許多跨國(guó)芯片公司會(huì)選擇低調(diào)地將部份工作轉(zhuǎn)移到中國(guó)，然而，也有一些業(yè)界領(lǐng)導(dǎo)公司選擇更加深入中國(guó)市場(chǎng)，而且，他們從不避諱談?wù)撈渲袊?guó)策略。 德州儀器(Texas Instruments, TI)，就是一個(gè)很好的例子。 德州儀器上海微控制器(MCU) 設(shè)計(jì)中心最近剛完成了首款本地設(shè)計(jì)的芯片tape out案例。雖然TI并不透露其上海設(shè)計(jì)中心團(tuán)隊(duì)規(guī)模，但TI微控制器副總裁Scott Roller在最近的一次接受本刊采訪(fǎng)時(shí)表示，“上海MCU設(shè)計(jì)中心是一個(gè)規(guī)模頗大的團(tuán)隊(duì)，有許多設(shè)計(jì)正在進(jìn)行?！? TI的上海MCU設(shè)計(jì)中心是在2011年初開(kāi)始運(yùn)作，除了TI位于德國(guó)、邦加羅爾和達(dá)拉斯以外，上海是目前最新一個(gè)設(shè)計(jì)中心。 值 得注意的是，TI在中國(guó)的工廠(chǎng)并不只是為了支持現(xiàn)有的MCU產(chǎn)品而存在。相反地，它主要負(fù)責(zé)一些來(lái)自中國(guó)的MCU產(chǎn)品線(xiàn)開(kāi)發(fā)工作?！霸谥袊?guó)，我們開(kāi)發(fā)專(zhuān)門(mén) 為中國(guó)市場(chǎng)而設(shè)計(jì)的微控制器，”Roller說(shuō)。這個(gè)設(shè)計(jì)中心聚集了所有必要的工作人員，包括銷(xiāo)售、應(yīng)用軟件開(kāi)發(fā)、系統(tǒng)和處理器設(shè)計(jì)工程師，以及現(xiàn)場(chǎng)應(yīng)用工程師等。“我們包含了前端到后端的完整流程，”Roller說(shuō)。

德州儀器微控制器副總裁Scott Roller

為何在中國(guó)設(shè)計(jì)？ 當(dāng)被問(wèn)及為何選在中國(guó)設(shè)計(jì)時(shí)，Roller給了我兩個(gè)理由：“首先，你可以更靈活，速度更快。其次，選擇在本地進(jìn)行設(shè)計(jì)，將大幅減少因誤解而出錯(cuò)的機(jī)率。” 盡管理論上來(lái)說(shuō)，此舉可能深具意義，但并不是每一家跨國(guó)公司都愿意公開(kāi)承諾他們將在中國(guó)進(jìn)行設(shè)計(jì)。 月前，在美國(guó)德州San Antonio舉辦的飛思卡爾技術(shù)論壇(Freescale Technology Forum)上，飛思卡爾新任總裁暨CEO Gregg Lowe 指出，中國(guó)市場(chǎng)正在從低成本的電子產(chǎn)品制造基地，轉(zhuǎn)變?yōu)榫邆湓O(shè)計(jì)能力，可為當(dāng)?shù)貜V大內(nèi)需市場(chǎng)開(kāi)發(fā)所需產(chǎn)品的真正高科技樞紐。 想在中國(guó)銷(xiāo)售產(chǎn)品的芯片供貨商，就必須在當(dāng)?shù)卦琇owe說(shuō)。IC供貨商并不一定要在中國(guó)設(shè)計(jì)產(chǎn)品，但他們必須在當(dāng)?shù)卦O(shè)有應(yīng)用工程師和系統(tǒng)工程師，他表示。 然而，ARM中國(guó)總裁吳雄昂說(shuō)，中國(guó)的根本區(qū)別就在于中國(guó)的速度。就他觀察，一般在中國(guó)的SoC設(shè)計(jì)周期會(huì)更短?！皬脑O(shè)計(jì)開(kāi)始到真正tape out，有時(shí)只需要五到六個(gè)月。中國(guó)公司做決定的速度非?？?，對(duì)市場(chǎng)的反應(yīng)速度更迅速?！? 而在進(jìn)行一款特殊設(shè)計(jì)時(shí)，跨國(guó)公司通常要回報(bào)總部做決定，這種做法并不適合中國(guó)本地市場(chǎng)，吳雄昂說(shuō)。 TI的Roller強(qiáng)調(diào)，出于成本考量，TI過(guò)去并沒(méi)有在中國(guó)建立MCU設(shè)計(jì)中心。他指出，中國(guó)的工程技術(shù)人才可能會(huì)“便宜一點(diǎn)，但你不會(huì)真的省下很多錢(qián)?！倍瘢琓I在中國(guó)成立設(shè)計(jì)中心的主因，是因?yàn)椤拔覀兿Ｍ茉O(shè)計(jì)和生產(chǎn)出更貼近本地需求的產(chǎn)品?！? 對(duì)TI來(lái)說(shuō)，現(xiàn)階段中國(guó)MCU市場(chǎng)中一個(gè)最關(guān)鍵的領(lǐng)域，就是中國(guó)正在發(fā)展的國(guó)家電網(wǎng)計(jì)畫(huà)所帶動(dòng)的智能電表。TI已經(jīng)推出了首款產(chǎn)品，目前已提供樣品，Roller說(shuō)?！斑@對(duì)我們來(lái)說(shuō)，是非常龐大的商機(jī)?！? TI在中國(guó)銷(xiāo)售的兩款主力MCU產(chǎn)品分別是超低功耗16位微控制器 MSP430，以及32位的C2000微控制器。 當(dāng)問(wèn)到還有哪些跨國(guó)公司也大力投入中國(guó)MCU市場(chǎng)時(shí)，Roller提到了飛思卡爾(Freescale)和瑞薩(Renesas)。 飛思卡爾已進(jìn)軍中國(guó)多年，TI的Roller說(shuō)。但他也表示，在新的CEO上任后，情況可能會(huì)改變。Roller口中的新任CEO指的是從TI跳到飛思卡爾的Lowe。 飛思卡爾的Lowe任職于TI時(shí)，曾帶領(lǐng)TI位在達(dá)拉斯的模擬事業(yè)部。而TI已經(jīng)在中國(guó)設(shè)立了數(shù)個(gè)模擬設(shè)計(jì)中心。 在飛思卡爾技術(shù)論壇中，飛思卡爾的Lowe在接受采訪(fǎng)時(shí)承認(rèn)，“設(shè)計(jì)決策將會(huì)逐漸轉(zhuǎn)移?！彼硎?，“你不能只待在硅谷，還有其它重要的地方得去，但你必須了解客戶(hù)的需求，才能設(shè)計(jì)出真正貼近客戶(hù)的芯片。 距我首次踏上中國(guó)土地已經(jīng)有一個(gè)月了。這段時(shí)間以來(lái)，我陸續(xù)把我對(duì)中國(guó)的一些觀察寫(xiě)成報(bào)告，分享給讀者們。接在，也在網(wǎng)站我文章的留言評(píng)論版上，看到很多朋友、同事和讀者對(duì)這一系列中國(guó)報(bào)導(dǎo)的反映。事實(shí)上，每一則留言所揭露的信息，都比我的報(bào)導(dǎo)要多得多。 不過(guò)，我完全沒(méi)有被冒犯的感覺(jué)，因?yàn)槲疑钣型小? 本文授權(quán)編譯自EE Times，版權(quán)所有，謝絕轉(zhuǎn)載 本文下一頁(yè)：中國(guó)功夫，唯快不破

相關(guān)閱讀：
• 聯(lián)發(fā)科：讓千元智能機(jī)擁有iPhone 4S的功能
• 本土無(wú)晶圓廠(chǎng)業(yè)者們真正需要的策略
• http://www.zgwwsyw.com/ART_8800121647_1100_2700_0_4300_dde103af.HTMn0lesmc

{pagination} 寫(xiě)在上文之后 今天早上，我收到一位現(xiàn)居北京的前同事來(lái)信：“我很欣賞因?yàn)閵叺闹袊?guó)系列報(bào)導(dǎo)而在網(wǎng)站上出現(xiàn)的所有留言。這些留言有些很有趣，但也有些令人不安......而令人感到可怕的，是雙方對(duì)彼此的了解程度竟然這么低?！? 我完全同意。 對(duì)包括我在內(nèi)的大多數(shù)人來(lái)說(shuō)，中國(guó)在很大程度上仍然是相當(dāng)神秘的。我知道的非常少，盡管我采訪(fǎng)了很多人，雖然我知道他們的真實(shí)故事，但我對(duì)他們的認(rèn)識(shí)程度仍然有限。 同樣，我也感覺(jué)到中國(guó)人對(duì)我們的認(rèn)識(shí)是有限的，尤其是在談到美國(guó)工程師們感受到對(duì)中國(guó)的恐懼時(shí)。 雙方對(duì)彼此都抱持著先入為主的觀念，將傳聞(特別是對(duì)自己有利的消息)當(dāng)作加強(qiáng)自己認(rèn)為足夠了解對(duì)方的左證。 現(xiàn)在，是我們深呼吸，各退一步的時(shí)候了。 我先舉幾個(gè)例子。每當(dāng)我寫(xiě)到有關(guān)中國(guó)的設(shè)計(jì)和生產(chǎn)速度時(shí)──我一直認(rèn)為這是中國(guó)獲勝的關(guān)鍵──此時(shí)，我就會(huì)接到讀者的響應(yīng)，大多是批評(píng)中國(guó)產(chǎn)品的品質(zhì)。這類(lèi) 批評(píng)非常多，我得到許多的反饋，不只是在網(wǎng)站留言，甚至在不經(jīng)意的對(duì)話(huà)間，都會(huì)有人說(shuō)出誰(shuí)才剛剛在中國(guó)買(mǎi)了一部 Android 手機(jī)，但卻故障了….諸如此類(lèi)的評(píng)論。 當(dāng)我在北京拜訪(fǎng)一位美國(guó)芯片公司的工程部副總裁時(shí)，他對(duì)我說(shuō)：“順子，這些手機(jī)的品質(zhì)并不好。但這不是因?yàn)樗麄儾荒茏龊?，而是因?yàn)樗麄儾幌搿！?我對(duì)我聽(tīng)到的話(huà)有點(diǎn)難以理解，因此，我問(wèn)他：“你是說(shuō)，他們是故意做出品質(zhì)不大好的手機(jī)嗎？” 不，他說(shuō)?！八麄冎粚?duì)如何更快速地賣(mài)出更多手機(jī)感興趣。因而他們并不想花時(shí)間做出一只完美的手機(jī)。” 他接著以蘋(píng)果產(chǎn)品為例指出：“你想我的 iPhone 來(lái)自哪里？它們都是由富士康(Foxconn)在中國(guó)的工廠(chǎng)所生產(chǎn)的。所以，中國(guó)人并不是不知道如何制造品質(zhì)優(yōu)良的產(chǎn)品?！? 今天下午，我又收到一封來(lái)自ARM中國(guó)區(qū)總裁吳雄昂的電子郵件。他是因?yàn)槲疑衔闹械膱?bào)導(dǎo)給我寫(xiě)信。他對(duì)我說(shuō)：“我對(duì)你的“中國(guó)速度”一說(shuō)完全同意。產(chǎn)品不一定要在中國(guó)設(shè)計(jì)，但企業(yè)必須有能力響應(yīng)“中國(guó)速度”，才能在這個(gè)市場(chǎng)獲得成功?！碧貏e是在中國(guó)已然成為芯片產(chǎn)業(yè)的關(guān)鍵市場(chǎng)之際，他解釋道。 吳雄昂指出，“如果你去觀察硅谷的新創(chuàng)公司，我會(huì)說(shuō)除了熱情和承諾等要素以外，“速度”、快速做出決策以便對(duì)市場(chǎng)需求做出反應(yīng)，也是他們決勝的關(guān)鍵?！? 吳雄昂的結(jié)論是，硅谷原則同樣能在中國(guó)贏得市場(chǎng)，不僅贏得市場(chǎng)，也贏得員工的心?？偨Y(jié)起來(lái)，“中國(guó)速度”可歸納為因應(yīng)本地市場(chǎng)和環(huán)境而調(diào)整的能力。 他也指出，用速度來(lái)驅(qū)動(dòng)中國(guó)和硅谷是不錯(cuò)，但也有導(dǎo)致失敗的可能。我趕緊同意。 不過(guò)，就我觀察，中國(guó)和硅谷還是存在著不同。中國(guó)的速度要快上好幾倍，主因是中國(guó)有著自己的“生態(tài)系統(tǒng)”──從零件供貨商、軟件開(kāi)發(fā)商到合約制造商都包含在內(nèi)。而他們所生產(chǎn)的產(chǎn)品也能以最快速度上市。 我們會(huì)繼續(xù)探討更多有關(guān)“中國(guó)速度 vs. 中國(guó)品質(zhì)”的故事。當(dāng)然，隨時(shí)歡迎你提供意見(jiàn)或分析──甚至包括你聽(tīng)到的傳聞──任何消息都非常歡迎。但是，在我們更了解彼此之前，讓我們都學(xué)著，別太過(guò)在意對(duì)方的評(píng)斷。 編譯： Joy Teng 本文下一頁(yè)：參考英文原文：Why TI does MCU designs in Shanghai，by Junko Yoshida，Dylan McGrath; China speed vs. China quality ，by Junko Yoshida

相關(guān)閱讀：
• 聯(lián)發(fā)科：讓千元智能機(jī)擁有iPhone 4S的功能
• 本土無(wú)晶圓廠(chǎng)業(yè)者們真正需要的策略
• http://www.zgwwsyw.com/ART_8800121647_1100_2700_0_4300_dde103af.HTMn0lesmc

{pagination} China speed vs. China quality Junko Yoshida It’s been only a month since I first traveled to China, and have started reporting on tidbits about China and things Chinese from the ground up. Many of my friends, colleagues and readers told me that a flood of comments posted by EE Times community members’ after each of my China stories is actually more revealing than my reporting. I take no offense, because I feel the same way. A former colleague of mine, now resident in Beijing, wrote to me this morning: “I am enjoying the comments your China stories are eliciting almost as much as the stories. Some are funny and some disturbing… A bit frightening is how little both sides understand each other.” I couldn’t agree more. China remains largely a mystery to most of us – me included. I still know very little, and the more people I interview, the less I feel as though I know their real stories. Similarly, I find Chinese people’s understanding is limited, especially when it comes to how fearful the U.S. engineering community is about China. Both sides remain caught up in pre-conceived notions of what the other side is like; they cherry-pick anecdotal evidence (which is often a valuable piece in a bigger picture), and use it to reinforce what they think they know. I think it’s time to take a deep breath and step back – for all of us. Here are a few good examples. Every time I write anything about China’s speed of design and production – which to me is one of the vital elements of Chinese success, I get reactions from our readers about the poor quality of China’s products. And this is a salient criticism. I get a steady flow of feedback – not just on our site, but in casual conversation – about an Android phone someone just bought in China; and already it’s going haywire. When I met in Beijing a vice president of engineering who works for a U.S. chip company, he was emphatic: “Junko, the quality of some of those phones is bad. But this isn’t because they can’t get it right; it’s because they don’t want to.” A little stumped by what I thought I heard him say, I asked, “What do you mean? Are you saying that they are intentionally making bad phones?” No, he said. “They are more interested in turning out more phones they can sell -- quickly. It’s just that they don’t want to take the time to make a perfect phone.” He pointed out the glaring exception to the stereotype: “Guess where all the iPhones are coming from? They are made by Foxconn in China. It’s not like Chinese don’t know how to manufacture good quality products.” Meanwhile, I received an e-mail this afternoon from Allen Wu, ARM China’s president. Referring to my story, "Why TI does MCU design in Shanghai," Wu wrote to me: “I fully agree with your comments on ‘China Speed.’ Products don’t necessarily need to be designed in China, but companies needs to respond in ‘China Speed’ to be successful,” especially as China becomes one of the leading markets for the chip industry, he explained. Wu noted, “If you look at the start-up culture in the Valley, I would put ‘speed,’ fast decision-making and response to market, as one of the clear top reasons for success --along with passion, commitment, etc.” Wu’s conclusion is that the principles of Silicon Valley work equally well in China, from winning markets to winning the hearts and minds of employees. It really comes down to the ability to adjust execution to local markets and environments in ‘China speed.’ His point that China and Silicon Valley are driven – perhaps to a fault -- by “speed” is well taken. I hasten to agree. What’s different, though, in my observation is that China’s speed is multiplied several times now, largely because China has elements in its ‘ecosystem’ – ranging from component suppliers to software developers and contract manufacturers – who can accelerate even further. Their end products eventually get to the market faster, too. The story of China speed vs. China quality is something we’re going to keep probing at EE Times. Your comments and analysis – even the anecdotal stuff – are always welcome. But let’s hold off, on both sides, being too judgmental about each other — until we’ve all learned a little more. Why TI does MCU designs in Shanghai Junko Yoshida SHANGHAI – Multinational chip companies who’ve consigned some of their design work to China tend to do so as quietly as possible; but a few leading companies are clearly making deeper inroads in China, and they’re not especially shy about it. Texas Instruments is a good example. Its newest MCU design center based in Shanghai recently saw its first locally designed device successfully taped out. While TI is not disclosing the size of its team, Scott Roller, vice president of TI Microcontrollers, in a recent interview with EE Times, described the MCU design center in Shanghai as “a sizeable team, with multiple designs going on.” TI’s Shanghai-based MCU design center, whose operation started in early 2011, is the newest addition to the company’s three others worldwide. They include design centers in Germany, Bangalore and Dallas. It’s important to note that TI’s facility in China is not there just to support existing MCU products. Rather, it actually executes some MCU product line development from China. “This is to develop MCUs – built in China for China,” said Roller. The design center performs every job – ranging from sales, application developers to system and processor designers and application field engineers – necessary to do the design work in one place. “We cover everything -- from front to back end,” said Roller. Scott Roller, vice president of microcontrollers at Texas Instruments Why design in China? Asked why design in China, Roller gave two straightforward reasons: “First, you can move much faster. Second, by designing it locally, there will be less room for misinterpretation.” Although in theory, that may make sense, not every multinational has gone that far in their public commitment to designing in China. Earlier this month at the Freescale Technology Forum in San Antonio, Texas, Gregg Lowe, Freescale Semiconductor's new president and CEO, said the China market is in transition from a low-cost manufacturer of electronics products designed in the West to a true high-tech hub with the design capability to create products for its massive domestic market. Chip vendors who want to sell products in China need a strong presence there, Lowe said. IC vendors don't necessarily need to design their products in China, but they need to have applications engineers and system engineers right there to be successful there, Lowe said. The fundamental difference in China, however, is China’s speed, said Allen Wu, president of ARM China. Design cycles for SoCs in China are generally much shorter, he observed. “From design starts to tape-out, sometimes it takes only five to six months. [Local companies] make decisions much more quickly and they react to the market very fast.” So, by the time multinationals finally reach a decision on a specific design back at headquarters, that may have already become irrelevant on the local Chinese market, explained Wu. TI’s Roller stressed that TI did not build its MCU design center in China “because of the cost.” He said that engineering talents may be “a bit cheaper but you don’t really save money.” TI has done this “because we want to design and produce products closer to where the demand is.” One of the biggest MCU market segments TI is after in China is a smart metering device based on China’s state grid program. TI taped out the first product and is sampling it now, said Roller. “This is a huge opportunity for us.” TI’s two key MCUs used in China include MSP430 ultra-low-power 16-bit microcontrollers and C2000 32-bit microcontrollers. Asked which other multinationals are commanding a strong presence in the MCU market in China, Roller mentioned Freescale and Renesas. Freescale – with a long history in China – may be pulling back a little, TI’s Roller said. But he quickly added: “That may change with the new CEO,” referring to Freescale's Lowe who was ex-TI executive. Freescale’s Lowe, while at TI, most recently led the Dallas-based company’s analog business. TI has already established several analog design centers in China. At the interview held at the Freescale Technology Forum, Freescale’s Lowe acknowledged "Design decisions will move." He said, "You can't just go to Silicon Valley anymore. That's still a very important place to go, but you need to be in all of the places where chips are being designed into systems to be close to customers and understand customer requirements."