Have we reached the limit of computer power? - Sajan Saini and George Zaidan

翻译人员: Yip Yan Yeung 校对人员: XU Gloria

In the Netherlands, there’s an ambitious company

在荷兰， 有一家雄心勃勃的公司

that builds one of the most advanced and expensive tools in the world:

正在制造世界上最先进、 最昂贵的工具之一：

a single unit costs hundreds of millions of dollars.

每台成本高达数亿美元的计算机设备。

And when companies buy one,

如果有一家公司购买了一台设备，

they also need 250 engineers to install the 165-ton device

还需要 250 名工程师来安装 这台重达 165 吨的设备，

in a process that typically takes half a year.

通常耗时需要半年。

But despite this steep cost in time and money,

但是，尽管要耗费大量时间和金钱，

many microchip makers desperately want one of these machines.

许多微芯片制造商 还是迫切想要一台这样的机器。

The hundred-million-dollar question is: why?

这价值上亿美元的问题是：为什么？

The answer has to do with something called Moore’s Law.

答案与所谓的“摩尔定律”有关。

First coined by Intel co-founder Gordon Moore,

该定律最初由英特尔联合创始人 戈登·摩尔提出，

this law states that every 1 to 2 years

指出每隔 1 至 2 年，

the number of transistors that can fit on a given size computer chip will double.

一定尺寸的计算机芯片上 可以容纳的晶体管数量会翻倍。

And by extension, the rough number of calculations that chip can do per second

推而广之，芯片每秒 能够进行的粗略计算量

will also double.

也将翻一番。

Now, this law isn't a physical law like gravity.

这个定律不是像重力这样的物理定律。

It’s just a trend Moore observed during the early 1960s.

它只是摩尔在 1960 年代初 观察到的一种趋势。

But chipmakers turned that trend into a goal, and in turn,

但是芯片制造商将这种趋势 视作目标，于是反过来，

consumers learn to expect computing progress to continue

消费者学会了期待计算 将持续以这种指数级的速度发展。

at this exponentially fast pace.

And the amazing thing is, for six decades, it pretty much has.

令人惊奇的是，六十年来， 它差不多就是这么发展的。

Thanks to Moore’s Law, chips have gotten smaller, faster,

由于摩尔定律， 芯片变得更小、更快、

more efficient, and cheaper.

更高效、更便宜。

But today, there are four key problems that trip up this trend,

但是如今，有四个关键问题 阻碍了这一趋势，

potentially ending Moore’s Law

有可能会终结摩尔定律，

and fundamentally changing how we make progress in computing.

彻底改变我们在计算领域 取得进展的方式。

The first is transistor size.

首先是晶体管的尺寸。

Transistors are basically on/off switches,

晶体管其实就是开关，

and these building blocks of digital computing have been shrinking

自 1960 年代以来， 这个数字计算的构件一直在缩小。

since the 1960s.

But recently they’ve gotten so small,

但是最近它们变得如此之小，

quantum physics has begun to interfere with their functions.

量子物理开始干扰它们的功能。

When a transistor’s switch, or gate, is less than 20 nanometers,

当晶体管的开关或栅极 小于 20 纳米时，

electrons will tunnel along it continuously,

电子将不断流过这个通道，

turning a crisp on/off switch into a hazy dimmer.

将本该清晰开或关的开关 变成一个模糊的渐变开关。

The second problem is heat.

第二个问题是热量。

As chipmakers make components smaller and more complex,

随着芯片制造商 制造的组件更小、更复杂，

the copper lines that run between them need to be thinner and longer.

组件之间的铜线必须更细、更长。

This increases their electrical resistance and generates high heat

这会增加电阻、产生高温，

that impairs chip performance and can’t be easily dissipated.

从而损害芯片性能 且散热并不容易。

Today’s chips can already run hot enough to cook an egg,

当今的芯片运行的温度 已经高到可以煎鸡蛋，

and temperatures are only predicted to increase without new innovations.

如果没有更多创新， 预计温度只会继续升高。

While both these issues represent limits in the fundamental physics of chipmaking,

尽管这两个问题都代表了 芯片制造在基础物理上的局限性，

researchers haven't stopped trying to solve them.

但研究人员并没有停止 尝试解决这些问题。

Unfortunately, their solutions often exacerbate the third major problem:

不幸的是，他们的解决方案 往往会加剧第三个主要问题：

chipmaking’s environmental impact.

芯片制造对环境的影响。

For example, swapping copper lines for ruthenium

例如，将铜线换成钌

could help pack transistors more tightly and keep chips smaller—

有助于更密集地封装晶体管， 让芯片更小，

but that metal is far scarcer than copper

但是这种金属比铜稀缺得多，

and would require new mining infrastructure.

需要新的采矿基础设施。

Similarly, the technology currently used to make today’s smallest transistors

同样，目前用于 制造当今最小晶体管的技术

requires huge amounts of energy and chemicals

需要大量的能量和化学物质，

called perfluoroalkyl and polyfluoroalkyl substances,

称为全氟烷基和多氟烷基物质，

which can take thousands of years to break down in the environment.

这些物质可能需要数千年 才能在环境中降解。

Managing these first three problems contributes to the final issue: cost.

要解决前三个问题 会导致最后一个问题：成本。

To keep achieving Moore’s Law,

为了继续实现摩尔定律，

chipmakers have to make individual chip components smaller.

芯片制造商必须缩小每一个芯片组件。

And this is where that costly $400 million machine comes in.

这就说到了那台 耗资 4 亿美元的天价机器。

This marvel of chip-making science shoots a stream of tin droplets

这个芯片制造科学的奇迹 将一注锡滴

into a vacuum chamber before blasting them with a high-energy laser

射入真空室，然后用高能激光

that vaporizes the tin to create plasma.

将其炸开，使锡蒸发， 产生等离子体。

In turn, the plasma emits a 13.5 nanometer wavelength of ultraviolet light

等离子体转而发射 波长为 13.5 纳米的紫外线，

that can be used to produce incredibly small transistors.

可用于生产非常小的晶体管。

This remarkable feat of engineering has helped chipmakers

这一非凡的工程壮举 帮助芯片制造商

keep up with Moore’s Law.

跟上了摩尔定律。

But as chips keep getting denser,

但随着芯片的密度不断提高，

intricate manufacturing plants keep getting more expensive.

精密的制造工厂越来越昂贵。

This trend has been so consistent, it’s actually earned the nickname

这种趋势一直很稳定，甚至取得了

Moore’s Second Law.

“摩尔第二定律”的绰号。

Obviously, all these trajectories are unsustainable.

显然，这些发展趋势 全都是不可持续的。

Manufacturing plants can’t keep increasing in price,

制造工厂不能继续涨价，

our ecosystems can’t endure endless mining and pollution,

我们的生态系统 无法承受无止境的采矿和污染，

and the laws of physics are unlikely to change anytime soon.

物理定律不太可能在短期内改变。

Fortunately, Moore's Law is flexible,

幸运的是，摩尔定律很灵活，

and there’s no reason we can’t introduce new goals

我们依然可以引入新的目标 来取得负责任的计算进步。

to keep making computing progress responsibly.

Perhaps we could introduce a new Sustainability Law?

也许我们可以提出 新的可持续发展定律？

Smaller transistors already use less material and produce less e-waste,

更小的晶体管使用更少的材料， 产生更少的电子垃圾，

and advancements in electronic-photonic integration

电子光子集成的进步

are allowing chips to use less energy and generate less heat.

使芯片消耗更少的能量， 产生更少的热量。

So perhaps chips should be made twice as sustainable every several years?

那么，也许芯片的可持续性 应该每隔几年翻一番呢？

Whatever the answer is,

不管答案是什么，

we make the laws, so the future is up to us.

我们提出定律， 所以未来在我们手中。