摩爾定律如何預測了計算機的未來

SAN FRANCISCO — On April 19, 1965, just over 50 years ago, Gordon Moore, then the head of research for Fairchild Semiconductor and later one of the co-founders of Intel, was asked by Electronics Magazine to submit an article predicting what was going to happen to integrated circuits, the heart of computing, in the next 10 years. Studying the trend he’d seen in the previous few years, Moore predicted that every year we’d double the number of transistors that could fit on a single chip of silicon so you’d get twice as much computing power for only slightly more money. When that came true, in 1975, he modified his prediction to a doubling roughly every two years. “Moore’s Law” has essentially held up ever since — and, despite the skeptics, keeps chugging along, making it probably the most remarkable example ever of sustained exponential growth of a technology.

舊金山——就在50年前的1965年4月19日，《電子雜誌》(Electronics Magazine)請當時擔任仙童半導體公司(Fairchild Semiconductor)研發負責人、後來成爲英特爾(Intel)聯合創始人的戈登·摩爾(Gordon Moore)寫一篇文章，預測在接下來的10年裏，作爲計算核心的集成電路將會經歷什麼樣的發展。在對之前幾年看到的趨勢進行了研究後，摩爾預言，單個硅片上安放的晶體管數每年都會增加一倍，因此只需稍微多花點錢，就能獲得翻倍的計算能力。當這一點在1975年成爲現實後，摩爾將自己的預測改爲大致每兩年增加一倍。此後，“摩爾定律”(Moore’s Law)基本上一直得到應驗，而且儘管遭到了一些質疑，它仍在不斷推進。該定律因而可能是有史以來反映一項技術持續呈指數增長最著名的例子。

For the 50th anniversary of Moore’s Law, I interviewed Moore, now 86, at the Exploratorium in San Francisco, at a celebration in his honor co-hosted by the Gordon and Betty Moore Foundation and Intel. I asked him what he’d learned most from Moore’s Law having lasted this long.

在摩爾定律誕生50週年之際，我在舊金山探索博物館(Exploratorium)採訪了現已86歲高齡的摩爾。那是一場爲摩爾舉辦的慶祝活動，由戈登與貝蒂·摩爾基金會(Gordon and Betty Moore Foundation)同英特爾聯合舉辦。我問他，摩爾定律的有效性持續了這麼長時間，他最大的收穫是什麼。

“I guess one thing I’ve learned is once you’ve made a successful prediction, avoid making another one,” Moore said. “I’ve avoided opportunities to predict the next 10 or 50 years.”

“我想，我學到的一條是，一旦預測成功了，就要避免再做預測，”摩爾說。“我回避了預測接下來10年或50年會是什麼情況的各種機會。”

But was he surprised by how long it has been proved basically correct?

但在這麼長時間裏，摩爾定律都被證明基本正確，他覺得意外嗎？

“Oh, I’m amazed,” he said. “The original prediction was to look at 10 years, which I thought was a stretch. This was going from about 60 elements on an integrated circuit to 60,000 — a thousandfold extrapolation over 10 years. I thought that was pretty wild. The fact that something similar is going on for 50 years is truly amazing. You know, there were all kinds of barriers we could always see that [were] going to prevent taking the next step, and somehow or other, as we got closer, the engineers had figured out ways around these. But someday it has to stop. No exponential like this goes on forever.”

“啊呀，我挺吃驚的，”他說。“最初是預測未來10年的情況，這我都以爲是有所發揮了。也就是說，一個集成電路上的晶體管會從大約60個增加到6萬個——在10年時間裏提升到之前的1000倍。我以爲這已經很瘋狂了。類似的情形持續了50年，確實不可思議。你知道，我們總能發現，會有各種障礙阻止我們進入下一階段，可莫名其妙地，當我們越來越接近障礙時，工程師又找到了避開它們的辦法。但總有一天得停下來。像這樣的指數增長不會永遠持續下去。”

But what an exponential it’s been. In introducing the evening, Intel’s C.E.O., Brian Krzanich summarized where Moore’s Law has taken us. If you took Intel’s first generation microchip, the 1971 4004, and the latest chip Intel has on the market today, the fifth-generation Core i5 processor, he said, you can see the power of Moore’s Law at work: Intel’s latest chip offers 3,500 times more performance, is 90,000 times more energy efficient and about 60,000 times lower cost.

然而，這是一種多麼驚人的指數增長啊。在主持當晚的活動時，英特爾首席執行官科再奇(Brian Krzanich)對摩爾定律對我們的影響行了總結。他說，看看英特爾的1971年發佈第一代微芯片4004和目前市場上的最新一代英特爾芯片，也就是第五代酷睿i5(Core i5)處理器，就能看到摩爾定律的力量：最新一代芯片的性能提高了3500倍，能耗是原來的九萬分之一，成本降至先前的六萬分之一。

To put that another way, Krzanich said Intel engineers did a rough calculation of what would happen had a 1971 Volkswagen Beetle improved at the same rate as microchips did under Moore’s Law: “Here are the numbers: [Today] you would be able to go with that car 300,000 miles per hour. You would get two million miles per gallon of gas, and all that for the mere cost of 4 cents! Now, you’d still be stuck on the [Highway] 101 getting here tonight, but, boy, in every opening you’d be going 300,000 miles an hour!”

換個角度來看呢？科再奇表示，英特爾的工程師進行了大致的計算，如果按照摩爾定律下微芯片的發展速度，1971年推出的一款大衆甲殼蟲(Volkswagen Beetle)汽車的性能會有怎樣的改進：“數據如下：（現在）這款車能開到每小時30萬英里（約合48萬公里）。1加侖汽油（約合4升）可以跑200萬英里，而且只要花4分錢！時下，雖然今晚你可能還是會在趕來這裏的途中堵在101公路上，但天哪，只要能動，你們可是會開到30萬英里的時速呢！”

What is most striking in Moore’s 1965 article is how many predictions he got right about what these steadily improving microchips would enable. The article, entitled “Cramming More Components Onto Integrated Circuits,” argued that: “Integrated circuits will lead to such wonders as home computers — or at least terminals connected to a central computer — automatic controls for automobiles, and personal portable communications equipment. The electronic wristwatch needs only a display to be feasible today. ... In telephone communications, integrated circuits in digital filters will separate channels on multiplex equipment. [They] will also switch telephone circuits and perform data processing.”

摩爾1965年的那篇文章中最令人吃驚的是，他對這些穩步改善的微芯片所能帶來的變化所做的許多預測都是準確的。題爲《讓集成電路填滿更多元件》(Cramming More Components Onto Integrated Circuits)的這篇文章稱：“集成電路將催生奇蹟，比如家用電腦——或者至少是與中央計算機相連的終端——汽車自動控制和個人便攜式通信設備。如今電子手錶只需要一個顯示屏就可以了……電話通信領域，數字濾波器中的集成電路將在多路傳輸設備上有單獨的通道。[它們]還將轉換電話線路，進行數據處理。”

Moore pretty much anticipated the personal computer, the cellphone, self-driving cars, the iPad, Big Data and the Apple Watch. How did he do that? (The only thing he missed, I jokingly told him, was “microwave popcorn.”)

基本上，摩爾這是預言了個人電腦、手機、自動駕駛汽車、iPad、大數據和Apple Watch的出現。他是怎麼做到的？（我跟摩爾開玩笑，他唯一漏掉的就是“微波爐爆米花”了。）

“Well,” said Moore, “I had been looking at integrated circuits — [they] were really new at that time, only a few years old — and they were very expensive. There was a lot of argument as to why they would never be cheap, and I was beginning to see, from my position as head of a laboratory, that the technology was going to go in the direction where we would get more and more stuff on a chip and it would make electronics less expensive. ... I had no idea it was going to turn out to be a relatively precise prediction, but I knew the general trend was in that direction and had to give some kind of a reason why it was important to lower the cost of electronics.”

“這個嘛，”摩爾回答，“我一直在研究集成電路——[它們]當時還是新生事物，剛出現沒幾年——而且非常昂貴。有很多觀點認爲，它們永遠都不會便宜下來。從一名實驗室負責人的角度，我開始發現，這種技術未來的方向是，一枚芯片上會容納越來越多的東西，而這會讓電子產品變得更便宜……我不知道這會是一個比較準確的預測，但我知道大體趨勢是這樣，而它一定會提供降低電子產品成本的某種理由。”

Can it continue? Every year someone predicts the demise of Moore’s Law, and they’re wrong. With enough good engineers working on it, he hoped, “we won’t hit a dead end. ... It’s [a] unique technology. I can’t see anything really comparable that has gone on for this long a period of time with exponential growth.”

這樣的趨勢還會繼續嗎？每年都有人預測摩爾定律將被打破，但他們都錯了。摩爾希望，憑藉着足夠多的優秀工程師的努力，“我們不會走投無路……這是[一項]獨特的技術。我沒有發現任何能與之類比的技術可以持續這麼長時間的指數增長。”

But let’s remember that it was enabled by a group of remarkable scientists and engineers, in an America that did not just brag about being exceptional, but invested in the infrastructure and basic scientific research, and set the audacious goals, to make it so. If we want to create more Moore’s Law-like technologies, we need to invest in the building blocks that produced that America.

不過，別忘了它是靠一批傑出的科學家和工程師來實現的，而那時的美國也不會只知道炫耀自己的卓越，而是會投資於基礎設施和基礎科學研究，設置大膽的目標，使之成爲現實。如果我們想要創造更多符合摩爾定律的技術，我們就需要投資於成就了當時的美國的基本要素。

Alas today our government is not investing in basic research the way it did when the likes of Moore and Robert Noyce, the co-inventor of the integrated circuit and the other co-founder of Intel, were coming of age.

可惜啊，今天的政府不像摩爾和羅伯特·諾伊斯(Robert Noyce)成長的時代那樣投資基礎科學研究了。諾伊斯是集成電路的聯合發明人，與摩爾一起創立了英特爾。

“I’m disappointed that the federal government seems to be decreasing its support of basic research,” said Moore. “That’s really where these ideas get started. They take a long time to germinate, but eventually they lead to some marvelous advances. Certainly, our whole industry came out of some of the early understanding of the quantum mechanics of some of the materials. I look at what’s happening in the biological area, which is the result of looking more detailed at the way life works, looking at the structure of the genes and one thing and another. These are all practical applications that are coming out of some very fundamental research, and our position in the world of fundamental science has deteriorated pretty badly. There are several other countries that are spending a significantly higher percentage of their G.N.P. than we are on basic science or on science, and ours is becoming less and less basic.”

“讓我感到失望的是，聯邦政府似乎在減少對基礎科研的支持，”摩爾說。“基礎科研真的是這些想法誕生的地方。它們需要很長時間來孕育，但最終會促成驚人的進步。要知道，我們整個行業都來自於對一些材料的量子力學的早期認識。我觀察了一下生物領域目前的發展，也是來自對生命的原理和基因的結構等一系列問題的細緻研究。這些都是一些非常基礎的研究帶來的實際應用，而我們在基礎科學領域的地位在嚴重下滑。其他一些國家在基礎科學或大的科學領域上的投入與國民生產總值的比值，要遠遠高於我們，而我們的研究越來越脫離基礎科學。”

How did he first get interested in science, I asked?

我問他，一開始是怎麼對科學產生興趣的？

“My neighbor got a chemistry set and we could make explosives,” he said. “In those days, chemistry sets had some really neat things in them, and I decided about then I wanted to be a chemist not knowing quite what they did, and I continued my work in a home laboratory for some period of time. Got to the point where I was turning out nitroglycerin in small production quantities and turning it to dynamite. ... A couple ounces of dynamite makes a marvelous firecracker. That really got my early interest in it. You couldn’t duplicate that today, but there are other opportunities. You know, I look at what some of my grandkids are doing, for example, those robotics and the like. These are spectacular. They’re really making a lot of progress.”

“我的鄰居有一套簡易化學實驗玩具，我們可以做炸藥，”他說。“那個時候的實驗玩具裏有很棒的東西，我大概就是當時決定要當一名化學家，雖然並不知道化學家究竟是做什麼的。我後來在家庭實驗室裏繼續倒騰，做了一段時間。然後我就可以少量地生產硝化甘油了，把它們變成炸藥……幾盎司的炸藥就能做成一個很棒的鞭炮。這真的讓我早早對它產生了興趣。你不可能在今天覆制這一切，但現在有其他的機會。比如，我看到孫輩在鼓搗機器人之類的東西。這些都很棒。他們的確取得了不少進步。”

Looking back on Moore’s Law and the power of computing that it has driven, I asked Moore what he thought was its most important contribution over the past 50 years.

我問摩爾，回過頭來看摩爾定律以及受它驅動的計算能力，他本人認爲它在過去50年裏的最大貢獻是什麼？

“Wow!” he said. “You know, just the proliferation of computing power. We’ve just seen the beginning of what computers are going to do for us.”

“哇！”他說。“這個，就是計算能力的爆炸性增長。關於計算機會爲我們做什麼，我們目前看到的不過是個開頭。”

How so?

何以見得？

“Oh, I think incrementally we see them taking over opportunities that we tried to do without them before and were not successful,” he added. “It’s kind of the evolution into the machine intelligence, if you wish, and this is not happening in one step. To me, it’s happening in a whole bunch of increments. I never thought I’d see autonomous automobiles driving on the freeways. It wasn’t many years ago [they] put out a request to see who could build a car that could go across the Mojave Desert to Las Vegas from a place in Southern California, and several engineering teams across the country set out to do this. Nobody got more than about 300 yards before there was a problem. Two years later, they made the full 25-mile trip across this desert track, and which I thought was a huge achievement, and from that it was just a blink before they were driving on the freeways. I think we’re going to see incremental advances like that in a variety of other areas.”

“恩，我想我們會逐漸看到，以前沒有計算機時我們試圖去抓住但卻沒能成功的一些機會，現在正被計算機拿去實現，”他補充道。“有點像是逐漸演化出機器智能，如果你希望如此的話；而這個過程並非一蹴而就。在我看來，它是一大堆進步的累積。我以前從沒想過會在高速公路上見到無人駕駛汽車。就在沒多少年前，[他們]發出徵集令，看誰能造出一輛無人駕駛汽車，讓它從加州南部出發，穿過莫哈維沙漠，抵達拉斯維加斯。來自全國各地的幾個工程師團隊開始着手做這件事。起初，沒有哪個團隊的車子能開出300碼不出問題。兩年後，車子就可以在沙漠裏整整行駛25英里。我當時認爲那是巨大的進步。再一眨眼，它們已經在高速公路上穿行了。我認爲，我們會在其他各種領域看到這類日積月累的進步。”

Did he worry, I asked Moore, whose own microprocessors seemed as sharp as ever, that machines would really start to replace both white-collar and blue-collar labor at a scale that could mean the end of work for a lot of people?

摩爾的思維看起來與昔日一樣敏捷。我問他，是否擔心，機器真的開始大規模取代藍領和白領勞動者，致使一大批人失去工作？

“Don’t blame me!” he exclaimed! “I think it’s likely we’re going to continue to see that. You know, for several years, I have said we’re a two-class society separated by education. I think we’re seeing the proof of some of that now.”

“別怪我！”他大聲說道。“我想，我們可能會持續看到這種現象。你知道，我們的社會以教育程度爲界分成兩個階層，這話我已經說了好幾年。我想，我們現在看到了這方面的一些證據。”

When was the moment he came home and said to his wife, Betty, “Honey, they’ve named a law after me?”

他第一次回到家告訴妻子貝蒂，“親愛的，他們用我的名字命名了一條定律”是在什麼時候？

Answered Moore: “For the first 20 years, I couldn’t utter the terms Moore’s Law. It was embarrassing. It wasn’t a law. Finally, I got accustomed to it where now I could say it with a straight face.”

摩爾答道：“頭20年，我一直無法把‘摩爾定律’這幾個字說出口。太尷尬了。它不是什麼定律。最後我終於習以爲常了，現在我可以面不改色心不跳地提到它。”

Given that, is there something that he wishes he had predicted — like Moore’s Law — but did not? I asked.

我問，既然如此，有沒有什麼是他希望自己能預測到——就像摩爾定律那樣——但實際上並未預測到的？

“The importance of the Internet surprised me,” said Moore. “It looked like it was going to be just another minor communications network that solved certain problems. I didn’t realize it was going to open up a whole universe of new opportunities, and it certainly has. I wish I had predicted that.”

“互聯網的重要性讓我感到驚訝，”摩爾說。“當初看來，它不過是會成爲一種用於解決特定問題的不太重要的通信網絡。我沒意識到它會帶來無數新機會，但它顯然做到了。我真希望自己當年能預測到這種情況。”