转发: 99%死定了?华为员工曝“备胎”芯片真相

bonjour

资深人士
注册
2004-02-01
消息
2,129
荣誉分数
644
声望点数
273
5月17日凌晨,华为海思总裁何庭波给员工写信称多年芯片备胎"一夜转正",人民日报等中共媒体纷纷转发。然而,华为芯片真的可以完全替代进口吗?希望之声记者获取的一份华为资深员工的文章,或者可以让您一睹究竟:

我在华为的时候,有一段时间在中软研究院工作。其中有几个部门, 是专们做备胎的,比如尽管华为在用谷歌的安卓,但自己同时也在开发手机操作系统。感觉很不解,现代科技都是分工协作,所有的企业都是攻自己的专项,依靠其他公司提供上下流产品。大而全是一件效率很低的事情。分在备胎技术部门的兄弟,工作看不到前景,没有效益,分红也比主流部门的少,工作自然不安心,产品可想而知如何了。后来明白,任老闆知道帝国主义跟咱尿不到一个壶里,早就存了分道扬镳的心,看到中兴的下场,于是早早做了准备。反正中国人工便宜, 搞个人海战术,广泛撒网,没准就捞着一条大鱼了。

备胎是准备了,但这备胎能不能用,好不好用,就要另说了。现在硬软件都是一个复杂系统,开发不是闷在黑屋,闭门造车就能折腾出来的。产品开发要经过长期的试错。首先你得需要一个环境,就是我们常说的生态系统。有上下游产品支持,有成千成万的用户。否则你的产品都是Bug,没有用户会有耐心花钱陪你玩。

芯片制造包括指令集架构和其他设计 IP 专利,测试设备和工具。比如龙芯的 MIPS 指令架构和 Android 手机上的 ARM 指令架构。这些都需要国外厂商的授权和许可证。你可以埋头苦干四五年,发展自己的新指令架构,做到不仰人鼻息。但是如果你的独门秘笈没有其他人跟随,生态系统不能发展,上下游(编译器、操作系统、芯片方案、终端、应用程序等等)都得不到开发,最后必然是死路一条。类似的例子像中国的 3G 无线移动系统 TD-CDMA。

另外,很多应用芯片就是计算架构和计算算法的硬件化。大多架构和算法中国都没有知识产权,需要得到国外厂商的授权和许可证。

华为的麒麟是 SoC(System on a Chip)芯片,包括 CPU, GPU 和无线电基带和射频。其中 CPU 和 GPU 用的是 ARM 的架构, 基带和射频使用很多高通的技术和专利,这都需要向国外厂商购买或者得到授权,这也是麒麟不外卖的原因,因为没有授权。央视说百分之一百的中国知识产权, 绝对是忽悠。

仔细读了华为海思总裁致员工的信,一看那语言,还真是华为的风格。 风萧萧兮易水寒,壮士出征兮不复还,那是走夜路,吹口哨,自己给自己壮胆。特别是海思半导体公司关于启动所谓备胎芯片的那封信, 对于业内人士来说,就是在忽悠、打鸡血的。也许写这封信的人根本就是个外行。

设计芯片和设计线路板使用芯片,是两回事。以为自己的团队已经开发了三五成的替代芯片,或是只是设计出了核心功能芯 片,虽然功能差一些,觉得也能凑合着用。但是一个产品上的线路板, 少一个芯片就没办法生产。而一个产品少则几十,多则成百上千个芯片,其中大多都是通用芯片,是欧美公司几十年的累积经验研发出来的。我相信中国人的聪明才智,华为更是聚集了国内菁英。如果华为集中人力,花几年的时间,开发出一个管脚和功能相同的通用芯片, 这已经是很了不起的成就了。现在国内自己研发的芯片,几乎都是专用芯片 ASIC (Application-Specific Integrated Circuit),也就是根据自己产品的需求,量身定制的。这种芯片不愁销路,没有竞争对手,随着自家产品而进入市场,而其芯片本身并没有商品的竞争对手、价格、 供需等特性。

大公司对于产品选用芯片的流程几乎相同,首先设计所用的元器件(包括芯片、电阻、电容器、晶振、二极管等等)都需要从 AVL(Approval Vendor List)的厂家中选取,除了保障产品质量,还有一 点就是保障供应的可靠连续性。除此之外,这些元器件大多需要 second source, 就是管脚功能一模一样的第二、第三个厂家。只有极 个别的可以例外,如 Intel 的 CPU、PCH,Broadcom、Marvell 的通 信芯片系列等。除了它们的芯片提供主要功能,独一无二的之外,还有就是它们都是芯片大厂,可靠性有保障。

对于其它辅助功能的芯片, 虽然有两三个厂家,但也常碰到芯片 EOL(end of life)的通知,这时候通常根据自己产品预计生产的年限和所需量做一次性的 LTB (last time buy)。当然如果买的芯片不够,或预计自己的产品的生命周期在几年内都不会结束,那么这时则需要找功能类似的芯片,修改原理图和线路板,同时也要修改 BOM(build of material) 。

等新的线路板生产回来后,还需各种测试(dvt,dmt,emi?)以及后续的 report,快则两三个月,慢则半年。这还只是少一个芯片,并能找到类似功能的情况。如果找不到类似功能的芯片,那么这个产品就可以洗洗睡了。

对于华为这样一个使用大量通用芯片的公司,断芯就意味着断粮断水。 海思是绝无丝毫可能设计出满足华为产品的全部芯片,甚至于连一半所需芯片也设计不出来。如果误以为海思把华为所需的关键核心芯片的备胎设计出来,就可以救华为,那就是太无知了。少了不起眼的辅助芯片,如供电芯片、桥芯片、数模转换芯片等等,就像少了空气一 样,虽然平常不觉得它存在,但一旦缺失,就会窒息而死。

所以只要美国断芯(美国占有全世界高端芯片百分之八十以上), 华为是百分之九十九死定。那百分之一生存的可能,就是中国闭关锁国。对了,还忘了说芯片设计辅助软件,美国在这方面几乎垄断了。

现代每个电子产品都是一个系统,包括几万到几十万个芯片。一个公司不可能全部自己开发所有的芯片,特别是通用芯片,必须依靠其他专业厂家。从经济上说,也是不合算的。世界上最有钱的苹果公司, 为了基带,不得不跟高通和解。而半导体领域的老江湖英特尔,能造出世界上最好的 CPU,却造不出一个兼容性没有问题的基带。华为可能在一些 ASIC 和少量通用芯片上开发备胎,但绝不可能,也不会在所有的通用芯片上自己开发备胎。一个产品只要被断供 5%的芯片, 找不到其他厂商的替代品,肯定玩完。至于大部分智能手机制造商之 所以使用安卓操作系统,是因为其广泛的使用人数和完整的应用生态系统。过去包括三星电子在内的几乎所有智能手机制造企业都曾投入巨资开发代替安卓的操作系统,但无一例外,全部失败了。任老闆要所有芯片、软件、操作系统全部自主研发,那真是要创造人类科技史上的奇迹了。
 
虽然话不中听,但是的确有些道理。
没经过实用的东西不会是好东西。看华为任老板如何破局吧。
 
"芯片其实是简单的东西,它就是在硼、硅、磷这三个简单的原子组成的化合物上,不断地用激光进行切割"。。。

:good:

可见金灿荣又露一个破绽:说中国禁了稀土出口,美国芯片就做不成了,因为稀土是芯片的原料。这是扯JB淡。好多反驳金五毛的也接受了他这个陷阱,在陷阱之上批驳他。。。 芯片确实用到一些元素,但是没有一种是稀土元素,芯片和稀土没有五毛钱关系。
我虽然也看不上金灿荣, 但是, 我也不喜欢你们这种铁口直断的说法。

是骡子是马, 牵出来溜溜嘛!

金说禁止稀土出口能够打击美国芯片;
你说不能打击美国芯片。

那就禁止出口嘛! 大家密切观察就是了, 看看到底中国稀土的出口与否, 是不是美国芯片的一个致命伤。

芯片和稀土有没有五毛钱关系, 你做几个实验不就知道了? 比如, 先禁止10年, 看看结果如何。
 
最后编辑:
我虽然也看不上金灿荣, 但是, 我也不喜欢你们这种铁口直断的说法。

是骡子是马, 牵出来溜溜嘛!

金说禁止稀土出口能够打击美国芯片;
你说不能打击美国芯片。

那就禁止出口嘛! 大家密切观察就是了, 看看到底中国稀土的出口与否, 是不是美国芯片的一个致命伤。

芯片和稀土有没有五毛钱关系, 你做几个实验不就知道了? 比如, 先禁止10年, 看看结果如何。
稀土这招中国用过,事实证明根本不好用
2010年在钓鱼岛中日两国闹翻,中国向日本禁运过稀土,结果反而促进了日本发展稀土替代技术,禁运前日本每年进口稀土超过3万吨,到了2012年锐减到1.4万吨,中国的稀土反而砸手里卖不出去了,禁运前国际市场一公斤稀土400美元,到了2012年,禁运了两年多反而暴跌到100美元,中国不得不灰头土脸地恢复了稀土供应
 
习近平访问一家稀土公司,跟稀土相关的股票已经疯涨,说明什么?

市场比瞎BB的你们聪明很多很多。
 
稀土这招中国用过,事实证明根本不好用
2010年在钓鱼岛中日两国闹翻,中国向日本禁运过稀土,结果反而促进了日本发展稀土替代技术,禁运前日本每年进口稀土超过3万吨,到了2012年锐减到1.4万吨,中国的稀土反而砸手里卖不出去了,禁运前国际市场一公斤稀土400美元,到了2012年,禁运了两年多反而暴跌到100美元,中国不得不灰头土脸地恢复了稀土供应
时代不一样了。

而且当时也没继续坚持禁运。再禁运个10年你看看,一切都会改变的。

关键是要拿出勇气来,坚持禁运!
 
interesting article. lol

HUAWEI’S ANDROID AND WINDOWS ALTERNATIVES ARE DESTINED FOR FAILURE
https://www.theverge.com/2019/5/23/18636841/huawei-android-windows-alternatives-europe-us-editorial
Not much is known about Huawei’s Android and Windows alternative, but it appears to be based on the open source version of Android (AOSP) and will include Huawei’s App Gallery store. This is Huawei’s alternative to Google’s Play Store, and every manufacturer that doesn’t license Google’s version of Android has to create its own app store or bundle one from a myriad of fragmented options. Huawei already has experience here, as the company’s phones run a forked AOSP-based version of Android without the Play Store in China, and Huawei’s been bundling the App Gallery on phones outside of China since early 2018.

Outside of China, phones running alternatives to Android and even those using AOSP haven’t fared well. Mozilla tried with its Firefox OS for years before giving up in 2015, Canonical pushed Ubuntu phones that never went anywhere, and Microsoft famously tried to create a third mobile operating system with Windows Phone. Even Samsung, once a big threat to Google’s version of Android, has all but given up on its Tizen operating system for phones, using it to power the company’s smartwatches and TVs instead. And let’s not even talk about what happened to BlackBerry.

All of these phone OS alternatives have failed for many different reasons, but chief among them is a common thread: competing with Google is very difficult. Google’s search market share is estimated at around 90 percent worldwide, with competitors like Bing, Yahoo, Baidu, and Yandex all making up single digits. This search market share has helped Google create and control a suite of apps like Chrome, Gmail, YouTube, Google Maps, Google Docs, and many other popular web services.

If you create a phone running the open source version of Android, it immediately comes without access to these key Google apps. As a manufacturer, you’re creating a device without the apps consumers demand in Europe, the US, and elsewhere.

moto g google play store japan line
THE PHONE CHALLENGE
This is the challenge Huawei now faces on the phone side. The only companies that have come close to creating a viable Android alternative outside of China are Microsoft or Amazon, and Amazon’s attempt is now limited to tablets. Microsoft created Windows Phone and was able to gain some impressive ground in certain countries in Europe, but Android still absolutely dominated overall. Microsoft took the approach of licensing its more closed operating system to phone makers for a fee per handset, while the “free” Android alternative came with its own royalty payments and a lot of customization options for phone makers and carriers.

Microsoft tried to offer a phone operating system without the support of Google’s apps and services. Windows Phone users were cut off from Google Maps temporarily, there was a bitter battle over a Windows Phone YouTube app, and Google surprised Microsoft by cutting off Gmail’s Exchange ActiveSync support for Windows Phone. Google also refused to develop Windows apps, undermining Microsoft’s mobile efforts to compete with Android.

Amazon has seen some limited success with its own implementation of Android. The company has created its own Android app store for its Fire tablets, and it’s managed to convince some app makers to list their popular apps in the store. Facebook, Spotify, Netflix, Skype, HBO Now, and others are all available, but Google’s apps are, predictably, nowhere to be seen. Amazon’s app store is still missing a lot of key apps, and even the ones available aren’t always updated at the same time as their main Android alternatives. Thankfully, the Fire tablet screen size lends itself to using the browser to access Google services more than the smaller display on phones. Amazon tried to compete with Android with its Fire Phone, but the company discontinued it quickly after poor sales.

Huawei P30 Pro
GOOGLE HAS AN IRON GRIP ON ANDROID
Google also has an iron grip on the definition of an Android device, including the open-source version of Android that Huawei will be using to compete. Most of the built-in alternatives to apps like search, Gmail, camera, calendar, Chrome, and even the keyboard are incredibly basic. Even a number of APIs like location, gaming, and in-app purchasing are proprietary, and third-party app developers use these in their apps. That makes it more difficult for developers to then have to support two different versions of their Android app, not knowing whether Amazon, Huawei, or anyone else has created replacements that are good enough.

Google has also helped stem the flow of fragmentation and open-source forks of Android outside of China in the past by bundling its own apps with access to the Play Store and requiring companies building phones or tablets that include the Play Store to only build phones and tablets that include the Play Store.

Huawei has leverage as the world’s second-largest phone maker to influence the future of Android, but Samsung had the same five years ago and was forced back into line. Google wasn’t impressed with Samsung’s Android software implementations back in 2014 and a series of meetings led to the two companies announcing a broad patent cross-licensing deal and an agreement on what the future of Android would look like.


GOOGLE IS FACING ANDROID CHALLENGES IN EUROPE
Huawei has been preparing for a similar moment for years, and now the US-China trade war has arrived to complicate the company’s relationship with Android. Huawei is well aware of the challenges, though. The company has been building a Play Store alternative in plain sight, and it reportedly pitched app makers last year on creating apps for its store with the offer of helping them make inroads into China and a “very significant” share of the app store’s revenue. Bloomberg reports that Huawei even claimed it would have 50 million European users of its app store by the end of 2018.

Huawei’s store plans and discussions come at a pivotal time for Android in Europe. Google could finally face some competition after a European Commission lawsuit fined the company over Android antitrust violations. Regulators allege that Google abused its Android market dominance by bundling its search engine and Chrome apps into Android, blocking phone makers from creating devices that run forked versions of Android, and making payments to manufacturers and network operators to bundle the Google search app on handsets.

As a result, Google will start charging Android device makers a fee for using its apps in Europe. This could open the door to rival app stores, a more competitive landscape for Android, and inevitable fragmentation. The more likely result is that manufacturers will simply continue to bundle google’s apps and services, since it’ll allow them to avoid those fees. There are still no popular alternatives to YouTube, Google Maps, or Google Search, after all, and consumers across Europe will reject phones that don’t have access to these apps. Phone makers also aren’t likely to want to maintain different versions of Android for Europe, China, the US, and elsewhere.


THE LAPTOP CHALLENGE
Over on the Windows side, Microsoft has maintained its dominance on desktop computing for more than two decades using the same kinds of bundling tactics as Google. US regulators famously grilled Microsoft over its bundling of Internet Explorer in Windows, and EU regulators also got involved some years later. The EU eventually forced the company to include a browser ballot with non-Microsoft browsers in an effort to improve competition.

Microsoft was also accused of illegally bundling its Windows Media Player with Windows, and the EU forced it to unbundle the app so that competitors could get a fair playing field. Microsoft created a special version of Windows for Europe without the app, but barely any manufacturers actually shipped machines with this version.

CHROME OS LOOKS LIKE THE MOST LIKELY WINDOWS ALTERNATIVE
There have been various forms of competition to the Windows dominance, including Macs, Linux-based netbooks, or even Google’s Chrome OS. Only mobile operating systems have managed to shake Microsoft’s dominance of computing in general, but Windows is still heavily used by businesses worldwide. Google’s Chrome OS looks like the most viable alternative for the masses, thanks to Google’s backing and its popular web services.

Many businesses still rely on Windows for its app compatibility, and to run apps and systems that aren’t just web-based. This has held back Microsoft from progressing with some of its own ARM-based laptop efforts, simplified versions of Windows, and even more restricted ones.

Huawei’s Plan B operating system would have to compete against both Windows 10 and Chrome OS in a market controlled by Lenovo, HP, Apple, Dell, Acer, and others. These companies have far more experience making and shipping PCs, and already have trusted brand recognition. We’ve seen Android-powered laptops arrive and disappear over the years, and it looks like Huawei is going down a similar path. If building an Android alternative for phones seems challenging, Huawei’s Windows alternative could be even harder to pull off in a market where it holds less sway.

Huawei now faces many tough decisions that are largely out of its control. None of this software even matters if the company doesn’t have the chips it needs. That’s a thorny problem that Huawei hasn’t shown any indication of being able to overcome. The ARM, Intel, and Qualcomm situations are far more dangerous for Huawei, but even without them, trying to compete with Google and Microsoft’s dominant operating systems seems an insurmountable challenge for a company better known for its hardware outside of China than software advancements
 
其实安卓根本不是大事,就算手机不做了华为还是华为。芯片器件断供才是硬伤。
 
牛拉多纳一个人技术再好,一帮猪队友,他很难有所作为。
华为这次会给人剁手剁脚,去掉大半条命。
不过靠走私芯片和国内市场,大概也死不了
 
后退
顶部