AMD and Intel slug, but they are convinced that Moore's law is not dead-Moore-IT information
AMD and Intel in years in to be cut-throat, K10, begins after the bulldozer processor can't keep up with times, has been losing money with Intel AMD can't do, especially in semiconductor technology, AMD no funding, no frequent upgrade strength process directly stopped playing. However, despite the different, but look at semiconductors golden rule "Moore's law" on the matter, AMD and Intel is the same--they don't think Moore's law is dead, it will continue to develop.
▲ AMD, Intel to fight to fight, but, in Moore's view on line
About Moore law a thing, we zhiqian also reported had times has--it is Intel veteran level founder one of of Gordon Moore in 1965 proposed of, he in a article articles in the predicted semiconductor chip Shang of transistor number will to annual turned times of speed growth, but this growth now has not too may has, so Moore law has had amendment, Intel official recognized of Moore law is "each 24 months transistor number turned times", this is Intel famous of " Tick-Tock "strategy was based on 2 years upgrading manufacturing processes.
Moore last year ushered in 50 years, 50 years ago, the law still can guide the development of the semiconductor industry? This issue has been discussed for many years, industry insiders view into two factions-the one that semiconductor development has slowed, impossible 2 years a technology upgrade, Moore's law is dead, Intel presented as Moore's naturally disagreed, that Moore's law will continue to exist, they are also trying to practice it.
So as the main rival AMD Intel how to think? AMD CTO Mark Papermaster also answered this question in an interview a few days ago, he believed that Moore's law was not dead, and live very well, he said that only a narrow-minded person would think technology upgrades are only related to transistor size.
He said, "it no longer have to do with transistors, each node, we cannot achieve improved transistors. It does bring the transistors improved, but also with our made in design, architecture, and how to mix with solution and so on, these guarantees Moore continues to advance. ”
PS: whatever Intel on the mouth to admit it or not, Moore really was a big challenge, 10nm and technology upgrade cycle after getting longer and longer, transistors also may not always be obvious upgrade, AMD said, Moore's law is not death improvements outside of the transistor.
AMD跟Intel争得你死我活，但他们都坚信摩尔定律未死 - 摩尔定律 - IT资讯
这么多年来AMD跟Intel斗得你死我活，从K10、推土机处理器之后开始跟不上趟了，一直亏损的AMD根本无力再跟Intel对干，特别是在半导体工艺上，AMD没资金、没实力频繁升级制程，直接甩手不玩了。不过双方虽然道不同，但在看待半导体金科玉律“摩尔定律”一事上，AMD的态度跟Intel倒是一致的——他们也不认为摩尔定律已死，它还会继续发展下去。
▲AMD、Intel争来争去，但在摩尔定律上看法一致
有关摩尔定律一事，我们之前也报道过多次了——它是Intel元老级创始人之一的戈登·摩尔在1965年提出的，他在一篇文章中预言半导体芯片上的晶体管数量会以每年翻倍的速度增长，不过这个增长速度现在已经不太可能了，所以摩尔定律有过修正，Intel官方认可的摩尔定律也是“每24个月晶体管数量翻倍”，这也是Intel著名的“Tick-Tock”战略的基础，2年升级一次制造工艺。
摩尔定律去年迎来50周年，50年前的定律到底还能不能指导半导体业发展？这个问题已经讨论过多年了，业内看法分成两派——一派认为半导体发展速度已经放缓，不可能再2年实现一次工艺升级，摩尔定律已经死了，而Intel作为摩尔定律的提出者，自然反对这种看法，认为摩尔定律还会继续存在，他们也在努力实践这一点。
那么作为Intel主要对手的AMD如何看呢？AMD CTO Mark Papermaster日前在采访中也回答了这个问题，他认为摩尔定律并没有死，而且活得很好，他说只有思路狭窄的人才会觉得工艺升级只跟晶体管大小有关。
他说“它不再与晶体管有什么关系，我们不可能在每个节点都实现晶体管改进。它确实会带来晶体管改善，但也与我们在设计、架构中所做所为以及怎样把解决方案混合在一起等有关，这些保证了摩尔定律继续前进。”
PS：不管Intel嘴上承认与否，摩尔定律其实真的受到了很大挑战，10nm及之后的工艺升级周期越来越长，晶体管也不可能每次都明显升级，AMD所说的摩尔定律不死说的也是晶体管之外的改进。