三泰虎

华为取得突破,实现14纳米级芯片的设计工具国产化

Huawei makes breakthroughs in design tools for 14nm chips -media

华为取得突破,实现14纳米级芯片的设计工具国产化

SHENZHEN, China, March 24 - Huawei Technologies Co Ltd (HWT.UL) has made breakthroughs in electronic design automation (EDA) tools for chips produced at and above 14-nanometre technology, Caiing reported on Friday, citing a speech by a senior executive.

3月24日,中国深圳——财经新闻周五援引一位高层的讲话报道,华为技术有限公司(HWT.UL)在14纳米及以上技术芯片的电子设计自动化(EDA)工具方面取得了突破。

Huawei will complete testing on the tools this year, rotating chairman Xu Zhijun said in a speech on Feb. 28, the Chinese financial news magazine reported. Huawei has developed 78 tools related to chip hardware and software, the report added.

据中国财经新闻杂志报道,华为轮值董事长徐直军在2月28日的一次演讲中表示,华为将在今年完成这些工具的测试。该报道补充说,华为已经研发出78个与芯片硬件和软件相关的工具。

 nimg.ws.126.jpg

以下是reddit网友的评论:

Edenz_

Well this thread stayed up much longer than usual! Must be from some riveting and insightful discussion?

这个帖子比其他帖子的置顶时间长多了!肯定有不少吸引眼球、颇具深度力的讨论吧?

 

 

June1994

I’ve commented that China has most components of the semiconductor supply chain in place already. People here simply don’t believe it, and Western pundits are asleep at the wheel because they don’t read Chinese.

I expect China to have the capability for mass production of fully domestic 28nm by 2025, perhaps even better than that.

我之前就说过中国已经拥有半导体供应链的大部分组件。可惜论坛里的人根本不相信,西方国家的专家们居功自傲,因为他们根本不懂中文。

我预计到2025年,中国将具备完全国产28纳米芯片的量产能力,甚至更进一步。

 

 

Cheeseblock27494356

The most sensitive area at the moment is probably photoresist, where China still has to produce sufficient quantities of it, but that’s a scaling problem, the R&D is already done as far as I can tell.

China might also attain EUV a lot faster than people think. Might have a working low rate production of the entire supply chain for EUV by 2030.

目前最主要的问题可能在于光刻胶,中国还需要生产足够数量的光刻胶,但这也仅是规模问题,据我所知,研发工作已经完成。

中国也可能会比人们想象的更快地实现EUV光刻技术。到2030年,整个EUV供应链可能会实现低速率生产。

 

 

Maleus_

Because usually that is a business thing, no one attempts it because the amount of money needed to be pumped in is not worth the risk of failing and eventual profits.

However, if the government is pum endless money into it because it's a strategic thing, then it's possible.

因为这是商业上的事情,没有人会轻易尝试,因为需要投入大额资金,可能要面临极高的失败风险,这样的投入产出比并不划算。

但如果政府认为这是战略性技术,不断地投入资金,那么是有可能成功的。

 

 

iopq

I don't think EUV will be easy, no Japanese or American competitor to ASML even really attempted it

我认为他们能轻松掌握EUV光刻技术,阿斯麦的日本、美国竞争对手都没有真正尝试过。

 

 

June1994

This isn’t a comparable situation whatsoever. China already makes a ton of its own equipment. AMEC or whatever the wafer etcher company is called, is a global leader in its field.

这不一样。中国已经制造了大量自己的设备。中微半导体设备有限公司或者不管这家晶片蚀刻机公司究竟叫什么名字,已经是该领域的全球领导者了。

 

 

Maleus_

They forgot to spend 40 years beforehand getting the tech from the west and stealing as much as they could. China didn't forget that step, they already produce semiconductors, just not the cutting edge.

他们从西方国家获得技术并竭尽全力盗窃技术时,忘了要先花上40年时间做做研究。中国没有忘记这一步,他们已经生产出半导体了,但不是最先进的那种。

 

 

carpcrucible

Keep in mind that the soviets tried to catch up on semiconductors during the cold war and dumped a lot of money into it, and could never get close.

Not saying it's impossible, but also "spend a lot of money" doesn't guarantee it, either.

请记住,苏联在冷战期间试图在半导体领域迎头赶上,并投入了大量资金,但从未成功过。

我不是说这种事不可能成功,但“靠巨额投资”是无法实现的。。

 

 

Physx32

Wrong. Nikon made experimental EUV machines back in 2012. They would've succeeded if the funding was more.

错了。尼康早在2012年就制造出了实验性的EUV光刻机。如果资金再充裕一些,他们可能就成功了。

 

 

SemanticTriangle

They made a tool, but they ran into an output power problem. They didn't have the funding to resolve that problem, and it looks like it was a fundamental constraint with the the method of light generation they chose.

他们制造出了一个工具,但遇到了输出功率的问题。他们没有资金来解决这个问题,看起来这是他们选择的光振荡方法的一个根本限制。

 

 

stevesbetting

People don't know that EUV can only be sold to a few customers ie TSMC Intel and Samsung.

People here also forget that TSMC Intel and Samsung are ASML shareholders.

If I came up with a nice EUV scanner, who are you going to sell to?

大家不知道EUV光刻机只能卖给少数客户,即台积电、英特尔和三星。

论坛里的人也忘了台积电、英特尔和三星都是阿斯麦的股东。

如果我设计出了一个很好的EUV扫描仪,你会卖给谁?

 

 

stevesbetting

People here seem to forget that making stuff like semiconductors is physics not some Caucasian magic.

大家似乎忘记了制造半导体这种东西是物理学范畴,并非白人的魔法。

 

 

kipperER1

Agreed. It's time for a multipolar world.

同意。现在已经是多极世界了。

 

 

MasatoWolff

And if they do manage to obtain it by 2030, ASML has had 7 more years of development so they will always be behind.

如果他们在2030年之前取得成功,阿斯麦也又有了7年的发展时间,所以他们永远都是落后的。

 

 

discwars

I expect your comment to be deleted and this post to be locked. Unfortunately, nuance or logic doesn’t seem to work on any posts involving China.

可惜只要是涉及中国的帖子,正常的逻辑思维似乎都不起作用。

 

 

hwgod

This head-in-sand denial is just laughable. When has pretending that something doesn't exst ever made it go away? God forbid someone acknowledge that there are smart people in China too, but I guess that's too much for the racists to stomach.

这种鸵鸟似的把头埋进沙子里否认事实的做法太可笑了。假装某样东西不存在,就能让它消失吗?上帝严禁人们承认中国也有聪明人,我想这对种族主义者来说太过分了。

 

 

Tyreal

It’s easy when you can just steal everything.

什么都可以偷,那当然很简单。

 

 

AnimalShithouse

You steal enough talent and tech, and it is bound to happen. I would echo that EUV will be hard for China to be "good at" before 2030, though.

如果你窃取了足够多的人才和技术,那么肯定能成功。我想说中国在2030年之前很难“拿下”EUV光刻技术。

 

 

More_Theory5667

EDA tools were probably always going to be one of the first parts of the sanctioned supply chain they make a breakthrough on as it has less to do with hardware.

EDA工具可能一直是他们在受制裁的供应链上取得突破的第一步,因为它跟硬件关系不大。

 

 

Yumewomiteru

It's really impressive how well Huawei has weathered the US sanctions, they've proven that US sanctions will not kill a great company.

华为安然躲过了美国的制裁,令人印象深刻,他们已经证明了美国的制裁无法根除一家优秀的企业。

 

 

hwgod

You should look at where the articles in the top scientific journals are coming from these days.

你应该看看现在顶级科学期刊上的论文都是来自于哪个国家的。

 

 

hwgod

China is great at stealing shit and hacking it to work, but they are really bad at original research and new discoveries.

中国很擅长偷窃,但他们真的不擅长原创研究和新发现。

 

 

SonOfHonour

Good luck competing in the modern world with such inferior technology.

We live in the age of technology, the levels of computing power a country has determines EVERYTHING.

你们在现代世界还在使用如此落伍的技术,呵呵。

我们生活在技术时代,一个国家的计算机能力水平决定了一切。

 

 

SonOfHonour

Self sufficiency is never enough if you want to be dominant.

如果你想称王称霸,只是自给自足是远远不够的。

 

 

iopq

You can run a webserver on 28nm, but only if fewer people want to use it. The scale of web servers in China is insane and they are all produced at TSMC and intel. In case of a full on trade blockade Chinese businesses would not be able to serve their customers the way they are now

你可以用28纳米级芯片运行web服务器,但前提是使用人数要少。中国网络服务器的规模十分吓人,都是由台积电和英特尔供应的。如果实施全面的贸易封锁,中国企业就无法像现在这样为客户提供服务了。

 

 

Maldiavolo

Huh? Samsung is only in the current chips game because they got 16nm delivered to them from a TSMC employee before TSMC reeled him back using the laws on the books and his contract. Samsung has had so many issues scaling downwards since then. Then we have Intel who has been stuck at 10nm for what seems like forever. There is no reality outside of TSMC that 7nm is an easy jump.

啥?三星之所以能杀进芯片市场,是因为他们从台积电的一名员工那里拿到了16纳米芯片,然后台积电利用法律和他的合同把他骗了回来。从那以后,三星在技术升级的过程中遇到了很多问题。还有英特尔,他们似乎在10纳米级别上举步维艰了。除了台积电,没有企业能轻松实现7纳米技术的飞跃。

 

 

uKnowIsOver

Everything is possible with money,resources,manpower and the right management. Look at TSMC, before the Taiwanese government and Apple money and engineers began to pour, they were a B tier fab at best with a very bad track record of axed or garbage nodes like their 20 and 40nm ones.

只要有资金、资源、人才和优秀的管理,一切皆有可能。看看台积电,在台湾政府和苹果的资金和工程师进入之前,充其量只能算事一家B级晶圆厂,连20和40纳米芯片都搞不定。

 

 

stevesbetting

Correct . It also becomes easy when you have a customer with deep pockets like Apple or Huawei giving you feedback and iterating until you get it right.

没错,如果有苹果或华为这样财力雄厚的客户不断提供反馈,并不断迭代,当然不难。

 

 

Teftell

Great, less chip monopoly.

太好了,破除芯片垄断。

三泰虎原创译文,禁止转载!:首页 > 大国 » 华为取得突破,实现14纳米级芯片的设计工具国产化

()
分享到: