 从完
The book return of True Perfect Diary is nearly 1.4 billion US dollars. Can it redeem the angel investment winter?
看
2.06K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-21
??
Samsung's Three Challenges Against TSMC
最近三星代工挑战台积电的话题引起我很大的兴趣,在过去三十年与代工厂密切合作中,我有自己不同的看法。正文:彭博社最近发布一篇名为“三星2020年超越台积电,加剧芯片战”的文章,这是“三星1160亿美元计划挑战台积电”文章的后续。据作者表述,三星和台积电将在3nm节点展开业务竞争,这意味着在2022年大批量生产。IBM使用了三星授权的GF 14nm,因此他们将继续使用三星7nm。
Samsung's Three Challenges Against TSMC
看
1.48K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-21
??
Intel's tortoise and rabbit race against AMD
英特尔与AMD的竞争是乌龟与野兔的竞争。这在AMD方面尤其困难,因为AMD并没有真正相当于英特尔的Ark提供单一的处理器列表,按代、类型和发布日期排序。随着近70%的单线程性能优势被英特尔的qx9650所取代,以及焦点突然回到单线程性能的王者地位,至少在这些高端CPU中是这样——AMD 21世纪的第一个黄金时代正式结束了。在整个2007年至2013年期间,英特尔都比AMD拥有单线程优势,而且通常是相当多的优势。
Intel's tortoise and rabbit race against AMD
看
1.93K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-11
??
Sorry, this post has been deleted
看
2.16K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-11
??
Sorry, this post has been deleted
看
1.79K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-11
??
Sorry, this post has been deleted
看
1.52K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-11
??
Sorry, this post has been deleted
看
2.58K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
年轻的太太
年轻的太太
·
2020-11-11
??
Sorry, this post has been deleted
看
2.17K
回复
Comment
点赞
Like
编组 21备份 2
Share
Report
Load more
Most Discussed
{"i18n":{"language":"en_US"},"isCurrentUser":false,"userPageInfo":{"id":"3567118012147142","uuid":"3567118012147142","gmtCreate":1604025983334,"gmtModify":1604025983334,"name":"年轻的太太","pinyin":"nqdttnianqingdetaitai","introduction":"","introductionEn":null,"signature":"","avatar":"https://static.laohu8.com/default-avatar.jpg","hat":null,"hatId":null,"hatName":null,"vip":1,"status":2,"fanSize":2,"headSize":67,"tweetSize":28,"questionSize":0,"limitLevel":999,"accountStatus":3,"level":{"id":1,"name":"萌萌虎","nameTw":"萌萌虎","represent":"呱呱坠地","factor":"评论帖子3次或发布1条主帖(非转发)","iconColor":"3C9E83","bgColor":"A2F1D9"},"themeCounts":1,"badgeCounts":0,"badges":[],"moderator":false,"superModerator":false,"manageSymbols":null,"badgeLevel":null,"boolIsFan":false,"boolIsHead":false,"favoriteSize":0,"symbols":null,"coverImage":null,"realNameVerified":"success","userBadges":[{"badgeId":"1026c425416b44e0aac28c11a0848493-1","templateUuid":"1026c425416b44e0aac28c11a0848493","name":"Debut Tiger","description":"Join the tiger community for 500 days","bigImgUrl":"https://static.tigerbbs.com/0e4d0ca1da0456dc7894c946d44bf9ab","smallImgUrl":"https://static.tigerbbs.com/0f2f65e8ce4cfaae8db2bea9b127f58b","grayImgUrl":"https://static.tigerbbs.com/c5948a31b6edf154422335b265235809","redirectLinkEnabled":0,"redirectLink":null,"hasAllocated":1,"isWearing":0,"stamp":null,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2022.05.07","exceedPercentage":null,"individualDisplayEnabled":0,"backgroundColor":null,"fontColor":null,"individualDisplaySort":0,"categoryType":1001}],"userBadgeCount":1,"currentWearingBadge":null,"individualDisplayBadges":null,"crmLevel":1,"crmLevelSwitch":0,"location":null,"starInvestorFollowerNum":0,"starInvestorFlag":false,"starInvestorOrderShareNum":0,"subscribeStarInvestorNum":0,"ror":null,"winRationPercentage":null,"showRor":false,"investmentPhilosophy":null,"starInvestorSubscribeFlag":false},"page":1,"watchlist":null,"tweetList":[{"id":399342760,"gmtCreate":1605968612850,"gmtModify":1704963264926,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/399342760","repostId":"2085230671","repostType":4,"isVote":1,"tweetType":1,"viewCount":2038,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":399342406,"gmtCreate":1605968573005,"gmtModify":1704963264598,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/399342406","repostId":"2085237064","repostType":4,"repost":{"id":"2085237064","kind":"news","pubTimestamp":1605923700,"share":"https://ttm.financial/m/news/2085237064?lang=en_US&edition=fundamental","pubTime":"2020-11-21 09:55","market":"us","language":"zh","title":"He Xiaopeng, CEO of XPeng Motors, suspected of responding to rumors of stealing technology from a distance","url":"https://stock-news.laohu8.com/highlight/detail?id=2085237064","media":"36氪","summary":"今日小鹏汽车CEO何小鹏发布朋友圈,疑似隔空回应马斯克指小鹏偷窃技术。何小鹏称,“看来昨天我们发布的包含激光雷达的小鹏下一代自动驾驶架构,让西边的某人很不爽,连续用pigu发声。我想说的是,造谣早就证明是无法打败任何竞争对手的,明年开始,在中国的自动驾驶你、要有思想准备被我们打的找不着东,至于国际,我们会相遇的。”\n\n\n\n海量资讯、精准解读,尽在新浪财经APP\n\n责任编辑:尹悦","content":"<p>Nov. 21, today<a href=\"https://laohu8.com/S/XPEV\">XPeng Motors</a>CEO He Xiaopeng released a circle of friends, suspected of responding to Musk's reference to XPeng stealing technology from a distance. He Xiaopeng said, \"It seems that the next-generation autonomous driving architecture of XPeng containing lidar we released yesterday made someone in the west very unhappy, and they continued to speak with pigu. What I want to say is that rumors have long proved that they can't defeat any competitors. Starting next year, you should be mentally prepared to be beaten by us in autonomous driving in China. As for the international community, we will meet.\"</p><p><img src=\"https://static.tigerbbs.com/c4a0ea4c2c3f7d8ebc73053b75126ed3\" tg-width=\"550\" tg-height=\"483\"></p>","source":"sina","collect":0,"html":"<!DOCTYPE html>\n<html>\n<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\" />\n<meta name=\"viewport\" content=\"width=device-width,initial-scale=1.0,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no\"/>\n<meta name=\"format-detection\" content=\"telephone=no,email=no,address=no\" />\n<title>He Xiaopeng, CEO of XPeng Motors, suspected of responding to rumors of stealing technology from a distance</title>\n<style type=\"text/css\">\na,abbr,acronym,address,applet,article,aside,audio,b,big,blockquote,body,canvas,caption,center,cite,code,dd,del,details,dfn,div,dl,dt,\nem,embed,fieldset,figcaption,figure,footer,form,h1,h2,h3,h4,h5,h6,header,hgroup,html,i,iframe,img,ins,kbd,label,legend,li,mark,menu,nav,\nobject,ol,output,p,pre,q,ruby,s,samp,section,small,span,strike,strong,sub,summary,sup,table,tbody,td,tfoot,th,thead,time,tr,tt,u,ul,var,video{ font:inherit;margin:0;padding:0;vertical-align:baseline;border:0 }\nbody{ font-size:16px; line-height:1.5; color:#999; background:transparent; }\n.wrapper{ overflow:hidden;word-break:break-all;padding:10px; }\nh1,h2{ font-weight:normal; line-height:1.35; margin-bottom:.6em; }\nh3,h4,h5,h6{ line-height:1.35; margin-bottom:1em; }\nh1{ font-size:24px; }\nh2{ font-size:20px; }\nh3{ font-size:18px; }\nh4{ font-size:16px; }\nh5{ font-size:14px; }\nh6{ font-size:12px; }\np,ul,ol,blockquote,dl,table{ margin:1.2em 0; }\nul,ol{ margin-left:2em; }\nul{ list-style:disc; }\nol{ list-style:decimal; }\nli,li p{ margin:10px 0;}\nimg{ max-width:100%;display:block;margin:0 auto 1em; }\nblockquote{ color:#B5B2B1; border-left:3px solid #aaa; padding:1em; }\nstrong,b{font-weight:bold;}\nem,i{font-style:italic;}\ntable{ width:100%;border-collapse:collapse;border-spacing:1px;margin:1em 0;font-size:.9em; }\nth,td{ padding:5px;text-align:left;border:1px solid #aaa; }\nth{ font-weight:bold;background:#5d5d5d; }\n.symbol-link{font-weight:bold;}\n/* header{ border-bottom:1px solid #494756; } */\n.title{ margin:0 0 8px;line-height:1.3;color:#ddd; }\n.meta {color:#5e5c6d;font-size:13px;margin:0 0 .5em; }\na{text-decoration:none; color:#2a4b87;}\n.meta .head { display: inline-block; overflow: hidden}\n.head .h-thumb { width: 30px; height: 30px; margin: 0; padding: 0; border-radius: 50%; float: left;}\n.head .h-content { margin: 0; padding: 0 0 0 9px; float: left;}\n.head .h-name {font-size: 13px; color: #eee; margin: 0;}\n.head .h-time {font-size: 12.5px; color: #7E829C; margin: 0;}\n.small {font-size: 12.5px; display: inline-block; transform: scale(0.9); -webkit-transform: scale(0.9); transform-origin: left; -webkit-transform-origin: left;}\n.smaller {font-size: 12.5px; display: inline-block; transform: scale(0.8); -webkit-transform: scale(0.8); transform-origin: left; -webkit-transform-origin: left;}\n.bt-text {font-size: 12px;margin: 1.5em 0 0 0}\n.bt-text p {margin: 0}\n</style>\n</head>\n<body>\n<div class=\"wrapper\">\n<header>\n<h2 class=\"title\">\nHe Xiaopeng, CEO of XPeng Motors, suspected of responding to rumors of stealing technology from a distance\n</h2>\n<h4 class=\"meta\">\n<p class=\"head\">\n<strong class=\"h-name small\">36氪</strong><span class=\"h-time small\">2020-11-21 09:55</span>\n</p>\n</h4>\n</header>\n<article>\n<p>Nov. 21, today<a href=\"https://laohu8.com/S/XPEV\">XPeng Motors</a>CEO He Xiaopeng released a circle of friends, suspected of responding to Musk's reference to XPeng stealing technology from a distance. He Xiaopeng said, \"It seems that the next-generation autonomous driving architecture of XPeng containing lidar we released yesterday made someone in the west very unhappy, and they continued to speak with pigu. What I want to say is that rumors have long proved that they can't defeat any competitors. Starting next year, you should be mentally prepared to be beaten by us in autonomous driving in China. As for the international community, we will meet.\"</p><p><img src=\"https://static.tigerbbs.com/c4a0ea4c2c3f7d8ebc73053b75126ed3\" tg-width=\"550\" tg-height=\"483\"></p>\n<div class=\"bt-text\">\n\n\n<p> source:<a href=\"https://finance.sina.com.cn/chanjing/gsnews/2020-11-21/doc-iiznctke2525746.shtml\">36氪</a></p>\n\n\n</div>\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"https://static.tigerbbs.com/4aa4c2246df411f68e16a21f6d365311","relate_stocks":{"XPEV":"小鹏汽车"},"source_url":"https://finance.sina.com.cn/chanjing/gsnews/2020-11-21/doc-iiznctke2525746.shtml","is_english":false,"share_image_url":"https://static.laohu8.com/b0d1b7e8843deea78cc308b15114de44","article_id":"2085237064","content_text":"11月21日讯,今日小鹏汽车CEO何小鹏发布朋友圈,疑似隔空回应马斯克指小鹏偷窃技术。何小鹏称,“看来昨天我们发布的包含激光雷达的小鹏下一代自动驾驶架构,让西边的某人很不爽,连续用pigu发声。我想说的是,造谣早就证明是无法打败任何竞争对手的,明年开始,在中国的自动驾驶你、要有思想准备被我们打的找不着东,至于国际,我们会相遇的。”","news_type":1,"symbols_score_info":{"XPEV":0.9}},"isVote":1,"tweetType":1,"viewCount":1954,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":399342238,"gmtCreate":1605968545524,"gmtModify":1704963264435,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/399342238","repostId":"1123040704","repostType":4,"repost":{"id":"1123040704","kind":"news","pubTimestamp":1605928876,"share":"https://ttm.financial/m/news/1123040704?lang=en_US&edition=fundamental","pubTime":"2020-11-21 11:21","market":"us","language":"zh","title":"The book return of True Perfect Diary is nearly 1.4 billion US dollars. Can it redeem the angel investment winter?","url":"https://stock-news.laohu8.com/highlight/detail?id=1123040704","media":"投中网","summary":"投资完美日记对真格基金来说意义重大,是可以记在这家机构投资历史上,乃至中国天使投资历史上的案例。\n\n完美日记创办四年即上市,又是个标准的风险投资的好故事。但从不同的视角出发,看到意义也不尽相同。\n从完","content":"<p><b>The perfect investment diary is of great significance to Zhen Fund, which can be recorded in the investment history of this institution and even in the history of Chinese angel investment.</b>Perfect Diary went public four years after its founding, and it is another good story of standard venture capital. But from different perspectives, the meaning is not the same.</p><p>From the perspective of the investment institutions behind Perfect Diary, Hillhouse and Hony are a successful and efficient attempt by PE institutions to invest in an earlier stage. For Gao Rong, \"investing in round A\" and \"new consumption\" are standard VC good story labels; But it is of great significance to Zhenge, which can be recorded in the investment history of this institution and even in the history of Chinese angel investment.</p><p>First, throw early.</p><p>Perfect Diary was established in August 2016. On December 30th of that year, it won the angel round financing of Hony and Zhen Fund.</p><p>In March 2017, Perfect Diary Taobao Store was officially launched; Five months later, Perfect Diary Taobao Store was upgraded to Tmall Store. It can be said that when Zhen Fund invested in 2016, it was only in the investment \"concept\" stage, and there was no clear product launch.</p><p>From the perspective of new consumer brand investment, CVSource investment data shows that in the 10 years from 2006 to 2016, China's makeup market was dominated by European and American brands, and the national hot flashes were beginning to show their forefront, and local brands were in their infancy. Among them, there were 106 total investment cases in the past 10 years, with an overall upward trend: 35 cases in 2015, and the beauty industry achieved a concentrated outbreak; In 2016, there were 16 cases, and the beauty industry resumed steady growth.</p><p>It is not difficult to see that ZhenFund has stepped on the wave of industry outbreak and selected the right enterprises with great potential.</p><p>Secondly, the multiple is high.</p><p>According to the Perfect Diary prospectus, as of the pre-IPO,<a href=\"https://laohu8.com/S/YSG\">Yixian E-commerce</a>The top three institutional shareholders are Hillhouse Capital, Zhen Fund and Gaorong Capital. Among them, ZhenFund holds 10.5% of the shares.</p><p><img src=\"https://static.tigerbbs.com/8049b0840698be7b5ba4a3e87a7e4194\" tg-width=\"920\" tg-height=\"594\"></p><p>By analyzing how the 10.5% shares came from, we can more clearly see the advancement and retreat, gains and losses of Zhenge at each node in this case.</p><p><img src=\"https://static.tigerbbs.com/f070cc17f655b95aa59e1751cbdfbfd4\" tg-width=\"1038\" tg-height=\"183\"></p><p>Drawing: Zheng Xuan</p><p>The investment amount of the angel round, before 36kr, had a report caliber of 1 million yuan, and a certain database also had a statement of 300,000 US dollars. However, based on the prospectus, the data was about 800,000 US dollars from three rounds of seed, A-1 and A-2.</p><p>In addition, in the round of financing in July 2019, Zhenge invested US$18 million, which is also a rare investment behavior with obviously high amount and risk and pressure for an angel fund. In addition, in more than a year from 2019 to 2020, Zhenge also cashed out 8.6 million, 43 million and 25.3 million ordinary shares in three installments, accounting for about 4% of the total share capital before IPO. However, the valuation of Perfect Diary soared in 2020, and the exit price at this time was obviously much higher than the follow-up price in 2019.</p><p>So it should be said that at this node in 2019, is the real retreat? Is staying? Or follow up? It was the decisive decision they made on the case of Perfect Diary.</p><p>More broadly, how can angel investment institutions save themselves or win back the right to speak in the market? The case of true investment perfect diary actually provides a core logic: vote early, hold it, keep up and retire accurately.</p><p>The share price of Perfect Diary's IPO soared two days after its listing. As of press time, the market value exceeded 13 billion USD, that is to say, the book market value of Zhenge Shareholding was nearly 1.4 billion USD. For this angel fund, Xu Xiaoping and Fang Aizhi, such a book return must be unprecedented.</p><p>At that time, Jumei Youpin, which landed on the New York Stock Exchange in 2014, brought more than 800 times of book return to Xu Xiaoping (founder of Zhen Fund). In the media caliber at that time, this was Xu Xiaoping's investment with the highest return \"so far\".</p><p>China Investment Network has sorted out all the IPO projects of Zhen Fund so far. In addition to Perfect Diary, more than 700 IPO projects have been sold by Zhen Fund<a href=\"https://laohu8.com/S/V03.SI\">Startups</a>Among them, 7 companies have been successfully listed. From the perspective of performance return, Zhen Fund's exit return rate on Jumei Youpin project reached 79,071.3% (about 790 times), ranking the highest.</p><p><img src=\"https://static.tigerbbs.com/f2982c1ec7f3a67312226bf1e91c9420\" tg-width=\"1080\" tg-height=\"389\"></p><p>Source: CVSource hit data</p><p>Apparently, after six years, this amazing return myth has been shattered in one fell swoop by Perfect Diary. It is foreseeable that in the future superstar projects of ZhenFund, it must be the perfect diary, not Jumei Youpin, that can redefine ZhenFund.</p><p>Looking at the big picture, if this return figure of nearly $1.4 billion is placed in the whole angel investment market, it is also a wonderful case.</p><p>Especially in the past two years, under the pessimistic tone of capital winter, the volume of angel investment is getting smaller and smaller. In the past, the investment enthusiasm of \"becoming famous in World War I\" and \"fast-in and fast-out\" carried by angel investors has been gradually swallowed up by reality, and there are very few exciting moments with high returns and high progress.</p><p>\"Some peers have gone to FA, some have chased blockchain, and some have completely transformed and left this industry.\" In 2019, when talking about angel investment, some investors described the current situation of the industry.</p><p>Nowadays, Perfect Diary, a legendary case of high yield created by angel investment, has been put on the table. The wealth effect has always been the explosive point and catalyst of the market. Next, should the market re-examine the value of angel investment? We can even pay further attention to this question: Can this phenomenal case, which may return more than $1 billion, redeem the cold winter of angel investment?</p>","source":"tzw","collect":0,"html":"<!DOCTYPE html>\n<html>\n<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\" />\n<meta name=\"viewport\" content=\"width=device-width,initial-scale=1.0,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no\"/>\n<meta name=\"format-detection\" content=\"telephone=no,email=no,address=no\" />\n<title>The book return of True Perfect Diary is nearly 1.4 billion US dollars. Can it redeem the angel investment winter?</title>\n<style type=\"text/css\">\na,abbr,acronym,address,applet,article,aside,audio,b,big,blockquote,body,canvas,caption,center,cite,code,dd,del,details,dfn,div,dl,dt,\nem,embed,fieldset,figcaption,figure,footer,form,h1,h2,h3,h4,h5,h6,header,hgroup,html,i,iframe,img,ins,kbd,label,legend,li,mark,menu,nav,\nobject,ol,output,p,pre,q,ruby,s,samp,section,small,span,strike,strong,sub,summary,sup,table,tbody,td,tfoot,th,thead,time,tr,tt,u,ul,var,video{ font:inherit;margin:0;padding:0;vertical-align:baseline;border:0 }\nbody{ font-size:16px; line-height:1.5; color:#999; background:transparent; }\n.wrapper{ overflow:hidden;word-break:break-all;padding:10px; }\nh1,h2{ font-weight:normal; line-height:1.35; margin-bottom:.6em; }\nh3,h4,h5,h6{ line-height:1.35; margin-bottom:1em; }\nh1{ font-size:24px; }\nh2{ font-size:20px; }\nh3{ font-size:18px; }\nh4{ font-size:16px; }\nh5{ font-size:14px; }\nh6{ font-size:12px; }\np,ul,ol,blockquote,dl,table{ margin:1.2em 0; }\nul,ol{ margin-left:2em; }\nul{ list-style:disc; }\nol{ list-style:decimal; }\nli,li p{ margin:10px 0;}\nimg{ max-width:100%;display:block;margin:0 auto 1em; }\nblockquote{ color:#B5B2B1; border-left:3px solid #aaa; padding:1em; }\nstrong,b{font-weight:bold;}\nem,i{font-style:italic;}\ntable{ width:100%;border-collapse:collapse;border-spacing:1px;margin:1em 0;font-size:.9em; }\nth,td{ padding:5px;text-align:left;border:1px solid #aaa; }\nth{ font-weight:bold;background:#5d5d5d; }\n.symbol-link{font-weight:bold;}\n/* header{ border-bottom:1px solid #494756; } */\n.title{ margin:0 0 8px;line-height:1.3;color:#ddd; }\n.meta {color:#5e5c6d;font-size:13px;margin:0 0 .5em; }\na{text-decoration:none; color:#2a4b87;}\n.meta .head { display: inline-block; overflow: hidden}\n.head .h-thumb { width: 30px; height: 30px; margin: 0; padding: 0; border-radius: 50%; float: left;}\n.head .h-content { margin: 0; padding: 0 0 0 9px; float: left;}\n.head .h-name {font-size: 13px; color: #eee; margin: 0;}\n.head .h-time {font-size: 12.5px; color: #7E829C; margin: 0;}\n.small {font-size: 12.5px; display: inline-block; transform: scale(0.9); -webkit-transform: scale(0.9); transform-origin: left; -webkit-transform-origin: left;}\n.smaller {font-size: 12.5px; display: inline-block; transform: scale(0.8); -webkit-transform: scale(0.8); transform-origin: left; -webkit-transform-origin: left;}\n.bt-text {font-size: 12px;margin: 1.5em 0 0 0}\n.bt-text p {margin: 0}\n</style>\n</head>\n<body>\n<div class=\"wrapper\">\n<header>\n<h2 class=\"title\">\nThe book return of True Perfect Diary is nearly 1.4 billion US dollars. Can it redeem the angel investment winter?\n</h2>\n<h4 class=\"meta\">\n<p class=\"head\">\n<strong class=\"h-name small\">投中网</strong><span class=\"h-time small\">2020-11-21 11:21</span>\n</p>\n</h4>\n</header>\n<article>\n<p><b>The perfect investment diary is of great significance to Zhen Fund, which can be recorded in the investment history of this institution and even in the history of Chinese angel investment.</b>Perfect Diary went public four years after its founding, and it is another good story of standard venture capital. But from different perspectives, the meaning is not the same.</p><p>From the perspective of the investment institutions behind Perfect Diary, Hillhouse and Hony are a successful and efficient attempt by PE institutions to invest in an earlier stage. For Gao Rong, \"investing in round A\" and \"new consumption\" are standard VC good story labels; But it is of great significance to Zhenge, which can be recorded in the investment history of this institution and even in the history of Chinese angel investment.</p><p>First, throw early.</p><p>Perfect Diary was established in August 2016. On December 30th of that year, it won the angel round financing of Hony and Zhen Fund.</p><p>In March 2017, Perfect Diary Taobao Store was officially launched; Five months later, Perfect Diary Taobao Store was upgraded to Tmall Store. It can be said that when Zhen Fund invested in 2016, it was only in the investment \"concept\" stage, and there was no clear product launch.</p><p>From the perspective of new consumer brand investment, CVSource investment data shows that in the 10 years from 2006 to 2016, China's makeup market was dominated by European and American brands, and the national hot flashes were beginning to show their forefront, and local brands were in their infancy. Among them, there were 106 total investment cases in the past 10 years, with an overall upward trend: 35 cases in 2015, and the beauty industry achieved a concentrated outbreak; In 2016, there were 16 cases, and the beauty industry resumed steady growth.</p><p>It is not difficult to see that ZhenFund has stepped on the wave of industry outbreak and selected the right enterprises with great potential.</p><p>Secondly, the multiple is high.</p><p>According to the Perfect Diary prospectus, as of the pre-IPO,<a href=\"https://laohu8.com/S/YSG\">Yixian E-commerce</a>The top three institutional shareholders are Hillhouse Capital, Zhen Fund and Gaorong Capital. Among them, ZhenFund holds 10.5% of the shares.</p><p><img src=\"https://static.tigerbbs.com/8049b0840698be7b5ba4a3e87a7e4194\" tg-width=\"920\" tg-height=\"594\"></p><p>By analyzing how the 10.5% shares came from, we can more clearly see the advancement and retreat, gains and losses of Zhenge at each node in this case.</p><p><img src=\"https://static.tigerbbs.com/f070cc17f655b95aa59e1751cbdfbfd4\" tg-width=\"1038\" tg-height=\"183\"></p><p>Drawing: Zheng Xuan</p><p>The investment amount of the angel round, before 36kr, had a report caliber of 1 million yuan, and a certain database also had a statement of 300,000 US dollars. However, based on the prospectus, the data was about 800,000 US dollars from three rounds of seed, A-1 and A-2.</p><p>In addition, in the round of financing in July 2019, Zhenge invested US$18 million, which is also a rare investment behavior with obviously high amount and risk and pressure for an angel fund. In addition, in more than a year from 2019 to 2020, Zhenge also cashed out 8.6 million, 43 million and 25.3 million ordinary shares in three installments, accounting for about 4% of the total share capital before IPO. However, the valuation of Perfect Diary soared in 2020, and the exit price at this time was obviously much higher than the follow-up price in 2019.</p><p>So it should be said that at this node in 2019, is the real retreat? Is staying? Or follow up? It was the decisive decision they made on the case of Perfect Diary.</p><p>More broadly, how can angel investment institutions save themselves or win back the right to speak in the market? The case of true investment perfect diary actually provides a core logic: vote early, hold it, keep up and retire accurately.</p><p>The share price of Perfect Diary's IPO soared two days after its listing. As of press time, the market value exceeded 13 billion USD, that is to say, the book market value of Zhenge Shareholding was nearly 1.4 billion USD. For this angel fund, Xu Xiaoping and Fang Aizhi, such a book return must be unprecedented.</p><p>At that time, Jumei Youpin, which landed on the New York Stock Exchange in 2014, brought more than 800 times of book return to Xu Xiaoping (founder of Zhen Fund). In the media caliber at that time, this was Xu Xiaoping's investment with the highest return \"so far\".</p><p>China Investment Network has sorted out all the IPO projects of Zhen Fund so far. In addition to Perfect Diary, more than 700 IPO projects have been sold by Zhen Fund<a href=\"https://laohu8.com/S/V03.SI\">Startups</a>Among them, 7 companies have been successfully listed. From the perspective of performance return, Zhen Fund's exit return rate on Jumei Youpin project reached 79,071.3% (about 790 times), ranking the highest.</p><p><img src=\"https://static.tigerbbs.com/f2982c1ec7f3a67312226bf1e91c9420\" tg-width=\"1080\" tg-height=\"389\"></p><p>Source: CVSource hit data</p><p>Apparently, after six years, this amazing return myth has been shattered in one fell swoop by Perfect Diary. It is foreseeable that in the future superstar projects of ZhenFund, it must be the perfect diary, not Jumei Youpin, that can redefine ZhenFund.</p><p>Looking at the big picture, if this return figure of nearly $1.4 billion is placed in the whole angel investment market, it is also a wonderful case.</p><p>Especially in the past two years, under the pessimistic tone of capital winter, the volume of angel investment is getting smaller and smaller. In the past, the investment enthusiasm of \"becoming famous in World War I\" and \"fast-in and fast-out\" carried by angel investors has been gradually swallowed up by reality, and there are very few exciting moments with high returns and high progress.</p><p>\"Some peers have gone to FA, some have chased blockchain, and some have completely transformed and left this industry.\" In 2019, when talking about angel investment, some investors described the current situation of the industry.</p><p>Nowadays, Perfect Diary, a legendary case of high yield created by angel investment, has been put on the table. The wealth effect has always been the explosive point and catalyst of the market. Next, should the market re-examine the value of angel investment? We can even pay further attention to this question: Can this phenomenal case, which may return more than $1 billion, redeem the cold winter of angel investment?</p>\n<div class=\"bt-text\">\n\n\n<p> source:<a href=\"https://mp.weixin.qq.com/s/t6Uj0_grllpRto4xxRGYvQ\">投中网</a></p>\n\n\n</div>\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"https://static.tigerbbs.com/769dc8865e0dc507ddea5079c2429001","relate_stocks":{"YSG":"逸仙电商"},"source_url":"https://mp.weixin.qq.com/s/t6Uj0_grllpRto4xxRGYvQ","is_english":false,"share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","article_id":"1123040704","content_text":"投资完美日记对真格基金来说意义重大,是可以记在这家机构投资历史上,乃至中国天使投资历史上的案例。\n\n完美日记创办四年即上市,又是个标准的风险投资的好故事。但从不同的视角出发,看到意义也不尽相同。\n从完美日记背后投资机构来看,高瓴和弘毅是PE机构向更早期投资的一次成功且有效率的尝试,对高榕来说,“投A轮”、“新消费”都是标准的VC好故事标签;但对真格来说则意义重大,是可以记在这家机构投资历史上,乃至中国天使投资历史上的案例。\n首先,投得早。\n完美日记成立于2016年8月,当年12月30日,便斩获了弘毅和真格基金的天使轮融资。\n2017年3月,完美日记淘宝店才正式上线;5个月后,完美日记淘宝店升级为天猫店铺。可以说,2016年真格基金投资之时,还只是在投资“概念”阶段,并未有明确的产品推出。\n从新消费品牌投资角度,CVSource投中数据显示,2006至2016的10年间,中国彩妆市场由欧美品牌主导,国潮热初露锋芒,本土品牌处于萌芽阶段。其中,10年间总投资案例为106起,整体呈上升趋势:2015年为35起,美妆行业实现集中爆发;2016年为16起,美妆行业恢复稳步增长。\n不难看出,真格基金踩准了行业爆发的浪潮,并选对了潜力巨大的企业。\n其次,倍数高。\n根据完美日记招股书,截至IPO前,逸仙电商占股前三的机构股东为高瓴资本、真格基金及高榕资本。这其中,真格基金持股比例为10.5%。\n\n分析一下这10.5%的股份怎么来的,就能更清楚地看到真格在这个案子上每个节点的进退与得失。\n\n制图:郑玄\n天使轮的投资额度,之前36kr曾有个报道口径是100万元,某家数据库也有个30万美金的说法,但基于招股书的话,数据是从种子、A-1、A-2三轮放了约80万美元进来。\n此外在2019年7月的那一轮融资,真格跟投了1800万美元,这对一家天使基金来说,也是个罕见的、明显额度偏高、顶着风险和压力的投资行为。此外,在2019-2020一年多的时间里,真格也分三次分别套现了860万、4300万以及2530万普通股,约占IPO前总股本的4%,但2020年完美日记估值猛涨,此时的退出价格显然高过2019年跟投价格很多。\n所以应该这么说,2019年这个节点,真格是退?是留?还是跟进?这是他们对完美日记这个案子做的决定性决策。\n往更大了说,天使投资机构如何自救抑或赢回市场话语权?真格投资完美日记的案例其实提供了一个核心逻辑:投得早、拿得住、跟得上、退得准。\n完美日记IPO上市两天股价猛涨,截至发稿,市值超过130亿美元,也就是说,真格持股的账面市值近14亿美元,对于这家天使基金和徐小平、方爱之来说,这样的账面回报一定是前所未有。\n当年,2014年登陆纽交所的聚美优品曾给徐小平(真格基金创始人)带来了800多倍的账面回报。在当时的媒体口径里,这是徐小平“迄今为止”回报最高的一笔投资。\n投中网梳理了下真格基金至今为止所有的IPO项目,除了完美日记之外,在真格基金出手的700余家创业公司中,成功上市的企业共有7家。从业绩回报角度,真格基金在聚美优品项目上的退出回报率达79071.3%(约790倍),排名最高。\n\n来源:CVSource投中数据\n显然,时隔六年,这一惊人的回报神话被完美日记一举打破了。可以预见的是,在日后真格基金的超明星项目里,能够重新定义真格基金的一定是完美日记,而非聚美优品。\n往大了看,若将这近14亿美元的回报数字置于整个天使投资市场上,也不失为一个精彩的好案例。\n特别是近两年,在资本寒冬的悲观论调下,天使投资的声量相比越来越小,过往天使投资人怀揣着的“一战成名”、“快进快出”的投资热情也逐渐被现实吞噬,那些高倍回报、高歌猛进的激动时刻更是寥寥无几。\n“有些同行去做FA了,有些去追区块链了,还有一些彻底转型离开了这个行业。”2019年,在谈到天使投资时,有投资人就如此描述行业现状。\n如今,完美日记这一由天使投资创造的高收益传奇案例已经摆在了台面上,财富效应从来都是市场的爆点和催化剂,接下来,市场是否应该重新审视一下天使投资的价值?甚至可以进一步关注这个问题:这单可能超过10亿美金回报的现象级案例,能够救赎天使投资的凛冽寒冬吗?","news_type":1,"symbols_score_info":{"YSG":0.9}},"isVote":1,"tweetType":1,"viewCount":2062,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":399342665,"gmtCreate":1605968520257,"gmtModify":1704963263944,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/399342665","repostId":"2085023000","repostType":4,"repost":{"id":"2085023000","kind":"news","pubTimestamp":1605933480,"share":"https://ttm.financial/m/news/2085023000?lang=en_US&edition=fundamental","pubTime":"2020-11-21 12:38","market":"us","language":"zh","title":"Samsung's Three Challenges Against TSMC","url":"https://stock-news.laohu8.com/highlight/detail?id=2085023000","media":"半导体行业观察","summary":"最近三星代工挑战台积电的话题引起我很大的兴趣,在过去三十年与代工厂密切合作中,我有自己不同的看法。正文:彭博社最近发布一篇名为“三星2020年超越台积电,加剧芯片战”的文章,这是“三星1160亿美元计划挑战台积电”文章的后续。据作者表述,三星和台积电将在3nm节点展开业务竞争,这意味着在2022年大批量生产。IBM使用了三星授权的GF 14nm,因此他们将继续使用三星7nm。","content":"<p><b>Author's Preface:</b>Semiconductor foundry has recently become a hot topic in semiconductor industry news, which is understandable. After all, the miracle of chips can't be separated from wafer foundry. Unfortunately, most of the \"uplifting\" news has been exaggerated. recently<a href=\"https://laohu8.com/S/SMSN.UK\">Samsung</a>Foundry challenges<a href=\"https://laohu8.com/S/TSM\">TSMC</a>The topic is of great interest to me, and in working closely with foundries over the past thirty years, I have my own different opinions.<b>Text:</b>Bloomberg recently published an article titled \"Samsung surpasses TSMC in 2020, intensifies chip war\" (Samsung Intensifies Chip Wars With Bet ItCan Catch TSMC by 2022) The article, which is \"Samsung's $116 billion plan to challenge TSMC\" (Samsung takes Another Step in $116 Billion Plan to Take on TSMC) Follow-up to the article. The authors of the two articles are the same person, working in Bloomberg Korea. Although they have no background and experience in semiconductor education, they must know Samsung and can get in touch with Samsung's senior executives directly. It can be judged that this kind of article comes from Samsung's mouth.</p><p>According to the authors, Samsung and TSMC will compete for business at 3nm nodes, which means high-volume production (HVM) in 2022. The very important point here is that TSMC 3nm and Samsung 3nm will be using very different technologies. TSMC is expanding their 5nm FinFET-based process, while Samsung has released a new process technology (GAA) at 3nm.</p><p>In my opinion, Samsung has three challenges in the 3nm process:</p><p>The first challenge is ecology, with TSMC using a tried-and-true technology, supported by a vast ecosystem of EDA, IP and service companies. Hundreds of verified IPs will be immediately available to TSMC 3nm customers, while Samsung must build a new GAA ecosystem. It certainly isn't that easy.</p><p>The second challenge is trust. There are many forms of foundry trust: trust that your IP is safe and reliable, trust that the foundry will not compete unfairly with you, trust that the foundry can provide good PPA (power/performance/volume) technology.</p><p>The third challenge is yield. GAA is a new process technology, and Samsung has been plagued by yield problems. The first person to eat crabs in new technology deserves respect and a symbol of glory, and I personally pay deep respect to Samsung's technical strength. But I question Samsung's process of introducing a new process and achieving mass production, which must be difficult in the middle. Customers must also trust the foundry to fulfill its promise and provide good wafers to meet the agreed chip delivery schedule.</p><p>Finally, the author of that article suggested: \"If Samsung succeeds, it will break<a href=\"https://laohu8.com/S/AAPL\">Apple</a>、<a href=\"https://laohu8.com/S/AMD\">AMD</a>Rely on TSMC's balance and grab some orders. These companies are extremely dependent on TSMC. \"</p><p>Apparently, Apple and AMD today are TSMC's exclusive customers. This exclusivity has brought Apple and AMD into the circle within TSMC, where collaboration is at its highest level. Samsung's big customer is<a href=\"https://laohu8.com/S/NVDA\">NVIDIA</a>,<a href=\"https://laohu8.com/S/QCOM\">Qualcomm</a>And<a href=\"https://laohu8.com/S/IBM\">IBM</a>None of them are in TSMC's internal cooperation circle.</p><p>From an insider's point of view, Qualcomm and Nvidia used to be TSMC's best partners, but Qualcomm competes with Apple, while Nvidia competes with AMD, so there is a disagreement in the ranks. IBM used the Samsung licensed GF 14nm, so they will continue to use the Samsung 7nm.</p><p>Summary: Can Samsung foundry really compete with TSMC? Sorry, not today, not 3nm. TSMC's 3nm PDK (process design kit) has been used in top semiconductor companies worldwide and is fully supported by the ecosystem. On the other hand, the Samsung 3nm PDK and the tools and IPs it supports are still evolving. Of course just my observations, experiences and opinions.</p>","source":"bdthygc","collect":0,"html":"<!DOCTYPE html>\n<html>\n<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\" />\n<meta name=\"viewport\" content=\"width=device-width,initial-scale=1.0,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no\"/>\n<meta name=\"format-detection\" content=\"telephone=no,email=no,address=no\" />\n<title>Samsung's Three Challenges Against TSMC</title>\n<style type=\"text/css\">\na,abbr,acronym,address,applet,article,aside,audio,b,big,blockquote,body,canvas,caption,center,cite,code,dd,del,details,dfn,div,dl,dt,\nem,embed,fieldset,figcaption,figure,footer,form,h1,h2,h3,h4,h5,h6,header,hgroup,html,i,iframe,img,ins,kbd,label,legend,li,mark,menu,nav,\nobject,ol,output,p,pre,q,ruby,s,samp,section,small,span,strike,strong,sub,summary,sup,table,tbody,td,tfoot,th,thead,time,tr,tt,u,ul,var,video{ font:inherit;margin:0;padding:0;vertical-align:baseline;border:0 }\nbody{ font-size:16px; line-height:1.5; color:#999; background:transparent; }\n.wrapper{ overflow:hidden;word-break:break-all;padding:10px; }\nh1,h2{ font-weight:normal; line-height:1.35; margin-bottom:.6em; }\nh3,h4,h5,h6{ line-height:1.35; margin-bottom:1em; }\nh1{ font-size:24px; }\nh2{ font-size:20px; }\nh3{ font-size:18px; }\nh4{ font-size:16px; }\nh5{ font-size:14px; }\nh6{ font-size:12px; }\np,ul,ol,blockquote,dl,table{ margin:1.2em 0; }\nul,ol{ margin-left:2em; }\nul{ list-style:disc; }\nol{ list-style:decimal; }\nli,li p{ margin:10px 0;}\nimg{ max-width:100%;display:block;margin:0 auto 1em; }\nblockquote{ color:#B5B2B1; border-left:3px solid #aaa; padding:1em; }\nstrong,b{font-weight:bold;}\nem,i{font-style:italic;}\ntable{ width:100%;border-collapse:collapse;border-spacing:1px;margin:1em 0;font-size:.9em; }\nth,td{ padding:5px;text-align:left;border:1px solid #aaa; }\nth{ font-weight:bold;background:#5d5d5d; }\n.symbol-link{font-weight:bold;}\n/* header{ border-bottom:1px solid #494756; } */\n.title{ margin:0 0 8px;line-height:1.3;color:#ddd; }\n.meta {color:#5e5c6d;font-size:13px;margin:0 0 .5em; }\na{text-decoration:none; color:#2a4b87;}\n.meta .head { display: inline-block; overflow: hidden}\n.head .h-thumb { width: 30px; height: 30px; margin: 0; padding: 0; border-radius: 50%; float: left;}\n.head .h-content { margin: 0; padding: 0 0 0 9px; float: left;}\n.head .h-name {font-size: 13px; color: #eee; margin: 0;}\n.head .h-time {font-size: 12.5px; color: #7E829C; margin: 0;}\n.small {font-size: 12.5px; display: inline-block; transform: scale(0.9); -webkit-transform: scale(0.9); transform-origin: left; -webkit-transform-origin: left;}\n.smaller {font-size: 12.5px; display: inline-block; transform: scale(0.8); -webkit-transform: scale(0.8); transform-origin: left; -webkit-transform-origin: left;}\n.bt-text {font-size: 12px;margin: 1.5em 0 0 0}\n.bt-text p {margin: 0}\n</style>\n</head>\n<body>\n<div class=\"wrapper\">\n<header>\n<h2 class=\"title\">\nSamsung's Three Challenges Against TSMC\n</h2>\n<h4 class=\"meta\">\n<p class=\"head\">\n<strong class=\"h-name small\">半导体行业观察</strong><span class=\"h-time small\">2020-11-21 12:38</span>\n</p>\n</h4>\n</header>\n<article>\n<p><b>Author's Preface:</b>Semiconductor foundry has recently become a hot topic in semiconductor industry news, which is understandable. After all, the miracle of chips can't be separated from wafer foundry. Unfortunately, most of the \"uplifting\" news has been exaggerated. recently<a href=\"https://laohu8.com/S/SMSN.UK\">Samsung</a>Foundry challenges<a href=\"https://laohu8.com/S/TSM\">TSMC</a>The topic is of great interest to me, and in working closely with foundries over the past thirty years, I have my own different opinions.<b>Text:</b>Bloomberg recently published an article titled \"Samsung surpasses TSMC in 2020, intensifies chip war\" (Samsung Intensifies Chip Wars With Bet ItCan Catch TSMC by 2022) The article, which is \"Samsung's $116 billion plan to challenge TSMC\" (Samsung takes Another Step in $116 Billion Plan to Take on TSMC) Follow-up to the article. The authors of the two articles are the same person, working in Bloomberg Korea. Although they have no background and experience in semiconductor education, they must know Samsung and can get in touch with Samsung's senior executives directly. It can be judged that this kind of article comes from Samsung's mouth.</p><p>According to the authors, Samsung and TSMC will compete for business at 3nm nodes, which means high-volume production (HVM) in 2022. The very important point here is that TSMC 3nm and Samsung 3nm will be using very different technologies. TSMC is expanding their 5nm FinFET-based process, while Samsung has released a new process technology (GAA) at 3nm.</p><p>In my opinion, Samsung has three challenges in the 3nm process:</p><p>The first challenge is ecology, with TSMC using a tried-and-true technology, supported by a vast ecosystem of EDA, IP and service companies. Hundreds of verified IPs will be immediately available to TSMC 3nm customers, while Samsung must build a new GAA ecosystem. It certainly isn't that easy.</p><p>The second challenge is trust. There are many forms of foundry trust: trust that your IP is safe and reliable, trust that the foundry will not compete unfairly with you, trust that the foundry can provide good PPA (power/performance/volume) technology.</p><p>The third challenge is yield. GAA is a new process technology, and Samsung has been plagued by yield problems. The first person to eat crabs in new technology deserves respect and a symbol of glory, and I personally pay deep respect to Samsung's technical strength. But I question Samsung's process of introducing a new process and achieving mass production, which must be difficult in the middle. Customers must also trust the foundry to fulfill its promise and provide good wafers to meet the agreed chip delivery schedule.</p><p>Finally, the author of that article suggested: \"If Samsung succeeds, it will break<a href=\"https://laohu8.com/S/AAPL\">Apple</a>、<a href=\"https://laohu8.com/S/AMD\">AMD</a>Rely on TSMC's balance and grab some orders. These companies are extremely dependent on TSMC. \"</p><p>Apparently, Apple and AMD today are TSMC's exclusive customers. This exclusivity has brought Apple and AMD into the circle within TSMC, where collaboration is at its highest level. Samsung's big customer is<a href=\"https://laohu8.com/S/NVDA\">NVIDIA</a>,<a href=\"https://laohu8.com/S/QCOM\">Qualcomm</a>And<a href=\"https://laohu8.com/S/IBM\">IBM</a>None of them are in TSMC's internal cooperation circle.</p><p>From an insider's point of view, Qualcomm and Nvidia used to be TSMC's best partners, but Qualcomm competes with Apple, while Nvidia competes with AMD, so there is a disagreement in the ranks. IBM used the Samsung licensed GF 14nm, so they will continue to use the Samsung 7nm.</p><p>Summary: Can Samsung foundry really compete with TSMC? Sorry, not today, not 3nm. TSMC's 3nm PDK (process design kit) has been used in top semiconductor companies worldwide and is fully supported by the ecosystem. On the other hand, the Samsung 3nm PDK and the tools and IPs it supports are still evolving. Of course just my observations, experiences and opinions.</p>\n<div class=\"bt-text\">\n\n\n<p> source:<a href=\"https://mp.weixin.qq.com/s/q1hnB15lQQ2cN6sWDS_Bcw\">半导体行业观察</a></p>\n\n\n</div>\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"https://static.tigerbbs.com/91dab16b33effe5ffcc01a35b561c0e3","relate_stocks":{"03145":"华夏亚洲高息股","SMSN.UK":"三星","TSM":"台积电","EWT":"台湾ETF-iShares MSCI"},"source_url":"https://mp.weixin.qq.com/s/q1hnB15lQQ2cN6sWDS_Bcw","is_english":false,"share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","article_id":"2085023000","content_text":"作者序:半导体代工近期成为了半导体行业新闻的热门,这无可厚非,毕竟芯片的奇迹离不开晶圆代工厂。不幸的是,大多数“令人振奋”的消息已经被夸大了。最近三星代工挑战台积电的话题引起我很大的兴趣,在过去三十年与代工厂密切合作中,我有自己不同的看法。\n\n正文:彭博社最近发布一篇名为“三星2020年超越台积电,加剧芯片战”(Samsung Intensifies Chip Wars With Bet ItCan Catch TSMC by 2022)的文章,这是“三星1160亿美元计划挑战台积电”(SamsungTakes Another Step in $116 Billion Plan to Take on TSMC)文章的后续。两篇文章的作者都是同一个人,在彭博社韩国区工作,虽说没有半导体教育背景和经验,但一定了解三星,且能与三星高层直接取得联系,可以下一个判断,该类文章就是出自三星之口。\n据作者表述,三星和台积电将在3nm节点展开业务竞争,这意味着在2022年大批量生产(HVM)。这里非常重要的一点是,台积电3nm和三星3nm将是采用非常不同的技术。台积电正在扩展他们的基于5nm FinFET的工艺,而三星则在3nm发布了一种新的工艺技术(GAA)。\n我认为,三星在3nm工艺上存在三个挑战:\n第一个挑战是生态,台积电使用的是一种久经考验的切实成真的技术,由EDA、IP和服务公司组成的庞大生态系统所支持。数以百计得到验证的IP将立即提供给TSMC 3nm客户,而三星必须建立一个新的GAA生态系统。这肯定没那么容易。\n第二个挑战就是信任,代工厂的信任有很多形式:相信你的IP安全可靠,相信代工厂不会与你不公平竞争,相信代工厂能提供良好的PPA(功率/性能/体积)技术。\n第三个挑战就是良率,GAA是一种新的工艺技术,且三星一直被良率问题所困扰。新技术第一个吃螃蟹的人值得尊重,也是荣耀的象征,且我本人对三星的技术实力深表敬意。但我对三星引入一种新工艺并实现大批量量产的过程产生质疑,这中间一定困难重重。客户也必须要相信代工厂能良好的履行承诺,提供好的晶圆,以满足商定的芯片交货时间表。\n最后,那篇文章的作者建议:“如果三星成功,将会打破苹果、AMD依赖台积电的平衡,而抢部分订单。这些公司极度依赖台积电。”\n显然,如今的苹果和AMD是台积电独有的客户。这种排他性使苹果和AMD进入了台积电内部的圈子,在该圈子中,合作达到了最高水平。三星的大客户是英伟达,高通和IBM,它们都不在台积电内部合作圈。\n从内部人士的角度来看,高通和英伟达曾经是台积电的最佳搭档,但高通和苹果竞争,而英伟达与AMD竞争,因此在队伍中存在分歧。IBM使用了三星授权的GF 14nm,因此他们将继续使用三星7nm。\n总结: 三星代工厂真的可以与台积电竞争吗?抱歉,不是今天,不是3nm。台积电3nm PDK(process design kit,流程设计工具包)已在全球顶级半导体公司中使用,并得到了生态系统的全面支持。另一方面,三星3nm PDK及其支持的工具和IP仍在发展。当然只是我的观察,经验和意见。","news_type":1,"symbols_score_info":{"03145":0.9,"EWT":0.9,"TSM":0.9,"SMSN.UK":0.9}},"isVote":1,"tweetType":1,"viewCount":1481,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":399342865,"gmtCreate":1605968484157,"gmtModify":1704963264107,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/399342865","repostId":"2085230767","repostType":4,"repost":{"id":"2085230767","kind":"news","pubTimestamp":1605934193,"share":"https://ttm.financial/m/news/2085230767?lang=en_US&edition=fundamental","pubTime":"2020-11-21 12:49","market":"us","language":"zh","title":"Intel's tortoise and rabbit race against AMD","url":"https://stock-news.laohu8.com/highlight/detail?id=2085230767","media":"半导体行业观察","summary":"英特尔与AMD的竞争是乌龟与野兔的竞争。这在AMD方面尤其困难,因为AMD并没有真正相当于英特尔的Ark提供单一的处理器列表,按代、类型和发布日期排序。随着近70%的单线程性能优势被英特尔的qx9650所取代,以及焦点突然回到单线程性能的王者地位,至少在这些高端CPU中是这样——AMD 21世纪的第一个黄金时代正式结束了。在整个2007年至2013年期间,英特尔都比AMD拥有单线程优势,而且通常是相当多的优势。","content":"<p><a href=\"https://laohu8.com/S/INTC\">Intel</a>With<a href=\"https://laohu8.com/S/AMD\">AMD</a>The rivalry is the rivalry between the tortoise and the hare. But which company is the tortoise? Which company is Rabbit?</p><p>The commentary battle between Intel and AMD fans has been raging over the last couple of launch cycles, with plenty of digital ink devoted to discussing which company has improved significantly over the years, or not. There's also no shortage of opinions on the raw performance of both companies' fastest processors. We thought it would be interesting to delve into the archived performance benchmarking of each company's fastest desktop/CPU to give a good idea of how each company has actually performed over the years and even to see if there are some patterns to gather or some bets to make on the future.</p><p>Before diving into the charts, let's start with a few tables – this way you can see which CPUs we use as milestones each year. During the course of our research, there were some irregularities in the data; We'll talk about those as well, talking about something that a simple chart can't show.</p><p><b>Twenty years of enthusiast calculation</b></p><p><img src=\"https://static.tigerbbs.com/582eb6403400c9c07e9f29f893cb848e\" tg-width=\"1044\" tg-height=\"1193\"></p><p>While Intel and AMD launch a large number of processors each year for different price points and target markets, we are limited to the fastest desktop or \"enthusiast\" processors each year. That means no server processors, and no high-end desktop (HEDT) processors — so we're not looking at Threadrippers or the latest model XE series Intel parts.</p><p>Even for someone like me who has been a system builder for the entire period, putting together a list like this is a huge pain, let alone matching test results. This is especially difficult with AMD, which doesn't really equal Intel's Ark to offer a single list of processors, sorted by generation, type, and release date. If you think I picked the wrong \"top processor\" of the year, let me know in the comments.</p><p>We should address some of the anomalies that appeared in the chart – first of all, no, it wasn't an oversight for AMD not to introduce new high-performance CPUs in 2014, 2015 and 2016. The 2013 FX-9590 was widely (and rightfully) criticized for its staggering 225W TDP and mediocre performance — Team Red's fastest CPU in four years of operation. During this time, AMD introduced several generations of low-power, inexpensive desktop devices, but none of them outperformed the FX-9590.</p><p>Intel also experienced several missteps between 2013 and 2017, although not as severe or lasting as long. The fifth-generation Core series essentially died in 2014, although updates to the fourth-generation Core i7 series do offer a significant performance boost. Two years later, there's technically another bandwagon-the Core i7-7700K-equipped Kaby Lake didn't really launch until January 2017. We fabricated it a bit to get Kaby Lake into the 2016 ranks, because otherwise it would have disappeared entirely, with Coffee Lake and the i7-8700K showing up later that year.</p><p><b>AMD Athlon The rise and fall of: 2001-2007</b></p><p><img src=\"https://static.tigerbbs.com/21293fdcf9ca6782d2363d3e07b49a97\" tg-width=\"963\" tg-height=\"699\"></p><p>It's hard to believe these days, but once upon a time, multi-threaded workloads were far more punishing than single-threaded ones. But it wasn't that long ago.</p><p><img src=\"https://static.tigerbbs.com/dd4a36a1969b1c15787bc029dfeb2353\" tg-width=\"962\" tg-height=\"705\"></p><p>In 2001-2005, the era of Athlon XP and Athlon 64 vs Pentium 4-AMD dominated on dual-threaded workloads. In 2006, Intel's Penti um965 changed that balance.</p><p><img src=\"https://static.tigerbbs.com/57b7ecaf8abbcd19ad7621b7bf815f6b\" tg-width=\"955\" tg-height=\"712\"></p><p>Even in the Athlon era, AMD struggled to keep up with Intel on single-threaded performance. Its Even in the Athlon era, AMD struggled to catch up with Intel on single-threaded performance. Its Athlon 64 X2 processor held the lead at one point, but Intel recaptured it with the Core 2 Extreme QX9650.</p><p>This is something many of us forgot in the early 2000s-multitasking performance is so poor. In the first chart above, we are looking at the ratio between multi-threaded and single-threaded Passmark CPU benchmark scores. For 2001-2005 CPUs, single-threaded scores are actually higher than multi-threaded, meaning you can do more computing on a single thread in a given amount of time, rather than breaking down into parallel threads.</p><p>You might think it's because they're still single-core, single-threaded CPUs, but you're wrong. Intel introduced Hyper-Threading at the 1c/2t Pentium 4 2.8GHz in 2002, which had little impact on actual multithreading capabilities. AMD introduced its first true dual-core desktop CPU in 2005 – the Athlon 64 X2 4800+, but it didn't make much of a difference either.</p><p>Prior to Intel's 2006 Pentium Extreme 965 (dual-core/quad-threaded CPU), the ratio of multithreading did not break 100% – meaning multithreaded tasks were completed faster than equivalent single-threaded tasks. The following year, both AMD and Intel introduced true quad-core desktop CPUs and permanently ended the era of high-performance single-threaded processors.</p><p>Even in this golden age of AMD, Intel has beaten AMD in single-threaded performance. The Pentium 4 architecture is rounded out and deservedly mocked as the weaker successor to the Pentium III, which tends to outperform on a time-by-time basis. But, the Pentium III can't clock at the level of the Pentium 4, and, in any case, the highest-performing workloads of the era are single-threaded, and Intel's Pentium 4 is ahead of AMD's Athlon in performance. Despite the weak CPU performance of the latter.</p><p>Unfortunately for Intel, the entire world had moved to multitasking operating systems long before 2001, with the emergence of multithreaded workloads, as demonstrated in the adoption of tabbed browsing by Firefox predecessor Phoenix in 2002. As CPUs saturate and users become accustomed to an environment where many programs are open throughout the day, multi-threaded performance (rather than single-threaded) becomes a priority.</p><p>In 2007, both AMD and Intel introduced true quad-core desktop CPUs. Multithreading performance from enthusiast-grade CPUs from both vendors is at its peak — and while both sides are reaping huge gains, Intel's Core 2 Extreme QX9650 beats AMD's new Phenom X4.</p><p>The new quad-core CPU has enough multitasking performance to handle both low-priority background tasks and high-priority foreground tasks without significantly compromising the foreground tasks, which shifts the focus to single-threaded performance. This is especially unfortunate for AMD, as not only did Intel improve on multi-threading performance, the QX9650 also made a breakthrough on single-threading performance.</p><p>With nearly 70% of the single-threaded performance advantage replaced by Intel's qx9650, and the focus suddenly returning to the king of single-threaded performance, at least among these high-end CPUs – AMD's first golden age of the 21st century is officially over.</p><p><b>Intel Strikes Back: 2007-2013</b></p><p><img src=\"https://static.tigerbbs.com/a78acaa1e711a3d5ec740f5bb24fccfe\" tg-width=\"984\" tg-height=\"713\"></p><p>Intel's 2009 i7-975 Extreme isn't much of an improvement over the 2008 i7-965 Extreme, but the 2010 hexadecimal core i7-980X is a melting pot. Meanwhile, AMD, while making steady progress year after year, remains at a disadvantage.</p><p><img src=\"https://static.tigerbbs.com/a2ab806f71746d51dc4e97b6d1a230a9\" tg-width=\"973\" tg-height=\"719\"></p><p>Intel had a single-threaded advantage over AMD throughout 2007-2013, and usually quite a bit.</p><p>2007-2013 was an exhilarating time in terms of improving CPU performance and functionality, but rather boring in terms of the battle between Intel and AMD. Intel started with a healthy lead in single and multi-threaded performance and stayed ahead.</p><p>Throughout the operation of the Phenom and Phenom II architectures, AMD has achieved steady growth in both multi-threaded and single-threaded performance. In this way, it can almost catch up to Intel in single-threaded performance, but in multi-threading, it never gets close.</p><p>Both companies released their first six-core CPUs in 2010, but AMD's X6 1100T uses the same K10 architecture as 2009's X4 965, and Intel's i7-980X both achieves chip reduction and adds a new set of AES instructions. This has almost tripled Intel's multithreading performance while still maintaining a small single-thread boot time.</p><p>In 2011, AMD went all-in on multithreading with its ill-fated octa-core Bulldozer architecture. While Bulldozer did begin to gain significant ground on Intel in the multithreading space, it came at a cost. The new FX-8150 actually goes backwards in single-threaded performance.</p><p>By 2012, AMD's Bulldozer had become a Piledriver and had almost caught up with Intel in multithreading, but that was only because Intel reduced its core count from six to four when it scaled back its process to 22nm. This gives AMD \"almost as good\" multi-threaded performance with twice as many cores, while single-threaded performance is three years behind Intel.</p><p>In 2013, to make matters worse, AMD introduced the ultimate PileDriver CPU (the hateful FX-9590). Despite Intel's increasing lead in engineering, the CPU represents a Hail Mary approach to staying competitive-it mainly keeps pace with the Intel i7-4770K, but only through clock frequencies and voltages that typically remain in the \"extreme overclocker\" realm.</p><p>With a rated TDP of 225W compared to the 125W of the FX-8350 and the 84W of the i7-4770K, the FX-9590 is a nearly unusable CPU. Air-cooling it is unchallenging in typical circumstances, and the noise and waste heat from the fan is annoying that even a hardcore AMD fan can't make excuses for it, especially when its competition still easily outperforms it.</p><p>The FX-9590 is AMD's last enthusiast CPU in four years. Starting in 2014-2017, Red Team's only new CPU versions are budget CPUs and APUs, none of which are better than the FX-9590, let alone compete.</p><p><b>Change of Destiny: 2013-2020</b></p><p><img src=\"https://static.tigerbbs.com/bb79e3d31da7a59fa8039acf96334a87\" tg-width=\"957\" tg-height=\"714\"></p><p><img src=\"https://static.tigerbbs.com/541d74b2eda21ab98636a35c01fb9b78\" tg-width=\"959\" tg-height=\"716\"><img src=\"https://static.tigerbbs.com/d54fa6e7ae94137a28f970172b0b9ea8\" tg-width=\"953\" tg-height=\"716\"><img src=\"https://static.tigerbbs.com/09c34b9c8f96a5c34c1e4d01439ea430\" tg-width=\"957\" tg-height=\"711\"></p><p>The fact that AMD survived a three-year slump without releasing a new CPU speaks to a fraction of the market — and the profit margins of each vendor — where the biggest, fastest CPU is the real thing. In 2014-2017, the Red team went on to introduce new desktop CPUs – but all the new designs targeted \"more fiscally reasonable\" market segments, focusing on cheaper APUs.</p><p>Despite the AMD fiasco, Intel was taken aback. To their credit, Blue Team did continue to release new performance CPUs and achieved incremental gains in single-threaded performance during that time, but not much for single-threaded gains. In 2016, it was difficult to upgrade from a 5 year old CPU.</p><p>In 2017, everything changed when AMD finally released a new, warm and friendly architecture codenamed Zen. The Zen architecture brings power and thermal efficiency back into line while bringing raw multithreading performance significantly above Intel's best. Although the Zen still lags Intel in single-threaded performance, it also made several big jumps to reclaim that crown.</p><p>The 2018 Zen + was a minor improvement at best in terms of multi-threaded performance, but it held a modest share of Intel's single-threaded lead compared to the original Zen. Intel's first viable i9 product, the i9-9900K, briefly enabled Red Team to reclaim the crown of multithreaded performance, but the lead didn't last long.</p><p>In 2019, Intel made a big fuss in the field of zealots – its best CPU is the i9-9900KS, which has hardly any improvement over the i9-9900K of 2018. Meanwhile, AMD's Ryzen 9 3950x has delivered another huge multi-threading leap and a powerful single-threading growth. The Ryzen 9 3950x is a 16-core/32-thread giant processor that demonstrates the best performance of its new Zen 2 architecture. AMD hasn't quite caught up with Intel's single-threading lead yet, but it more than doubled the multi-threading performance of the i9-9900KS.</p><p>Which brings us to this year. Intel, still struggling with delayed 10nm process processes, released the i9-10900K, which, although still running on the same architecture as the 9900K and 9900KS, has made significant improvements in both single-and multi-threaded workloads. Unfortunately for Blue Team, AMD has not taken a break-the Zen 3 architecture of 2020 greatly improved the Zen 2's single-threaded performance.</p><p>At this point, AMD is leading the way in single-threaded performance. Furthermore, for the second year in a row, it is way ahead of the curve in multi-threaded performance, almost twice the best Intel can offer.</p><p><b>Long term: Generation-to-generation improvements from 2001-2020</b></p><p><img src=\"https://static.tigerbbs.com/6d8d6c2cce8e14dfd53f525a9e9eb85a\" tg-width=\"964\" tg-height=\"711\"></p><p>See an anomaly with Intel in 2012? The company went from a six-core CPU to a quad-core-and no one cared, as single-threaded performance was improved accordingly.</p><p><img src=\"https://static.tigerbbs.com/e304b2e92eea398c235146f2e8697739\" tg-width=\"964\" tg-height=\"713\"></p><p>AMD's largest single-threaded spike corresponds to the introduction of the Athlon 64 X2 architecture in 2005. In 2007, Intel launched its Core family of products.</p><p><img src=\"https://static.tigerbbs.com/9a9036b443970aa7f71dd940b4a7b5bc\" tg-width=\"957\" tg-height=\"712\"></p><p>At first glance, there isn't much difference between this chart and pure multithreading, but note that AMD's<a href=\"https://laohu8.com/S/42T.SI\">Trend line</a>It will be consistent between 2016 and 2020.</p><p>If we want to see how the engineering process works for Intel and AMD, it is instructive to proceed from the raw data and observe the generational improvement in performance-which means that this year is a few percent better than last year.</p><p>These charts amplify the impact of the respective architectural changes and engineering efforts of the two companies, making it easier to spot major changes. Multithreading performance has been the biggest and most exciting peak so far-with the introduction of quad-core CPUs in 2007, multithreading performance has also reached peak performance. On a single-threaded chart, almost nothing is visible.</p><p>Our favorite version of the generation chart is the third generation, which averages single-threaded and multi-threaded improvements to a straight line and gives both the same weight. Although it looks roughly similar to a pure multithreaded chart at first glance, there are some interesting features that pop up.</p><p>The first to jump out of here is Intel's brilliant performance in 2005, with less than 100% points on the chart. A drop below 100% may be small enough to make the error volatile-but, even if the Pentium 3.8GHz is not actively worse than the 2004 Pentium 4 3.6GHz, there is clearly no improvement.</p><p>It will be interesting to look again at AMD's 2017-2020 progress from Zen to Zen +, Zen 2 and Zen 3. Although the overall shape here is the same as the same pronounced M we see on the pure multithreaded chart, the inclusion of single-threaded performance clearly shows that each step is an important and continuous improvement beyond the last.</p><p><b>2021 and beyond</b></p><p>For Intel on the desktop, the situation is grim now, while another major architectural improvement from AMD (Zen 4) arrived in 2021, which AMD calls \"step by step\" with its process nodes shrunk to 5nm.</p><p>On the Intel side, we don't expect it to be as bad as it was from the i9-9900K to the i9-9900KS. But, that doesn't mean Blue Team is full of light. While the i9-10900K does represent a significant and surprising improvement over previous generations, it doesn't match the rise of AMD-it still represents the last breath of a dying architecture.</p><p>Intel's desktop CPUs are still stuck on the 14nm process, and next year's Rocket Lake will also maintain the 14nm process. The Cypress Cove architecture is essentially a variation of Intel's 10nm Sunny Cove notebook architecture that was reverse-ported to the earlier 14nm process.</p><p>While Intel has gained a significant performance boost with its latest Sunny Cove laptop processor (this year's Tiger Lake has only been seen in prototypes so far this year), its core and thread count is minimal, while high performance seems to correlate with higher power consumption.</p><p>It is too early to say whether the 10nm Sunny Cove's limitations in terms of core count and power efficiency will shift to the 14nm Cypress Cove. But, it seems unlikely to see Intel lead and largely sustain AMD's renaissance in 2021 like Zen and its successors did.</p><p>Still, bad news for Intel is likely to be good news for performance enthusiasts. Intel, a much smaller competitor, will need more than one or two dominant cycles to solidify relationships and market perception with large OEMs and system manufacturers. These business relationships — and the resulting orders, revenue, and profits — will be necessary to maintain AMD's strong R&D presence to ensure that AMD can deliver higher performance and better prices to consumers in the future.</p>","source":"bdthygc","collect":0,"html":"<!DOCTYPE html>\n<html>\n<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\" />\n<meta name=\"viewport\" content=\"width=device-width,initial-scale=1.0,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no\"/>\n<meta name=\"format-detection\" content=\"telephone=no,email=no,address=no\" />\n<title>Intel's tortoise and rabbit race against AMD</title>\n<style type=\"text/css\">\na,abbr,acronym,address,applet,article,aside,audio,b,big,blockquote,body,canvas,caption,center,cite,code,dd,del,details,dfn,div,dl,dt,\nem,embed,fieldset,figcaption,figure,footer,form,h1,h2,h3,h4,h5,h6,header,hgroup,html,i,iframe,img,ins,kbd,label,legend,li,mark,menu,nav,\nobject,ol,output,p,pre,q,ruby,s,samp,section,small,span,strike,strong,sub,summary,sup,table,tbody,td,tfoot,th,thead,time,tr,tt,u,ul,var,video{ font:inherit;margin:0;padding:0;vertical-align:baseline;border:0 }\nbody{ font-size:16px; line-height:1.5; color:#999; background:transparent; }\n.wrapper{ overflow:hidden;word-break:break-all;padding:10px; }\nh1,h2{ font-weight:normal; line-height:1.35; margin-bottom:.6em; }\nh3,h4,h5,h6{ line-height:1.35; margin-bottom:1em; }\nh1{ font-size:24px; }\nh2{ font-size:20px; }\nh3{ font-size:18px; }\nh4{ font-size:16px; }\nh5{ font-size:14px; }\nh6{ font-size:12px; }\np,ul,ol,blockquote,dl,table{ margin:1.2em 0; }\nul,ol{ margin-left:2em; }\nul{ list-style:disc; }\nol{ list-style:decimal; }\nli,li p{ margin:10px 0;}\nimg{ max-width:100%;display:block;margin:0 auto 1em; }\nblockquote{ color:#B5B2B1; border-left:3px solid #aaa; padding:1em; }\nstrong,b{font-weight:bold;}\nem,i{font-style:italic;}\ntable{ width:100%;border-collapse:collapse;border-spacing:1px;margin:1em 0;font-size:.9em; }\nth,td{ padding:5px;text-align:left;border:1px solid #aaa; }\nth{ font-weight:bold;background:#5d5d5d; }\n.symbol-link{font-weight:bold;}\n/* header{ border-bottom:1px solid #494756; } */\n.title{ margin:0 0 8px;line-height:1.3;color:#ddd; }\n.meta {color:#5e5c6d;font-size:13px;margin:0 0 .5em; }\na{text-decoration:none; color:#2a4b87;}\n.meta .head { display: inline-block; overflow: hidden}\n.head .h-thumb { width: 30px; height: 30px; margin: 0; padding: 0; border-radius: 50%; float: left;}\n.head .h-content { margin: 0; padding: 0 0 0 9px; float: left;}\n.head .h-name {font-size: 13px; color: #eee; margin: 0;}\n.head .h-time {font-size: 12.5px; color: #7E829C; margin: 0;}\n.small {font-size: 12.5px; display: inline-block; transform: scale(0.9); -webkit-transform: scale(0.9); transform-origin: left; -webkit-transform-origin: left;}\n.smaller {font-size: 12.5px; display: inline-block; transform: scale(0.8); -webkit-transform: scale(0.8); transform-origin: left; -webkit-transform-origin: left;}\n.bt-text {font-size: 12px;margin: 1.5em 0 0 0}\n.bt-text p {margin: 0}\n</style>\n</head>\n<body>\n<div class=\"wrapper\">\n<header>\n<h2 class=\"title\">\nIntel's tortoise and rabbit race against AMD\n</h2>\n<h4 class=\"meta\">\n<p class=\"head\">\n<strong class=\"h-name small\">半导体行业观察</strong><span class=\"h-time small\">2020-11-21 12:49</span>\n</p>\n</h4>\n</header>\n<article>\n<p><a href=\"https://laohu8.com/S/INTC\">Intel</a>With<a href=\"https://laohu8.com/S/AMD\">AMD</a>The rivalry is the rivalry between the tortoise and the hare. But which company is the tortoise? Which company is Rabbit?</p><p>The commentary battle between Intel and AMD fans has been raging over the last couple of launch cycles, with plenty of digital ink devoted to discussing which company has improved significantly over the years, or not. There's also no shortage of opinions on the raw performance of both companies' fastest processors. We thought it would be interesting to delve into the archived performance benchmarking of each company's fastest desktop/CPU to give a good idea of how each company has actually performed over the years and even to see if there are some patterns to gather or some bets to make on the future.</p><p>Before diving into the charts, let's start with a few tables – this way you can see which CPUs we use as milestones each year. During the course of our research, there were some irregularities in the data; We'll talk about those as well, talking about something that a simple chart can't show.</p><p><b>Twenty years of enthusiast calculation</b></p><p><img src=\"https://static.tigerbbs.com/582eb6403400c9c07e9f29f893cb848e\" tg-width=\"1044\" tg-height=\"1193\"></p><p>While Intel and AMD launch a large number of processors each year for different price points and target markets, we are limited to the fastest desktop or \"enthusiast\" processors each year. That means no server processors, and no high-end desktop (HEDT) processors — so we're not looking at Threadrippers or the latest model XE series Intel parts.</p><p>Even for someone like me who has been a system builder for the entire period, putting together a list like this is a huge pain, let alone matching test results. This is especially difficult with AMD, which doesn't really equal Intel's Ark to offer a single list of processors, sorted by generation, type, and release date. If you think I picked the wrong \"top processor\" of the year, let me know in the comments.</p><p>We should address some of the anomalies that appeared in the chart – first of all, no, it wasn't an oversight for AMD not to introduce new high-performance CPUs in 2014, 2015 and 2016. The 2013 FX-9590 was widely (and rightfully) criticized for its staggering 225W TDP and mediocre performance — Team Red's fastest CPU in four years of operation. During this time, AMD introduced several generations of low-power, inexpensive desktop devices, but none of them outperformed the FX-9590.</p><p>Intel also experienced several missteps between 2013 and 2017, although not as severe or lasting as long. The fifth-generation Core series essentially died in 2014, although updates to the fourth-generation Core i7 series do offer a significant performance boost. Two years later, there's technically another bandwagon-the Core i7-7700K-equipped Kaby Lake didn't really launch until January 2017. We fabricated it a bit to get Kaby Lake into the 2016 ranks, because otherwise it would have disappeared entirely, with Coffee Lake and the i7-8700K showing up later that year.</p><p><b>AMD Athlon The rise and fall of: 2001-2007</b></p><p><img src=\"https://static.tigerbbs.com/21293fdcf9ca6782d2363d3e07b49a97\" tg-width=\"963\" tg-height=\"699\"></p><p>It's hard to believe these days, but once upon a time, multi-threaded workloads were far more punishing than single-threaded ones. But it wasn't that long ago.</p><p><img src=\"https://static.tigerbbs.com/dd4a36a1969b1c15787bc029dfeb2353\" tg-width=\"962\" tg-height=\"705\"></p><p>In 2001-2005, the era of Athlon XP and Athlon 64 vs Pentium 4-AMD dominated on dual-threaded workloads. In 2006, Intel's Penti um965 changed that balance.</p><p><img src=\"https://static.tigerbbs.com/57b7ecaf8abbcd19ad7621b7bf815f6b\" tg-width=\"955\" tg-height=\"712\"></p><p>Even in the Athlon era, AMD struggled to keep up with Intel on single-threaded performance. Its Even in the Athlon era, AMD struggled to catch up with Intel on single-threaded performance. Its Athlon 64 X2 processor held the lead at one point, but Intel recaptured it with the Core 2 Extreme QX9650.</p><p>This is something many of us forgot in the early 2000s-multitasking performance is so poor. In the first chart above, we are looking at the ratio between multi-threaded and single-threaded Passmark CPU benchmark scores. For 2001-2005 CPUs, single-threaded scores are actually higher than multi-threaded, meaning you can do more computing on a single thread in a given amount of time, rather than breaking down into parallel threads.</p><p>You might think it's because they're still single-core, single-threaded CPUs, but you're wrong. Intel introduced Hyper-Threading at the 1c/2t Pentium 4 2.8GHz in 2002, which had little impact on actual multithreading capabilities. AMD introduced its first true dual-core desktop CPU in 2005 – the Athlon 64 X2 4800+, but it didn't make much of a difference either.</p><p>Prior to Intel's 2006 Pentium Extreme 965 (dual-core/quad-threaded CPU), the ratio of multithreading did not break 100% – meaning multithreaded tasks were completed faster than equivalent single-threaded tasks. The following year, both AMD and Intel introduced true quad-core desktop CPUs and permanently ended the era of high-performance single-threaded processors.</p><p>Even in this golden age of AMD, Intel has beaten AMD in single-threaded performance. The Pentium 4 architecture is rounded out and deservedly mocked as the weaker successor to the Pentium III, which tends to outperform on a time-by-time basis. But, the Pentium III can't clock at the level of the Pentium 4, and, in any case, the highest-performing workloads of the era are single-threaded, and Intel's Pentium 4 is ahead of AMD's Athlon in performance. Despite the weak CPU performance of the latter.</p><p>Unfortunately for Intel, the entire world had moved to multitasking operating systems long before 2001, with the emergence of multithreaded workloads, as demonstrated in the adoption of tabbed browsing by Firefox predecessor Phoenix in 2002. As CPUs saturate and users become accustomed to an environment where many programs are open throughout the day, multi-threaded performance (rather than single-threaded) becomes a priority.</p><p>In 2007, both AMD and Intel introduced true quad-core desktop CPUs. Multithreading performance from enthusiast-grade CPUs from both vendors is at its peak — and while both sides are reaping huge gains, Intel's Core 2 Extreme QX9650 beats AMD's new Phenom X4.</p><p>The new quad-core CPU has enough multitasking performance to handle both low-priority background tasks and high-priority foreground tasks without significantly compromising the foreground tasks, which shifts the focus to single-threaded performance. This is especially unfortunate for AMD, as not only did Intel improve on multi-threading performance, the QX9650 also made a breakthrough on single-threading performance.</p><p>With nearly 70% of the single-threaded performance advantage replaced by Intel's qx9650, and the focus suddenly returning to the king of single-threaded performance, at least among these high-end CPUs – AMD's first golden age of the 21st century is officially over.</p><p><b>Intel Strikes Back: 2007-2013</b></p><p><img src=\"https://static.tigerbbs.com/a78acaa1e711a3d5ec740f5bb24fccfe\" tg-width=\"984\" tg-height=\"713\"></p><p>Intel's 2009 i7-975 Extreme isn't much of an improvement over the 2008 i7-965 Extreme, but the 2010 hexadecimal core i7-980X is a melting pot. Meanwhile, AMD, while making steady progress year after year, remains at a disadvantage.</p><p><img src=\"https://static.tigerbbs.com/a2ab806f71746d51dc4e97b6d1a230a9\" tg-width=\"973\" tg-height=\"719\"></p><p>Intel had a single-threaded advantage over AMD throughout 2007-2013, and usually quite a bit.</p><p>2007-2013 was an exhilarating time in terms of improving CPU performance and functionality, but rather boring in terms of the battle between Intel and AMD. Intel started with a healthy lead in single and multi-threaded performance and stayed ahead.</p><p>Throughout the operation of the Phenom and Phenom II architectures, AMD has achieved steady growth in both multi-threaded and single-threaded performance. In this way, it can almost catch up to Intel in single-threaded performance, but in multi-threading, it never gets close.</p><p>Both companies released their first six-core CPUs in 2010, but AMD's X6 1100T uses the same K10 architecture as 2009's X4 965, and Intel's i7-980X both achieves chip reduction and adds a new set of AES instructions. This has almost tripled Intel's multithreading performance while still maintaining a small single-thread boot time.</p><p>In 2011, AMD went all-in on multithreading with its ill-fated octa-core Bulldozer architecture. While Bulldozer did begin to gain significant ground on Intel in the multithreading space, it came at a cost. The new FX-8150 actually goes backwards in single-threaded performance.</p><p>By 2012, AMD's Bulldozer had become a Piledriver and had almost caught up with Intel in multithreading, but that was only because Intel reduced its core count from six to four when it scaled back its process to 22nm. This gives AMD \"almost as good\" multi-threaded performance with twice as many cores, while single-threaded performance is three years behind Intel.</p><p>In 2013, to make matters worse, AMD introduced the ultimate PileDriver CPU (the hateful FX-9590). Despite Intel's increasing lead in engineering, the CPU represents a Hail Mary approach to staying competitive-it mainly keeps pace with the Intel i7-4770K, but only through clock frequencies and voltages that typically remain in the \"extreme overclocker\" realm.</p><p>With a rated TDP of 225W compared to the 125W of the FX-8350 and the 84W of the i7-4770K, the FX-9590 is a nearly unusable CPU. Air-cooling it is unchallenging in typical circumstances, and the noise and waste heat from the fan is annoying that even a hardcore AMD fan can't make excuses for it, especially when its competition still easily outperforms it.</p><p>The FX-9590 is AMD's last enthusiast CPU in four years. Starting in 2014-2017, Red Team's only new CPU versions are budget CPUs and APUs, none of which are better than the FX-9590, let alone compete.</p><p><b>Change of Destiny: 2013-2020</b></p><p><img src=\"https://static.tigerbbs.com/bb79e3d31da7a59fa8039acf96334a87\" tg-width=\"957\" tg-height=\"714\"></p><p><img src=\"https://static.tigerbbs.com/541d74b2eda21ab98636a35c01fb9b78\" tg-width=\"959\" tg-height=\"716\"><img src=\"https://static.tigerbbs.com/d54fa6e7ae94137a28f970172b0b9ea8\" tg-width=\"953\" tg-height=\"716\"><img src=\"https://static.tigerbbs.com/09c34b9c8f96a5c34c1e4d01439ea430\" tg-width=\"957\" tg-height=\"711\"></p><p>The fact that AMD survived a three-year slump without releasing a new CPU speaks to a fraction of the market — and the profit margins of each vendor — where the biggest, fastest CPU is the real thing. In 2014-2017, the Red team went on to introduce new desktop CPUs – but all the new designs targeted \"more fiscally reasonable\" market segments, focusing on cheaper APUs.</p><p>Despite the AMD fiasco, Intel was taken aback. To their credit, Blue Team did continue to release new performance CPUs and achieved incremental gains in single-threaded performance during that time, but not much for single-threaded gains. In 2016, it was difficult to upgrade from a 5 year old CPU.</p><p>In 2017, everything changed when AMD finally released a new, warm and friendly architecture codenamed Zen. The Zen architecture brings power and thermal efficiency back into line while bringing raw multithreading performance significantly above Intel's best. Although the Zen still lags Intel in single-threaded performance, it also made several big jumps to reclaim that crown.</p><p>The 2018 Zen + was a minor improvement at best in terms of multi-threaded performance, but it held a modest share of Intel's single-threaded lead compared to the original Zen. Intel's first viable i9 product, the i9-9900K, briefly enabled Red Team to reclaim the crown of multithreaded performance, but the lead didn't last long.</p><p>In 2019, Intel made a big fuss in the field of zealots – its best CPU is the i9-9900KS, which has hardly any improvement over the i9-9900K of 2018. Meanwhile, AMD's Ryzen 9 3950x has delivered another huge multi-threading leap and a powerful single-threading growth. The Ryzen 9 3950x is a 16-core/32-thread giant processor that demonstrates the best performance of its new Zen 2 architecture. AMD hasn't quite caught up with Intel's single-threading lead yet, but it more than doubled the multi-threading performance of the i9-9900KS.</p><p>Which brings us to this year. Intel, still struggling with delayed 10nm process processes, released the i9-10900K, which, although still running on the same architecture as the 9900K and 9900KS, has made significant improvements in both single-and multi-threaded workloads. Unfortunately for Blue Team, AMD has not taken a break-the Zen 3 architecture of 2020 greatly improved the Zen 2's single-threaded performance.</p><p>At this point, AMD is leading the way in single-threaded performance. Furthermore, for the second year in a row, it is way ahead of the curve in multi-threaded performance, almost twice the best Intel can offer.</p><p><b>Long term: Generation-to-generation improvements from 2001-2020</b></p><p><img src=\"https://static.tigerbbs.com/6d8d6c2cce8e14dfd53f525a9e9eb85a\" tg-width=\"964\" tg-height=\"711\"></p><p>See an anomaly with Intel in 2012? The company went from a six-core CPU to a quad-core-and no one cared, as single-threaded performance was improved accordingly.</p><p><img src=\"https://static.tigerbbs.com/e304b2e92eea398c235146f2e8697739\" tg-width=\"964\" tg-height=\"713\"></p><p>AMD's largest single-threaded spike corresponds to the introduction of the Athlon 64 X2 architecture in 2005. In 2007, Intel launched its Core family of products.</p><p><img src=\"https://static.tigerbbs.com/9a9036b443970aa7f71dd940b4a7b5bc\" tg-width=\"957\" tg-height=\"712\"></p><p>At first glance, there isn't much difference between this chart and pure multithreading, but note that AMD's<a href=\"https://laohu8.com/S/42T.SI\">Trend line</a>It will be consistent between 2016 and 2020.</p><p>If we want to see how the engineering process works for Intel and AMD, it is instructive to proceed from the raw data and observe the generational improvement in performance-which means that this year is a few percent better than last year.</p><p>These charts amplify the impact of the respective architectural changes and engineering efforts of the two companies, making it easier to spot major changes. Multithreading performance has been the biggest and most exciting peak so far-with the introduction of quad-core CPUs in 2007, multithreading performance has also reached peak performance. On a single-threaded chart, almost nothing is visible.</p><p>Our favorite version of the generation chart is the third generation, which averages single-threaded and multi-threaded improvements to a straight line and gives both the same weight. Although it looks roughly similar to a pure multithreaded chart at first glance, there are some interesting features that pop up.</p><p>The first to jump out of here is Intel's brilliant performance in 2005, with less than 100% points on the chart. A drop below 100% may be small enough to make the error volatile-but, even if the Pentium 3.8GHz is not actively worse than the 2004 Pentium 4 3.6GHz, there is clearly no improvement.</p><p>It will be interesting to look again at AMD's 2017-2020 progress from Zen to Zen +, Zen 2 and Zen 3. Although the overall shape here is the same as the same pronounced M we see on the pure multithreaded chart, the inclusion of single-threaded performance clearly shows that each step is an important and continuous improvement beyond the last.</p><p><b>2021 and beyond</b></p><p>For Intel on the desktop, the situation is grim now, while another major architectural improvement from AMD (Zen 4) arrived in 2021, which AMD calls \"step by step\" with its process nodes shrunk to 5nm.</p><p>On the Intel side, we don't expect it to be as bad as it was from the i9-9900K to the i9-9900KS. But, that doesn't mean Blue Team is full of light. While the i9-10900K does represent a significant and surprising improvement over previous generations, it doesn't match the rise of AMD-it still represents the last breath of a dying architecture.</p><p>Intel's desktop CPUs are still stuck on the 14nm process, and next year's Rocket Lake will also maintain the 14nm process. The Cypress Cove architecture is essentially a variation of Intel's 10nm Sunny Cove notebook architecture that was reverse-ported to the earlier 14nm process.</p><p>While Intel has gained a significant performance boost with its latest Sunny Cove laptop processor (this year's Tiger Lake has only been seen in prototypes so far this year), its core and thread count is minimal, while high performance seems to correlate with higher power consumption.</p><p>It is too early to say whether the 10nm Sunny Cove's limitations in terms of core count and power efficiency will shift to the 14nm Cypress Cove. But, it seems unlikely to see Intel lead and largely sustain AMD's renaissance in 2021 like Zen and its successors did.</p><p>Still, bad news for Intel is likely to be good news for performance enthusiasts. Intel, a much smaller competitor, will need more than one or two dominant cycles to solidify relationships and market perception with large OEMs and system manufacturers. These business relationships — and the resulting orders, revenue, and profits — will be necessary to maintain AMD's strong R&D presence to ensure that AMD can deliver higher performance and better prices to consumers in the future.</p>\n<div class=\"bt-text\">\n\n\n<p> source:<a href=\"https://mp.weixin.qq.com/s/jthH9s2r55PX6uVCqh0Cgg\">半导体行业观察</a></p>\n\n\n</div>\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"https://static.tigerbbs.com/8a1a8c5698b6f32ee1637556a15bb35b","relate_stocks":{"INTC":"英特尔","03086":"华夏纳指","AMD":"美国超微公司","09086":"华夏纳指-U"},"source_url":"https://mp.weixin.qq.com/s/jthH9s2r55PX6uVCqh0Cgg","is_english":false,"share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","article_id":"2085230767","content_text":"英特尔与AMD的竞争是乌龟与野兔的竞争。但哪家公司是龟?哪家公司是兔?\n在过去的几个发布周期中,英特尔和AMD的粉丝之间的评论之战一直很激烈,大量的数字墨水都在讨论这些年来哪个公司有了显著的改善,或者没有明显的改善。对于两家公司最快处理器的原始性能,也不乏意见。我们认为,深入研究每个公司的最快桌面/ CPU的存档性能基准测试是很有趣的,这样可以很好地了解这些年来每个公司的实际表现,甚至可以看看是否有一些模式需要收集,或者对未来进行一些押注。\n在深入研究图表之前,让我们先从一些表开始——通过这种方式,您可以看到我们使用了哪些CPU作为每年的里程碑。在我们研究的过程中,数据中有一些不规则现象;我们也会讨论这些,讨论一些简单图表无法展示的东西。\n二十年来的发烧友计算\n\n虽然英特尔和AMD每年都会针对不同的价位和目标市场推出大量的处理器,但我们每年都局限于速度最快的台式机或“发烧友”处理器。这意味着没有服务器处理器,也没有高端台式机(HEDT)处理器——所以我们不会关注Threadrippers或最新型号XE系列英特尔部件。\n即使是像我这样整个时期都是作为一个系统构建者的人,组装这样的列表也是一个巨大的痛苦,更不用说匹配测试结果了。这在AMD方面尤其困难,因为AMD并没有真正相当于英特尔的Ark提供单一的处理器列表,按代、类型和发布日期排序。如果你认为我选错了一年中的“顶级处理器”,请在评论中告诉我。\n我们应该解决图表中出现的一些异常情况——首先,不,2014年、2015年和2016年AMD没有推出新的高性能CPU并不是疏忽。2013年的fx -9590被广泛(并且是理所应当)批评了它惊人的225W TDP和平庸的性能——在四年的运行中是Team Red最快的CPU。在此期间,AMD推出了几代低功耗、价格低廉的桌面设备,但没有一款性能优于FX-9590。\n英特尔在2013年至2017年期间也经历了几次失误,尽管没有那么严重或持续这么长时间。第五代Core系列基本上死于2014年,尽管第四代Core i7系列的更新确实提供了显著的性能提升。两年后,从技术上讲又是另一种风潮-配备Core i7-7700K的Kaby Lake直到2017年1月才真正启动。我们稍加捏造,让Kaby Lake进入2016年的行列,因为否则它会完全消失,同年晚些时候Coffee Lake和i7-8700K出现。\nAMD Athlon的兴衰:2001-2007年\n\n现在很难相信,但曾几何时,多线程工作负载远比单线程负载要punish的多。但这并不是很久以前。\n\n在2001-2005年,Athlon XP和Athlon 64 vs Pentium 4-AMD的时代在双线程工作负载上独占鳌头。2006年,英特尔的Penti um965改变了这一平衡。\n\nEven in the Athlon era, AMD struggled to keep up with Intel on single-threaded performance. Its即使在Athlon时代,AMD也难以在单线程性能上赶上英特尔。它的Athlon 64 X2处理器一度占据领先地位,但英特尔(Intel)凭借酷睿2 Extreme QX9650夺回了领先地位。\n这是我们许多人在21世纪初忘记的事情-多任务处理性能 太差了。在上面的第一个图表中,我们正在研究多线程和单线程Passmark CPU基准测试分数之间的比率。对于2001-2005年份的CPU,单线程得分实际上 高于多线程,这意味着您可以在给定的时间内在单线程上完成更多的计算工作,而不是分解为并行线程。\n您可能会认为这是因为它们仍然是单核、单线程的CPU,但是您错了。Intel在2002年引入了1c/2t Pentium 4 2.8GHz的超线程,这对实际的多线程能力影响很小。AMD在2005年推出了第一个真正的双核桌面CPU——Athlon 64 X2 4800+,但它也没有太大的区别。\n在英特尔2006年的Pentium Extreme 965(双核/四线程CPU)之前,多线程的比率没有突破100%——这意味着多线程任务比同等的单线程任务完成得更快。第二年,AMD和英特尔都推出了真正的四核桌面CPU,并永久地结束了高性能单线程处理器的时代。\n即使在AMD的这个黄金时代,英特尔在单线程性能方面也击败了AMD。Pentium 4架构作为Pentium III的较弱后继者而受到全面和应有的嘲讽,而PentiumIII往往会逐时胜过它。但是,Pentium III的时钟频率不能达到Pentium 4的水平,而且,无论如何,该时代性能最高的工作负载都是单线程的,并且英特尔的Pentium 4在性能上领先于AMD的Athlon。尽管后者的CPU性能很弱。\n对英特尔来说不幸的是,早在2001年之前,整个世界就已经转向了多任务操作系统,多线程的工作负载也随之出现,这在Firefox前身Phoenix在2002年采用选项卡式浏览中得到了证明。随着CPU饱和,用户逐渐习惯于全天打开许多程序的环境,多线程性能(而不是单线程)成为当务之急。\n2007年,AMD和英特尔都推出了真正的四核桌面CPU。两家厂商的发烧级CPU的多线程性能都达到了顶峰——尽管双方都获得了巨大的收益,但英特尔的Core 2 Extreme QX9650击败了AMD的新Phenom X4。\n新的四核CPU具有足够的多任务性能,可以同时处理低优先级的后台任务和高优先级的前台任务,而不会使前台任务明显受损,这就将焦点转移到了单线程性能上。这对AMD来说尤其不幸,因为Intel不仅在多线程性能上有所改进,QX9650在单线程性能上也取得了突破。\n随着近70%的单线程性能优势被英特尔的qx9650所取代,以及焦点突然回到单线程性能的王者地位,至少在这些高端CPU中是这样——AMD 21世纪的第一个黄金时代正式结束了。\n英特尔反击:2007-2013\n\n与2008年的i7-965 Extreme相比,英特尔2009年的i7-975 Extreme并没有太大的改进,但2010年的十六进制core i7-980X则是一个大熔炉。与此同时,AMD虽然年复一年地稳步前进,但仍然处于劣势。\n\n在整个2007年至2013年期间,英特尔都比AMD拥有单线程优势,而且通常是相当多的优势。\n就提高CPU性能和功能而言,2007-2013年是一个令人振奋的时代,但就英特尔与AMD之间的斗争而言,这是一个相当无聊的时代。英特尔从单线程和多线程性能的健康领先开始,一直保持领先地位。\n在整个Phenom和Phenom II架构的运行过程中,AMD在多线程和单线程性能方面均取得了稳步的增长。这样一来,它在单线程性能上几乎可以赶上英特尔,但在多线程方面,它从未接近过。\n两家公司都于2010年发布了他们的第一个六核CPU,但是AMD的X6 1100T与2009年的X4 965使用相同的K10架构,英特尔的i7-980X既实现了芯片缩减,又添加了一套新的AES指令。这使Intel的多线程性能几乎飙升了两倍,同时仍然保持了很小的单线程引导时间。\n2011年,AMD凭借其命运多舛的八核Bulldozer架构全力投入多线程。尽管Bulldozer确实在多线程领域开始在Intel上取得重要地位,但这是有代价的。新的FX-8150实际上在单线程性能上倒退了。\n到2012年,AMD的Bulldozer已成为Piledriver,并且在多线程方面几乎赶上了Intel,但这仅仅是因为Intel在将其工艺缩减至22nm时将其核心数量从6个减少到4个。这使AMD拥有“几乎一样好”的多线程性能,内核数量是原来的两倍,而单线程性能则比英特尔落后了三年。\n2013年,情况更糟的是,AMD推出了最终的PiledriverCPU(可恶的FX-9590)。尽管英特尔在工程技术上的领先优势不断提高,但该CPU仍代表了一种保持竞争力的Hail Mary方法-它主要与英特尔i7-4770K保持同步,但仅通过时钟频率和电压来实现,而时钟频率和电压通常会留在“极限超频器”领域。\n与FX-8350的125W和i7-4770K的84W相比,额定TDP为225W,FX-9590是几乎无法使用的CPU。在典型情况下,对其进行空气冷却是没有挑战性的,而且风扇的噪音和废热也令人讨厌,即使是铁杆AMD风扇也无法为其找借口,尤其是在其竞争仍然轻松胜过它的时候。\nFX-9590是AMD四年来最后一个发烧级CPU。从2014-2017年开始,Red Team唯一的新CPU版本是预算CPU和APU,没有一个比FX-9590更好,更不用说竞争了。\n命运的变化:2013-2020\n\n\n事实上,AMD在三年的低迷期中幸存下来,却没有发布一款新的CPU,这说明了市场的一小部分——以及每个厂商的利润率——最大、最快的CPU才是真正的。在2014-2017年,Red团队继续推出了新的桌面CPU——但所有的新设计都瞄准了“财政上更合理”的市场细分,集中在便宜的APU上。\n尽管AMD惨败,但英特尔却大为吃惊。值得称赞的是,Blue Team确实继续发布了新的性能CPU,并在这段时间内实现了单线程性能的增量提升,但单线程的提升并不多。在2016年,很难从5年的旧CPU上进行升级。\n2017年,一切都发生了改变,AMD最终发布了一款代号为Zen的全新、热情友好的架构。Zen架构将功率和热效率恢复了一致,同时使原始多线程性能显著高于英特尔的最佳性能。尽管Zen的单线程性能仍落后于英特尔,但它也实现了几次大幅跃升,夺回了这一桂冠。\n2018年的Zen +在多线程性能方面充其量仅是次要的改进,但与最初的Zen相比,它在Intel的单线程领先优势中所占的份额不大。英特尔的第一个可行的i9产品i9-9900K短暂地使Red Team夺回了多线程性能的桂冠,但领先优势并没有持续很长时间。\n2019年,英特尔在狂热者这一领域大做文章——它最好的CPU是i9-9900KS,与2018年的i9-9900K相比几乎没有任何改进。与此同时,AMD的Ryzen 9 3950x又实现了一次巨大的多线程飞跃和一次强大的单线程增长。Ryzen 9 3950x是一款16核/32线程的巨形处理器,展示了其新Zen 2架构的最佳性能。AMD还没有完全赶上英特尔的单线程领先地位,但它将i9-9900KS的多线程性能提高了一倍多。\n这使我们到了今年。英特尔仍在为延迟10纳米制程工艺而苦苦挣扎,于是发布了i9-10900K,尽管它仍在与9900K和9900KS相同的体系结构上运行,但在单线程和多线程工作负载方面均取得了显着改善。不幸的是,对于Blue Team而言,AMD并没有休息-2020年的Zen 3架构大大改善了Zen 2的单线程性能。\n在这一点上,AMD在单线程性能上处于领先地位。此外,连续第二年,它在多线程性能上遥遥领先,几乎是英特尔所能提供的最好性能的两倍。\n长远来看:2001-2020年的一代一代的改进\n\n看到英特尔在2012年出现反常现象吗?该公司从六核CPU降为四核-而且没有人在意,因为单线程性能得到了相应的提高。\n\nAMD最大的单线程spike对应的介绍,Athlon 64 X2架构在2005年。2007年,英特尔推出了其核心系列产品。\n\n乍一看,此图表与纯多线程之间没有太大区别,但请注意,AMD的趋势线在2016年至2020年之间将保持一致。\n如果我们想看看英特尔和AMD的工程流程如何进行,那么从原始数据继续进行,并观察性能的代际提升是有启发性的-这意味着,今年的情况要比去年好百分之几。\n这些图表放大了两家公司各自的架构变更和工程工作的影响,从而更容易发现重大变更。到目前为止,多线程性能是最大和最令人兴奋的峰值-随着2007年四核CPU的推出,多线程性能也达到了最高性能。在单线程图表上,几乎看不到任何东西。\n我们最喜欢的一代图表版本是第三代,它将单线程和多线程改进平均到一条直线上,并赋予两者相同的权重。尽管乍一看它看上去与纯多线程图表大致相似,但还是会弹出一些有趣的功能。\n第一个跳出这里的是英特尔在2005年的辉煌表现,在图表上只有不到100%的点。低于100%的跌幅可能很小,足以使误差造成波动-但是,即使Pentium 3.8GHz并不比2004年的Pentium 4 3.6GHz主动 恶化,但显然没有任何改善。\n再看看AMD从Zen到Zen +,Zen 2和Zen 3的2017-2020年进展,这很有趣。尽管此处的总体形状与我们在纯多线程图表上看到的相同发音M相同,但包含单线程性能清楚地表明,每个步骤都是超越最后一步的重要且持续的改进。\n2021年及以后\n对于台式机上的英特尔来说,现在形势严峻,而AMD的另一项重大架构改进(Zen 4)则在2021年面世,AMD称其为“按部就班”,其工艺节点缩小至5纳米。\n在英特尔方面,我们预计并不会像从i9-9900K到i9-9900KS那样糟糕。但是,这并不意味着Blue Team充满了光明。尽管i9-10900K确实比前几代产品有了显着且令人惊讶的改进,但它与AMD的兴起没有匹敌-它仍然代表着垂死的架构的最后一口气。\n英特尔的台式机CPU仍停留在14纳米制程上,明年的Rocket Lake也将保持14纳米制程。赛普拉斯Cove架构本质上是英特尔10纳米Sunny Cove笔记本电脑架构的一种变体,被反向移植到较早的14纳米制程。\n尽管英特尔通过其最新的Sunny Cove笔记本电脑处理器(今年至今仅在原型设计中看到了今年的Tiger Lake)获得了显着的性能提升,但其内核和线程数却很少,而高性能似乎与更高的功耗相关。\n现在说10nm Sunny Cove在内核数和功率效率方面的限制是否会转变为14nm Cypress Cove还为时过早。但是,似乎不太可能在2021年看到英特尔像Zen和它的继任者那样带动并很大程度上维持AMD的复兴。\n尽管如此,英特尔的坏消息很可能是性能爱好者的好消息。英特尔这个规模小得多的竞争对手,将需要不止一两个主导周期来巩固与大型OEM和系统制造商的关系和市场认知。这些业务关系——以及由此带来的订单、收入和利润——将是保持AMD强大的研发力量的必要条件,以确保未来AMD能够为消费者提供更高的性能和更好的价格。","news_type":1,"symbols_score_info":{"INTC":0.9,"AMD":0.9,"09086":0.9,"03086":0.9}},"isVote":1,"tweetType":1,"viewCount":1930,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":304936670,"gmtCreate":1605107051790,"gmtModify":1704959126060,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/304936670","repostId":"1125351879","repostType":4,"isVote":1,"tweetType":1,"viewCount":2162,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":304938513,"gmtCreate":1605106987032,"gmtModify":1704959124923,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/304938513","repostId":"1173267429","repostType":4,"isVote":1,"tweetType":1,"viewCount":1792,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":304938363,"gmtCreate":1605106950552,"gmtModify":1704959124277,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/304938363","repostId":"1144133852","repostType":4,"isVote":1,"tweetType":1,"viewCount":1521,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":304931153,"gmtCreate":1605106827872,"gmtModify":1704959122970,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/304931153","repostId":"1172372185","repostType":4,"isVote":1,"tweetType":1,"viewCount":2578,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":304933648,"gmtCreate":1605106754684,"gmtModify":1704959122323,"author":{"id":"3567118012147142","authorId":"3567118012147142","name":"年轻的太太","avatar":"https://static.laohu8.com/default-avatar.jpg","crmLevel":1,"crmLevelSwitch":0,"followedFlag":false,"authorIdStr":"3567118012147142","idStr":"3567118012147142"},"themes":[],"htmlText":"??","listText":"??","text":"??","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/304933648","repostId":"1199283278","repostType":4,"isVote":1,"tweetType":1,"viewCount":2174,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0}],"defaultTab":"posts","isTTM":true}